Method and system for determining an emotion or sensory perception related to exposure to a fragrance or aroma component

Abstract

A method (400) for determining an emotion or sensory perception related to exposure to a fragrance or aroma component, comprising a step (405) of inputting at least two digital representation identifiers of physical components of a fragrance or aroma on a computer interface, wherein the obtained input corresponds to a physical composition digital identifier representing the physical composition of the physical fragrance or aroma component; a step (410) of operating a computing device to associate at least one value with the input physical composition digital identifier, the value being a relative ranking for at least one determined perceived emotion or sensation among the input digital representation identifiers of the physical components of the fragrance or aroma, a value representing the perception for at least one determined perceived emotion or sensation for at least one input digital representation identifier of a physical component of a fragrance or aroma, and / or a value representing a class of fragrance or aroma components in the classification of fragrance or aroma components by the perceived emotion or sensation perception for at least one determined emotion or sensory perception, wherein the step (410) represents; a step (415) of providing on a computer interface at least one value obtained during the step of operating for the input physical composition digital identifier; and comprising.

Description

Technical Field

[0001] The present invention aims at a method and a system for determining emotional or sensory perception related to exposure to flavor or fragrance components.

[0002] It is particularly applicable to the flavor or fragrance industry.

[0003] Background Art In the field of flavor and fragrance design and production, one of the important industrial challenges is to provide flavors and fragrances that contribute to the commercial success of products. Such products range from dairy products to soft drinks and perfumes. The commercial success of a product depends, in an important part, on the psychophysical reaction of consumers to the flavor or fragrance of the product. Such psychophysical reactions are related to the emotional or sensory perception of the flavor or fragrance. Such emotions or sensory perceptions are typically associated with words (e.g., "energizing" or "romantic" scents).

[0004] Since the underlying reasons for psychophysical reactions can be complex to identify, at a broader level, basic psychophysical emotional or sensory reactions can be captured. Such psychophysical emotional or sensory reactions correspond, for example, to an explicit binary output from consumers (i.e., the consumer "likes" or "dislikes" the product, flavor, or fragrance).

[0005] In traditional approaches, before the mass industrialization of a product, flavor, or fragrance, in order to capture psychophysical emotional or sensory reactions, opinions are solicited from panelists representing the target audience of the product, flavor, or fragrance, and the product, flavor, or fragrance is presented. Based on this feedback, the product, flavor, or fragrance can enter production, be redesigned, or be completely discarded.

[0006] Such an approach is costly in terms of both the time required and the corresponding financial expenditure. Further, such an approach may provide inconsistent results due to the limited sample size of the panelists involved. Finally, such results may be inaccurate considering the lack of modeling of important parameters that affect (bias) the human response to the digital identifier of the physical composition of the fragrance or aroma.

[0007] For these reasons, current approaches are insufficient to reliably evaluate the psychophysical emotions or sensory responses of humans presented with a product, fragrance, or aroma.

[0008] Summary of the Invention The present invention aims to address all or some of these drawbacks.

[0009] According to a first aspect, the present invention aims to provide a method for determining an emotional or sensory perception related to an exposure to a fragrance or aroma component, the method comprising: inputting, on a computer interface, at least two digital representation identifiers of physical components of a fragrance or aroma, wherein the resulting input corresponds to a physical composition digital identifier representing a physical composition of a physical fragrance or aroma component; operating, by a computing device, instructions configured to associate at least one value with the input physical composition digital identifier, the value being: a relative ranking for at least one determined perceived emotion or sensation among the input digital representation identifiers of physical components of the fragrance or aroma; a value representing a perception for at least one determined perceived emotion or sensation with respect to at least one of the input digital representation identifiers of physical components of the fragrance or aroma, and / or a value representing a class of the fragrance or aroma component in the classification of the fragrance or aroma component by the perceived emotion or sensation perception for at least one determined emotion or sensation perception; and For an input physical composition digital identifier, providing, on a computer interface, at least one value obtained during the step of operating including.

[0010] Such measures enable an accurate determination of the actual physical reactions of humans exposed to a composition assembled (or embodied) according to the digital representation of said composition.

[0011] Such measures also enable the use of other algorithms in various scenarios all derived from the initial training of a trained neural network device or said neural network device. For example, in one scenario, a neural network device may be used to rank components with respect to their contribution to the perception of a particular human emotion or sensation. Such ranking may then be used by fragrance and flavor manufacturers when designing a physical composition digital identifier intended to evoke said human emotion or sensation. In another scenario, an absolute value representing the intensity of a perceived human emotion or sensation may be derived. Such a value may also be used by fragrance and flavor manufacturers when designing a physical composition digital identifier. In a third scenario, components may be classified into emotion or sensation impact categories that assist fragrance and flavor manufacturers when designing a physical composition digital identifier.

[0012] In certain embodiments, the step of operating is configured to operate a trained gradient boosting decision tree device on an input physical composition digital identifier.

[0013] Such measures enable the prediction of the psychophysical emotions or sensory responses of hidden physical composition digital identifiers without relying on panelists using a trained gradient boosting decision tree device. This significantly reduces the time and cost required to evaluate the emotional or sensory responses of consumers of designed physical composition digital identifiers or products incorporating physical composition digital identifiers. Furthermore, these measures enable a reliable prediction of the psychophysical emotions or sensory responses of said consumers.

[0014] In certain embodiments, the method subject matter of the present invention providing a set of exemplary data to a gradient boosting decision tree device or to a neural network device, the set of data at least two physical composition digital identifiers of at least two fragrances or aromas formed by at least two digital representations of physical components of the fragrance or aroma, and at least one emotional or sensory perception value for at least two of the fragrance or aroma components, the emotional or sensory perception value being associated with at least one emotional or sensory perception digital identifier, the emotional or sensory perception digital identifier representing a category of human emotional or sensory responses to the physical composition digital identifier embodied from among a finite list of emotional or sensory responses, at least one emotional or sensory perception value comprising, the step of operating a gradient boosting decision tree device or a neural network device based on the set of exemplary data, and obtaining a trained gradient boosting decision tree model or a trained neural network model comprising.

[0015] Such measures enable the prediction of psychophysical emotional or sensory responses of hidden physical composition digital identifiers without relying on panelists using a trained gradient boosting decision tree device. This significantly reduces the time and cost required to evaluate the emotional or sensory responses of consumers of designed physical composition digital identifiers or products incorporating physical composition digital identifiers. Further, these measures enable a reliable prediction of the psychophysical emotional or sensory responses of said consumers.

[0016] Preferably, any neural network used in a ranking configuration can be used, as opposed to a regular regression neural network model.

[0017] In certain embodiments, the exemplary data further includes at least one digital identifier associated with at least one physical composition digital identifier, and the at least one digital identifier is the gender of a human exposed to the embodied physical composition, the country of origin of a human exposed to the embodied physical composition, the type of use of the embodied physical composition, the compositional chemical base used to support the embodied physical composition, and / or the dosage of at least one physical component fragrance or aromatic physical component represented by the corresponding digital representation identifier and represents.

[0018] Such criteria affect the emotions or sensations perceived by exposure to the fragrance or aromatic physical composition digital identifier and are thus valuable parameters for training the gradient boosting decision tree device.

[0019] In certain embodiments, the method according to the present invention comprises, downstream of the step of operating, determining, by a computing device, a numerical value representing the influence of an emotional or sensory response to at least one inputted flavor or fragrance component, and providing, on a computer interface, the determined numerical value of the influence of the response.

[0020] In certain embodiments, the method according to the present invention comprises, downstream of the step of determining a numerical value representing the influence of an emotional or sensory response, providing at least one alternative digital representation identifier of a physical component of a flavor or fragrance for at least one inputted digital representation identifier of a physical component of a flavor or fragrance, the alternative digital representation identifier of a physical component of a flavor or fragrance being provided as a function of values representing the influence of the emotional or sensory response associated with the inputted and alternative digital representation identifiers of a physical component of a flavor or fragrance, for forming an alternative digital identifier of a physical composition.

[0021] In certain embodiments, the step of providing at least one alternative digital representation identifier of a physical component of a flavor or fragrance is configured to further provide at least one value representing the concentration of at least one of the alternative digital representation identifiers of a physical component of a flavor or fragrance.

[0022] Such embodiments enable an improved design of digital identifiers of physical compositions of flavors or fragrances, and enable perfumers and flavorists to dynamically select and improve digital identifiers of physical compositions.

[0023] In certain embodiments, the step of providing at least one alternative digital representation identifier of a physical component of a flavor or fragrance is configured to further provide a minimum value and / or a maximum value representing the concentration of at least one of the alternative digital representation identifiers of a physical component of a flavor or fragrance.

[0024] Such embodiments enable the proposal of components already present in the digital identifier of a physical composition, together with an adjusted concentration of a new component having a specific concentration.

[0025] Such an embodiment enables the provision of a value representing a threshold of performance of said alternative component.

[0026] In certain embodiments, the method subject of the present invention includes the step of assembling a physical composition corresponding to an input physical composition digital identifier or providing an input physical composition digital identifier to a system configured to assemble a physical composition.

[0027] Such an embodiment enables the realization of a designed physical composition digital identifier.

[0028] In certain embodiments, the method subject of the present invention comprises the step of assembling a database associating values representing measured human emotions or sensory perceptions for at least one group of at least two fragrance or aroma physical component digital representation identifiers, and the step of determining, by a computing device, a set of instructions to be executed during the step of operating and includes.

[0029] Such an embodiment enables the construction of a database that accurately represents the reality of emotions or sensations perceived by a human exposed to the composition.

[0030] In certain embodiments, the step of assembling the database comprises the step of exposing at least one human to a physical composition of a fragrance or aroma physical component, measuring the emotions or sensory perceptions of said at least one human exposed to said physical composition, and recording in the database a value representing the measured perceived emotion or sensation in association with a group of fragrance or aroma physical component digital representation identifiers representing the physical components used for the composition used during the exposing step and includes.

[0031] In certain embodiments, the method of the present invention includes the step of replacing an input digital representation identifier of a physical component of a flavor or aroma with a different equivalent digital representation identifier of a physical component of a flavor or aroma, said equivalence being defined in a database of equivalent digital representation identifiers of physical components of flavors or aromas.

[0032] Such embodiments allow for the dynamic substitution of components that are equivalents, and the use of one of said equivalent components is criticized, particularly for ecological reasons.

[0033] In certain embodiments, the step of operating comprises a first step of associating, by a computing device, at least one olfactory or gustatory descriptor with at least one input digital representation identifier of a physical component of a flavor or aroma, and a second step of associating, by a computing device, at least one perceived emotion or sensation as a function of at least one olfactory or gustatory descriptor associated with at least one input digital representation identifier of a physical component of a flavor or aroma and including.

[0034] Such embodiments allow for an indirect approach in which the method first determines descriptors associated with the components and then transforms these descriptors with the perceived emotions or sensations.

[0035] In certain embodiments, at least one emotion or sensation perception represents the perception of health or hygiene benefits associated with exposure to a physical composition.

[0036] According to a second aspect, the present invention aims at a system for determining an emotion or sensation perception related to exposure to a flavor or aroma component, the system comprising a step of inputting, on a computer interface, at least two digital representation identifiers of physical components of a flavor or aroma, the resulting input corresponding to a digital identifier of a physical composition representing a physical composition of a physical flavor or aroma component, the step and Operating a computing device to execute instructions configured to associate at least one value with an input physical composition digital identifier, the value being Among the input digital representations of the physical components of the aroma or fragrance, at least one relative ranking for a determined perceived emotion or sensation, For at least one input digital representation of the physical components of the aroma or fragrance, a value representing a perception for at least one determined perceived emotion or sensation, and / or A value representing a class of aroma or fragrance components in the classification of aroma or fragrance components by the perceived emotion or sensation perception for at least one determined emotion or sensation perception, And, Providing, on a computer interface, at least one value obtained during the operating step for the input physical composition digital identifier Including.

[0037] Such measures provide benefits similar to those of the first aspect of the present invention.

[0038] Other advantages, objects, and specific features of the present invention will become apparent from the following non-exhaustive description of at least one specific embodiment of the present invention in connection with the accompanying drawings.

Brief Description of the Drawings

[0039]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

[0040] Mode for Carrying Out the Invention Since each feature of one embodiment can be combined advantageously with any other feature of any other embodiment, this description is not exhaustive.

[0041] Various concepts of the present invention can be embodied as one or more methods, and examples thereof are provided. The operations performed as part of the method may be ordered in any suitable manner. Thus, although shown as consecutive operations in the exemplary embodiments, embodiments can be constructed in which the operations are performed in a different order than the example, including performing some operations simultaneously.

[0042] As used in this specification and the claims, the phrase "and / or" should be understood to mean "either or both" of the elements so joined, i.e., elements that may be present conjunctively in some cases and disjunctively in other cases. A plurality of elements listed with "and / or" should likewise be construed as "one or more" of the elements so joined. Other elements may optionally be present, whether or not specifically identified in relation to the elements specifically identified by the "and / or" clause. Thus, by way of non-limiting example, reference to "A and / or B", when used in combination with open-ended language such as "comprising", may in one embodiment refer to only A (optionally including elements other than B), in another embodiment refer to only B (optionally including elements other than A), and in yet another embodiment refer to both A and B (optionally including other elements), and so on.

[0043] As used in this specification and the claims, "or" should be understood to be inclusive.

[0044] As used in this specification and the claims, the phrase "at least one" in reference to a list of one or more elements means at least one element selected from any one or more of the elements in the list of elements, but does not necessarily include at least one of every element specifically listed in the list of elements, and is not to be construed as excluding any combinations of elements in the list of elements. This definition also allows for the possibility that elements other than those specifically identified in the list of elements referred to by the phrase "at least one" may optionally be present, whether or not they are related to specifically identified elements. Thus, by way of non-limiting example, "at least one of A and B" (or equivalently, "at least one of A or B", or equivalently, "at least one of A and / or B") can, in one embodiment, refer to at least one A, optionally including two or more, where B is absent (and optionally including elements other than B); in another embodiment, it can refer to at least one B, optionally including two or more, where A is absent (and optionally including elements other than A); and in yet another embodiment, it can refer to at least one A, optionally including two or more, and at least one B, optionally including two or more (and optionally including other elements), and so on.

[0045] In the claims and the above specification, all transitional phrases such as "comprising", "including", "carrying", "having", "containing", "involving", "holding", "composed of", etc. are to be understood to be open-ended, i.e., to mean including but not limited to. Only the transitional phrases "consisting of" and "consisting essentially of" are to be considered limiting or semi-limiting transitional phrases, respectively.

[0046] Note that at this point, the drawings are not to scale.

[0047] As used herein, the term "component" preferably refers to any component that exhibits flavor or aroma capabilities. The terms "compound" or "component" refer to the same item as "component". A component may be formed from one or more chemical molecules. Thus, a component may be considered "simple" if it consists mostly of a single compound component, or "complex" if it contains a plurality of compounds within a set ratio range between these compounds.

[0048] The term "physical composition digital identifier" refers to a liquid, solid, and / or gaseous aggregate of at least one component.

[0049] As used herein, "flavor" refers to the olfactory perception resulting from the sum of the activation, enhancement, and inhibition (if present) of (one or more) odorant receptors by at least one volatile component, via orthonasal and retronasal olfaction and the activation of taste buds containing taste receptor cells. Thus, by way of example, and not intending to limit the scope of the present disclosure in any way, "flavor" results from the sum of a first volatile component that activates odorant receptors or taste buds associated with the coconut note, a second volatile component that activates odorant receptors or taste buds associated with the celery note, and a third volatile component that inhibits odorant receptors or taste buds associated with the hay note, resulting in olfactory and taste bud perceptions.

[0050] As used herein, "aroma" refers to the olfactory perception resulting from the sum of activation, enhancement, and inhibition (if present) of odorant receptors by at least one volatile component. Thus, by way of example, and not intending to limit the scope of the present disclosure in any way, "aroma" results from the olfactory perception resulting from the sum of a first volatile component that activates odorant receptors associated with the tonality of coconut, a second volatile component that activates odorant receptors associated with the tonality of celery, and a third volatile component that inhibits odorant receptors associated with the tonality of hay.

[0051] As used herein, the term "means for inputting" is, for example, a keyboard, a mouse, and / or a touch screen adapted to interact with a computing system to collect user input. In a variant form, the means for inputting is essentially logical, such as a network port of a computing system configured to receive electronically transmitted input commands. Such input means may be associated with a GUI (Graphic User Interface) or an API (Application programming interface) presented to the user. In other variants, the means for inputting may be a sensor configured to measure specific physical parameters associated with the intended use case. Examples of the means for inputting are disclosed with respect to FIG. 8.

[0052] As used herein, the terms "computing system" or "computer system" refer to any electronic computing device that can receive numerical inputs and provide numerical outputs by any kind of interface, digital and / or analog, whether integrated or distributed. Typically, a computing system designates either a computer that executes software having access to data storage, or a client-server architecture where the client side functions as an interface while data and / or computations are executed on the server side. Examples of such computing systems are disclosed with respect to FIG. 8.

[0053] As used herein, the term "digital representation identifier" refers to any injective digital representation of a physical item such as a composition or a flavor or fragrance ingredient. Such a digital representation identifier may correspond, for example, to an entry in a database. The digital representation identifier may refer to, for example, a name, chemical structure, or label representing an internal reference of a flavor or fragrance ingredient.

[0054] As used herein, the term "emotional or sensory perception" refers to any measurable emotional or sensory manifestation of a perception induced by exposing a human to a composition physical composition digital identifier. Such measurements may be observed, for example, by neuroimaging techniques. Such perceptions may be related to emotions felt by the human (e.g., happiness, etc.), or to the perception of a sensation by the human (e.g., cold, etc.).

[0055] Such an emotion or sensation may belong to one of the following categories: - Sensuality (sensual, romantic, desire), - Energy (activated, energetic, refreshed), - Happiness (pleasantly surprised, happy, blissful), - Indulgence (tasty, thirsty, hungry), - Unpleasant (annoyed, irritated, surprised by unpleasantness), - Nostalgic (fond, melancholy, sad), - Interesting (fun, impressive, interesting), - Calmness (soothed, comforted, relaxed), and - Spirituality.

[0056] Such feelings or sensations (or attributes or impressions, which may also be called) may also belong to one of the following categories: - Aggressive, - Fun, - Indifferent, - Artificial, - Childish, - Caring, - Inexpensive, - Citrus, - Classical, - Clean, - Comforted, - Creamy, - Desire, - Unpleasant, - Easy to wear, - Edible, - Elegant / refined, - Energetic, - Feminine, - Flower-like, - Fresh, - Fruit-like, - Gourmet, - Green, - Happy, - Characteristic, - Heavy, - Impressive, - Interesting, - Irritated, - Light, - Preference, - Long-lasting, - Makes one feel good, - Masculine, - Medicinal, - Melancholic, - Latest, - Moisturizing, - Delicious-looking, - Natural, - Neutral, - New, - Nostalgic, - Old-fashioned, - Overwhelming, - Pleasantly surprised, - Pure, - Refreshed, - Relaxed, - Activated, - Romantic, - Sad, - Sensual, - Soapy, - Soft, - Healed, - Sour, - Spicy, - Sporty, - Strong, - Sweet, - Perfect for you, - Unpleasantly surprised, - Happiness, and / or - Woody.

[0057] In this specification, the terms "embodied" or "assembled" are intended to exist outside the digital environment of the present invention. "Embodied" can mean, for example, being readily found in nature or being synthesized in a laboratory or chemical plant. In any case, the embodied physical composition digital identifier presents a tangible reality. The terms "formulated" or "formulation" refer to the act of embodying a physical composition digital identifier, whether by extraction and assembly of components or by synthesis and assembly of components.

[0058] The embodiments disclosed below are presented in a general way.

[0059] As an example, the present invention can be used to predict the emotional or sensory response of a human user presented with a digital identifier of a physical composition intended to provide a "romantic" emotion or sensation.

[0060] Furthermore, the present invention can be used to predict the overall like / dislike of a physical composition digital identifier, or more specifically, to predict fragrance like / dislike and / or flavor like / dislike.

[0061] As a sample input for the following examples, data from 3500 tests can be used, each test including eight compositions tested with 80 - 100 panelists participating in each test. Such data can exhibit different human biases: - Country bias - In Mexico, panelist votes of 80 - 90% are normal, while in Germany, 50 - 60% is already a very high number. - Application bias - The same fragrance composition tested in two different applications can have different results. - Test bias - Compositions tested together can affect each other's results. A very sweet composition tested with other compositions can lower the other sweetness scores because all these comparisons are relative.

[0062] For example, the results of two different tests ranking sweetness among three compositions can be as follows: - Test 1 - Composition 1 11.4 - Composition 2 26.3 - Composition 3 19.7 - Test 2 - Composition 1 21.4 - Composition 2 37.3 - Composition 3 25.7

[0063] In this example, a very general situation can be seen where the values of both tests are different while the relative relationship (i.e., ranking) remains the same, i.e., one composition is sweeter than another composition in both tests and the scores of different tests can be different. Thus, the scoring values are only valid within one specific test, and rather than training the model object of the present invention to predict these scores (such as a regression model), the model is trained to correctly rank the compositions (thus, it is a ranking model). The advantage of the ranking model is to take each test separately and adapt it to rank the compositions that match their parameters to the ground truth ranking. As will be understood, such measures enable the prediction of the pairwise ranking of the components forming the physical composition digital identifier.

[0064] FIG. 1 shows a particular embodiment of the system 100 of the present invention. This system 100 for training a gradient boosting decision tree device or a neural network device that associates a perceived emotion or sensation with a physical composition digital identifier of the fragrance or aroma of at least one fragrance or aroma component comprises one or more computer processors and one or more computer-readable media operably coupled to the one or more computer processors, and when executed by the one or more computer processors, the one or more computer-readable media cause the one or more computer processors to perform at least the following steps, namely: - providing, at 105, a set of exemplary data to the gradient boosting decision tree device, the set of data comprising - at least two physical composition digital identifiers of at least two fragrances or aromas formed by at least two physical component digital representation identifiers of the fragrances or aromas, and - At least one emotional or sensory perception value 102 for at least two of the said aromatic or flavor components, where the emotional or sensory perception value is associated with at least one emotional or sensory perception digital identifier, and the emotional or sensory perception digital identifier represents a category of human emotional or sensory responses to the embodied physical composition digital identifier from among a finite list of emotional or sensory responses, at least one emotional or sensory perception value 102 and including, providing 105 and - Based on an exemplary data set, operating a gradient boosting decision tree device 110 and - Obtaining a trained gradient boosting decision tree device 115 and including.

[0065] System 100 itself may be formed from any combination of means for performing characteristic steps executed by a computer processor.

[0066] The step of providing 105 may be executed via a computer interface such as an API or any other digital input system. This step of providing 105 may be started manually or automatically. The exemplary data set may be assembled manually on a computer interface or automatically by a computing system from a larger set of exemplary data.

[0067] Exemplary data is, for example, - At least two fragrance or aroma physical composition digital identifiers 101 formed by at least two fragrance or aroma physical component digital representation identifiers and - At least one emotional or sensory perception value 102 for at least two of the aromatic or flavor components, wherein the emotional or sensory perception value is associated with at least one emotional or sensory perception digital identifier, and the emotional or sensory perception digital identifier represents a category of human emotional or sensory responses to the embodied physical composition digital identifier from among a finite list of emotional or sensory responses, at least one emotional or sensory perception value 102 comprising.

[0068] For example, the digital representation identifier of the physical component of the fragrance or aroma may correspond to a reference to the name of the component or an entry in a database of characteristics associated with the component. Such an identifier may be, for example, a Universal Unique Identifier (“UUID”).

[0069] The emotional or sensory perception value 102 may, for example, correspond to a numerical representation of the perception among panelists in a survey of the effect of the component on a determined emotion or sensory perception. Such a value 102 may be, for example, the difference between the components, where the difference refers to the quantification of the difference between the components with respect to the effect of the component on a determined emotion or sensory perception. In other embodiments, this emotional or sensory perception value 102 is an absolute quantification where the component is evaluated alone and not in relation to another component.

[0070] In more advanced embodiments, the exemplary set of data may also include - The gender 103 of the human exposed to the embodied physical composition, The country of origin 104 of the human exposed to the embodied physical composition, The type of use 106 of the embodied physical composition, The composition chemical base 107 used to support the embodied physical composition, and / or The dosage 108 of at least one physical component fragrance or aroma physical component represented by the corresponding digital representation identifier may be included.

[0071] Such a set of exemplary data may be obtained by assembling a physical composition digital identifier, presenting the physical composition digital identifier to a human user, and recording the emotional or sensory responses of those users. These records may be classified, for example, by type of emotional or sensory response ("relaxed", "romantic", "sad", "sensual", "thirsty", "healed"). This classification may be performed manually by associating tags with each emotional or sensory response, or may be performed automatically by a computing system configured to process text, audio, photo, or video recordings to determine the type of response from a list of possible types of responses. These types of emotional or sensory responses may then be associated with types of emotional or sensory response digital identifiers that can be used in the set of exemplary data.

[0072] Such emotions or sensory responses, or emotional or sensory perceptions, may also correspond to the perception of health or hygiene benefits associated with exposure to the physical composition.

[0073] In certain embodiments, the exemplary set may include values representing at least one of the following: - An identifier of the chemical composition of the fragrance, - A test stamp defining an identifier of the test, - The number of respondents indicating the number of consumers who accessed the test to evaluate the fragrance, - The fragrance dosage measuring the concentration (%) of the composition inside the tested product, - The country indicating the country where the test was conducted, - The gender indicating the type of consumer gender: both, male, female, - The age group attribute indicating the age group of the consumer: both, young, old, - The segment name defining the use of the product at a high level: body personal care, home fabric care, fine fragrance, - The sub-segment name defining the use of the product in a more specific way than the segment name: 13 attributes, - An application name that defines the use of the product in a more specific way than the sub-segment name. It corresponds to the leaf of the tree structure defined by the segment name, sub-segment name, and application name: 158 attributes, - A stage name tester that defines at which stage the product was tested, including 123 attributes, and this feature includes attributes such as neat squeeze bottles, wet clothes, during washing, etc., - The percentage score of the perceived emotion or feeling towards the following emotions, sensations, attributes, or preferences, - Emotions: interesting, comforted, desire, unpleasant, energetic, hungry, happy, impressive, interesting, irritated, melancholic, appetizing, nostalgic, pleasantly surprised, refreshed, relaxed, activated, romantic, sad, healed, unpleasantly surprised, happy, - Attributes: aggressive, cold, artificial, childish, caring, cheap, citrus, classical, clean, creamy, easy to wear, edible, elegant / refined, feminine, flowery, fresh, fruity, gourmet, green, characteristic, heavy, light, long-lasting, mood-enhancing, masculine, latest, moisturizing, natural, neutral, new, old-fashioned, overwhelming, pure, sensual, soapy, soft, sour, spicy, strong, sweet, perfect for you, unique, woody, and / or - Preference score.

[0074] In a specific variant form, the purpose of the system 100 of the present invention includes a step (not shown) of preprocessing the input. For example, this preprocessing step may utilize a sparse matrix called a physically compositional digital identifier matrix that stores the weighted partitions of the components used, and the weighted sum of each composition is 1. The physically compositional digital identifier matrix can include approximately one million compositions (rows) and 6351 components (columns), where the weight corresponds to the amount of a given component (column) within a given composition (row). It is also possible to use the physically compositional digital identifier matrix without modification and replace each composition identifier with its corresponding component physically compositional digital identifier.

[0075] Furthermore, in order to select an aroma with a dosage strictly higher than zero, binarize the encapsulation, use all ages, and remove the application name, a grid search algorithm can be additionally applied. This selects the best set of parameters for each emotion or sensation and thus maximizes the results for each emotion or sensation.

[0076] Furthermore, for each emotion or sensation, in order to maximize the information gain brought about by the use of target encoding, the most important features can be calculated from gender, age group attributes, segment name, sub-segment name, and aroma usage amount. If including the feature improves the quality metric, the feature is considered the most important, indicating that it is a prominent feature.

[0077] The step 110 of operating may be performed, for example, by a computer program executed on a computing system. During this step 110, the gradient boosting decision tree device is configured to determine the operating values of the parameters of the decision-making within the tree architecture, and the values are used with respect to future physical composition digital identifiers, providing specific knowledge regarding these physical composition digital identifiers.

[0078] Optionally, during the step 110 of operating, gradient boosting trees are sequentially constructed, such that each new tree minimizes the ranking loss in the context of a neural network device optimized for ranking prediction.

[0079] In other embodiments, any other neural network device technology can be used during the step 110 of operating. Such a neural network device may implement, for example, the RankNet, LambdaRank, or SoftRank algorithm.

[0080] In such an embodiment, the system 100 of the present invention is subject to the following steps, namely: Step 105 of providing a set of exemplary data to a neural network device, wherein the set of data comprises at least two digital identifiers of fragrance or aroma physical compositions formed by at least two digital identifiers of physical components of fragrance or aroma, and at least one emotional or sensory perception value for at least two of the fragrance or aroma components, wherein the emotional or sensory perception value is associated with at least one digital identifier of emotional or sensory perception, and the digital identifier of emotional or sensory perception represents a category of human emotional or sensory response to the digital identifier of the embodied physical composition selected from a finite list of emotional or sensory responses, at least one emotional or sensory perception value and Step 105; Step 110 of operating the neural network device based on the set of exemplary data; and Step 115 of obtaining a trained neural network model are included.

[0081] In previous systems, during each test, the product is successively tested by the same group of consumers. This introduces bias into the results. When testing a product, the sensations perceived by the consumer strongly depend on the products previously tested. Consumers make unconscious comparisons and thus give results that are influenced by these comparisons. As a result, the ranking algorithm based on test stamp attributes is an innovative idea for this problem and reproduces the actual behavior as closely as possible. Furthermore, the test stamp attributes define identifiers within the dataset that enable the use of the dataset without aggregation. Separating the groups by test stamp characteristics simultaneously forces a comparison between the compositions evaluated by the consumers. Considering the technical coding implementation, the training-validation-test split can be done by first applying a random split according to the test stamp (20% test set) and then applying a group k-fold split again according to the test stamp (20% validation set). The training of the decision tree neural network device may use CatBoost with the YetiRank and QuerySoftmax algorithms. YetiRank and QuerySoftmax belong to a family of pairwise learning methods that optimize an objective function defined on pairs of data samples retrieved for a given group. The important concept of high values introduced here is the accuracy measure related to the gap value. The gap value is the difference between the ground truth values of two compositions tested together. The accuracy measurement using the gap value is a new metric that can be used in combination with the proportion of correctly predicted pairs to measure the accuracy of the prediction.

[0082] CatBoost has been shown to function well for handling categorical variables. Hyperparameter tuning, done by maximizing the average gap score over the maximum gap calculated for each emotion or sensation, significantly improves the training output.

[0083] In certain embodiments, the system 100 of the present invention includes a post-processing step (not shown) of constructing a distribution of predicted values to determine the position of a particular composition in a particular context. The component impact value correlates with the amount for selecting the most feasible component. Other filters regarding component criteria may be applied.

[0084] The performance of the model can be measured according to weighted and properly ranked pairs. The ranking method measured by the ratio of weighted and properly ranked pairs is defined as follows: The performance of the model can be measured according to weighted and properly ranked pairs. The ranking method measured by the ratio of weighted and properly ranked pairs is defined as follows:

Number

[0085] The reason for using this score is that the percentage of properly ranked pairs is weighted by the number of pairs included in each test and passing the condition |ya - yb| > gap. This condition is to select only pairs where the absolute value difference of the sentiment or sensation scores is greater than the numerical gap value. Therefore, the typical pattern is an increase in the score when increasing the gap value.

[0086] Furthermore, in order to have a single numerical value for performing automatic parameter tuning, the ranking scores for gaps in the range of 0 to α*σ (sentiment or sensation distribution) can be averaged.

[0087] The output of the ranking model is a ranking score, which is a technical value representing a measurable difference in the perceived sentiment or sensory perception between physical composition digital identifiers. The higher the ranking value, the higher the ranking position is indicated. To associate these abstract ranking values with meaningful outputs, a quantile transformer can be used to indicate where the composition is ranked in a particular context. For example, the composition has a new quantile of 0.68 - this means that this composition is fresher than 68% of the target country and sub-segment compositions.

[0088] To determine which set of components is most influential for each sentiment or sensation, it is important to remember that the sentiment or sensation scores are affected by country biases. Therefore, component recommendations are provided for each country.

[0089] Recognized as a state-of-the-art method in the explainability of machine learning, Shappley Additive Explanations (SHAP) was published by Lundberg and Lee in 2017. SHAP assigns importance values to each feature of the input data set for a particular prediction.

[0090] Considering that components having a positive or negative relationship between the SHAP importance score and the component dosage need to be processed separately, selecting the most influential components is not a trivial task. The linear relationship determined using Spearman correlation enabled the separation of positive and negative relationships. The components that have the best effect on each emotion or sensation may be associated with defined boundary dosages given by the minimum and / or maximum amounts of a specific component. The minimum amount of a specific component can indicate the starting amount at which the component becomes an enhancer of a specific emotion or sensation. Beyond the minimum amount, the perception of a given emotion or sensation should increase until the maximum amount is reached. Adding an amount more than that indicated by the maximum amount, defined as the breakpoint, results in a slow increase or saturation or decrease in the influence of the component on a specific emotion or sensation. The components that have the most adverse effects also provide important additional information to flavorists. Adding a large amount of negative components may dramatically reduce the emotional or sensory attributes of a given emotion or sensation.

[0091] Such positive or negative relationships between component concentration and the perceived intensity of the perceived emotion or sensation are shown in FIG. 5, in which: - Two different components 601 and 602, footprints are shown, - The first component 601 is associated with a positive relationship to the perceived intensity of a specific emotion or sensation, the relationship being defined by: - A minimum concentration boundary 605 where no effect on the determined emotion or sensation is felt, - A maximum concentration boundary 610 where no additional effect on the determined emotion or sensation is felt, - An optimal range between the minimum concentration boundary 605 and the maximum concentration boundary 610 where additional component concentration affects the perceived emotion or sensation proportionally, - The second component 602 is associated with a negative relationship to the perceived intensity of a specific emotion or sensation, the relationship being defined by: - A maximum concentration boundary 615 where additional component concentration proportionally and negatively affects the perceived emotion or sensation, - A range below the maximum concentration boundary 615, within which no additional influence on the determined emotion or feeling is felt regardless of the concentration of the second component 602.

[0092] Considering either positive or negative relationships, calculate the final importance score for each component by taking the median of the positive or negative SHAP scores for the components for positive or negative samples exceeding 50. In this way, the component that contributes most positively or negatively is the one with the highest or lowest median score.

[0093] Intending to determine the quantity range and boundaries, for each feasible component, the plot of data points against quantity appearing in SHAP can be approximated by a piecewise linear function.

[0094] An example of composition optimization: For a shampoo composition having a ranking quantile = 0.265 according to the happiness model, targeting the US market to improve the happiness attribute (this composition brings more than 26% happiness of US hair care compositions and less than 74% happiness of US hair care compositions, meaning there is room for improvement). Shap values can be used to identify the top positive or negative component contribution factors in the component space. Since two of the top positive contribution factor components are already above the saturation quantity limit, these components may be maintained as they are, but the quantity of component 3 can be increased (the current quantity is 0.2%, but the recommended saturation limit or upper limit of this composition is 0.4%). Two negative components have quantities higher than the maximum proposed limit, so these components may be reduced to reduce their negative impact. The following table shows these effects individually and in combination:

Table 1

[0095] By manipulating these three components, it is possible to improve the preference quantile score from 0.26 to 0.45. Therefore, this composition is preferred more than 45% of US hair care compositions.

[0096] The additional output of the ranking model is a so-called "DNA" that models customer behavior in various countries. For example, the attribute "clean" differs in terms of the perception of a particular country regarding the olfactory profile. To search for these translations from emotional or sensory attributes to olfactory descriptors, the correlation between the predicted global ranking of the compositions of a particular country and the olfactory profile of the compositions can be calculated. The top positively and negatively correlated olfactory descriptors indicate what humans associate with a particular attribute and emotion or sensation. For some attributes, all countries are consistent - for example, the sweet attribute correlates with chocolate, caramel, and vanilla for almost all countries, while for other attributes, significant differences are observed. For example, for the attribute "natural" - among the most correlated olfactory descriptors in China are orange, grapefruit, lemon, citrus, and green, while in the case of the United States, there are fresh, calla lily, German calla lily, and green, and grapefruit & citrus are among the top negative correlations for natural.

[0097] An example of the classification output of the model can be shown in the following example: Based on the aggregation of different ranking results of related categories (emotion or sensation, attribute), the present invention can classify the compositions into groups related to the required context (sub-segment + country). For example, Composition 1 is in the top class of US body washes because it has high predicted quantiles for preference (0.74), clean (0.69), natural (0.81), and not aggressive (0.23).

[0098] For this particular use, the model input is a dictionary-form composition physical composition digital identifier {component1: quantity1, component2: quantity2,...} and a dictionary of additional information {country: United States, segment: personal care, sub-segment: hair care, gender: female}. This set of dictionaries that describe the compositions tested by the inventors is converted into a matrix form to feed the model. The target values given to the model are the results of blinded tests, and they can be in different forms depending on the case: for training rankings, "the fitness value within each test", that is, the percentage of participants who voted for a particular attribute and emotion or sensation, or the average value of the scores given by participants for a particular composition and a particular attribute may be used. For regression target values, they are the scores given by each user for classification, and in particular the predicted ranking scores for important attributes and emotions or sensations and hobbies.

[0099] For this particular use, the model output is the ranking quantiles for each attribute / emotion or sensation / preference in the ranking method, the average score of all participants in the test in the regression method, and the fitness class of the composition for a particular country and sub-segment in the classification method.

[0100] The obtaining step 115 may be performed via a computer interface such as an API or any other digital output system. The obtained trained gradient boosting decision tree device or neural network device may be stored in a data storage such as a hard drive or a database.

[0101] Figure 2 schematically shows a particular series of steps for the method 200 of the present invention. This method 200 for training a gradient boosting decision tree device or a neural network device that associates a perceived emotion or sensation with a physical composition digital identifier of a fragrance or aroma containing at least one fragrance or aroma component includes at least the following steps, namely, Step 105 of providing a set of exemplary data to a gradient boosting decision tree device or to a neural network device, the set of data comprising: at least two digital identifiers of fragrance or aroma physical compositions formed by at least two digital representations of physical components of the fragrance or aroma; and at least one emotional or sensory perception value for at least two of the fragrance or aroma components, the emotional or sensory perception value being associated with at least one digital identifier of emotional or sensory perception, the digital identifier of emotional or sensory perception representing a category of human emotional or sensory reactions to the digital identifier of the embodied physical composition selected from a finite list of emotional or sensory reactions; Step 105; Step 110 of operating a gradient boosting decision tree device or a neural network device based on the set of exemplary data; Step 115 of obtaining a trained gradient boosting decision tree model or a trained neural network model; and

[0102] As will be appreciated, the present invention also aims at a computer-implemented gradient boosting decision tree device or neural network device 305, characterized in that a neural network device is obtained by a computer-implemented method 200 as disclosed with respect to Figure 2.

[0103] As will be appreciated, the present invention also aims at a computer program product comprising instructions for performing the steps of a method as disclosed with respect to Figure 2 when executed on a computer.

[0104] As will be appreciated, the present invention also aims at a computer-readable medium storing instructions for performing the steps of a method 200 as disclosed with respect to Figure 2 when executed on a computer.

[0105] Figure 3 schematically shows a particular series of steps for the method 400 of the present invention. This method 400 for determining the emotional or sensory perception related to the exposure to flavor or fragrance components is Step 405 of inputting at least two digital representation identifiers of the physical components of flavor or fragrance on a computer interface, wherein the obtained input corresponds to a physical composition digital identifier representing the physical composition of the physical flavor or fragrance component, step 405; Step 410 of operating a computing device to associate at least one value with the input physical composition digital identifier, the value being Among the input digital representation identifiers of the physical components of the flavor or fragrance, the relative ranking for at least one determined human emotion or sensory perception, For at least one of the input digital representation identifiers of the physical components of the flavor or fragrance, a value representing the perception for at least one determined human emotion or sensory perception, and / or A value representing the class of the flavor or fragrance component in the classification of the flavor or fragrance component by the perception of human emotion or sensory perception for at least one determined human emotion or sensory reaction, Representing, step 410; Step 415 of providing on a computer interface at least one value obtained during the step of operating for the input physical composition digital identifier and Including.

[0106] The input step 405 is performed manually or automatically on a computer interface such as, for example, a graphical user interface (「GUI」). During this input step 405, digital representation identifiers of the physical components of the fragrance and / or aroma may be selected to form digital identifiers of the physical composition representing the equivalent real fragrance. In advanced embodiments, the relative or absolute amounts of the digital representation identifiers of the physical components of the fragrance and / or aroma may be input. At the beginning of this input step 405, a digital identifier of the physical composition is defined.

[0107] The operate step 410 is performed, for example, by a computing device. During such an operate step 410, the algorithms of the expert system may be executed, and a series of setting instructions are executed to provide values representing the following based on the input composition: - The relative ranking between the digital representation identifiers of the physical components of the input fragrance or aroma with respect to at least one determined human emotion or sensory perception, - For at least one digital representation identifier of the physical components of the input fragrance or aroma, a value representing the perception for at least one determined human emotion or sensory perception, and / or - A value representing the class of the fragrance or aroma components in the classification of the fragrance or aroma components by the perception of human emotions or sensory perceptions with respect to at least one determined human emotion or sensory reaction.

[0108] In other embodiments, during this operate step 410, a trained gradient boosting tree model or a trained neural network model may be operated to determine such values.

[0109] For example, during this operate step 410, the input digital identifier of the physical composition is supplied to a gradient boosting decision tree device, and the gradient boosting decision tree device returns at least one of the following: - The relative ranking between the input fragrance or aroma physical component digital representation identifiers for at least one determined emotion or sensory perception obtained from a trained gradient boosting decision tree device - For at least one input fragrance or aroma physical component digital representation identifier, a value representing the perception for at least one determined emotion or sensory perception obtained from a trained gradient boosting decision tree device, and / or - A value representing the class of the fragrance or aroma component in the classification of the fragrance or aroma component by the perception of emotion or sensory perception for at least one determined emotion or sensory perception.

[0110] The providing step 415 is performed, for example, by displaying on a computer screen the predicted human reaction output by a gradient boosting decision tree device.

[0111] In certain embodiments, such as those shown in FIG. 3, the method 400 of the present invention, downstream of step 410, includes a step 420 of determining, by a computing device, a numerical value representing the effect of an emotional or sensory reaction to at least one input fragrance or aroma component, and a step 415 of providing the numerical value of the determined reaction effect on a computer interface.

[0112] The effect of the reaction is a numerical representation of the ranking position determined for all other formulations with respect to a particular input (i.e., quantile value). For example, if there are no two formulations to compare and predict which is fresher, it can be compared to a context (e.g., all shampoos for women in the United States). For example, if a formulation has a quantile of 0.56 for freshness, this means that this formulation is fresher than 56% of the shampoo formulations for women in the United States and not fresher than 44% of the formulations for women in the United States.

[0113] In certain embodiments, such as those shown in FIG. 3, the method 400 of the present invention includes, downstream of step 420 of determining a numerical value representative of the effect of an emotion or sensory response, step 435 of providing at least one alternative digital representation identifier of a physical component of a fragrance or aroma, the physical composition digital identifier being provided to form as a function of the numerical values representative of the effect of the input and alternative fragrance or aroma components on an emotion or sensory response.

[0114] The alternative digital representation identifier of a physical component of a fragrance or aroma is selected, for example, as a function of the effect of the fragrance or aroma component on the determined emotion or sensation.

[0115] In certain embodiments, step 435 of providing is configured to provide a value of the amount or concentration of the alternative fragrance or aroma component.

[0116] During step 435 of providing, if the alternative digital representation identifier of a physical component of a fragrance or aroma is associated with a similar emotion or sensory perception value or effect, the alternative digital representation identifier of a physical component of a fragrance or aroma may be selected. Such a similar emotion or sensory perception value or effect may be determined if the emotion or sensory perception value or effect of the replacement component (or replacement candidate) is within an acceptable range from the emotion or sensory perception value or effect of the original component.

[0117] In a simple variant, the emotion or sensory perception value or effect of a component is evaluated with respect to only one emotion or sensory perception value.

[0118] In a more complex variant, several emotion or sensory perception values or effects of a component are evaluated to provide a more complete match between the emotion or sensory perception associated with exposure to the replaced component and the original component.

[0119] During the providing step 435, a single original component may be replaced by a plurality of candidate components, and similarly, a plurality of original components may be replaced by a single candidate component.

[0120] In certain variations, instead of or in addition to providing an alternative fragrance or aroma physical component digital representation identifier, the providing step 435 is configured to provide an alternative amount or concentration for at least one input fragrance or aroma physical component digital representation identifier of the physical composition digital identifier.

[0121] In certain embodiments, such as those shown in FIG. 3, the method 400 of the present invention includes a step 430 of assembling a composition represented by an input physical composition digital identifier.

[0122] Such an assembling step 430 may be performed in various ways, such as in a laboratory or a chemical plant.

[0123] In certain embodiments, the method 400 of the present invention includes a step 401 of assembling a database that associates values representing measured human emotions or sensory perceptions for at least one group of at least two fragrance or aroma physical component digital representation identifiers, and a step 407 of determining, by a computing device, a set of instructions to be executed during the operating step 410. and

[0124] The assembling step 401 can be performed manually, automatically, or semi-automatically using any input means. During this assembling step 401, a group of at least two fragrance or aroma physical component digital representation identifiers (corresponding to a composition or a part of a composition) are associated with values representing measured human emotions or sensory perceptions in a computer memory.

[0125] The nature of the determining step 407 depends on the type of instructions that are operative during the operating step 410. For example, this determining step 407 may correspond to method 200 as disclosed with respect to FIG. 2. In other embodiments, this determining step 407 may correspond to a regression algorithm used in an assembled database.

[0126] In certain embodiments, the step 401 of assembling a database includes exposing at least one human to a physical composition of physical components of a fragrance or aroma, measuring the emotions or sensory perceptions of the at least one human exposed to the physical composition, and recording in a database a value representing the measured human emotions or sensory perceptions, in association with a group of digital representation identifiers of physical components of a fragrance or aroma representing the physical components of the composition used during the exposing step. Including.

[0127] The measuring step 403 can be performed by any type of sensor commonly used to measure the intensity or presence of emotions or sensory perceptions in humans.

[0128] The recording step 404 can be performed by any input means adapted to a particular use case.

[0129] In certain embodiments, the method 400 of the present invention includes a step 406 of replacing an input digital representation identifier of physical components of a fragrance or aroma with a different equivalent digital representation identifier of physical components of a fragrance or aroma, the equivalence being defined in a database of equivalent digital representation identifiers of physical components of a fragrance or aroma.

[0130] Such a replacement step 406 may be performed by creating a dictionary of synonyms between digital representations of physical components of a fragrance or aroma and by searching for synonyms of an input digital representation of a physical component of a fragrance or aroma.

[0131] In certain embodiments, the step 410 of operating includes a first step 411 of associating, by a computing device, at least one olfactory or gustatory descriptor with at least one input digital representation of a physical component of a fragrance or aroma, and a second step 412 of associating, by a computing device, at least one human emotion or sensory perception as a function of at least one olfactory or gustatory descriptor associated with at least one input digital representation of a physical component of a fragrance or aroma. Including.

[0132] Emotions and attributes can be translated into a series of olfactory descriptors or tonality. It is possible to understand preferences in certain contexts represented by concepts other than the actually measured attributes (for example, in China, nature is associated with citrus descriptors (lemon, grapefruit), while in European countries, nature is somewhat milder like almond).

[0133] Such associations can be obtained by exposing the user to the formulation and asking which descriptor best fits this formula and which emotion or sensory perception best describes the perceived emotion or sensation by the user. From this sampling, general association rules can be determined statistically.

[0134] FIG. 5 schematically shows a particular embodiment of the system 500 of the present invention. This fragrance or aroma prediction emotion or sensory perception determination system 500 is Step 505 of entering at least two digital representation identifiers of physical components of a fragrance or aroma on a computer interface (510), wherein the obtained input corresponds to a physical composition digital identifier representing a physical composition of the physical fragrance or aroma component, step 505, Step 515 of operating a computing device (520) with instructions configured to associate at least one value with the input physical composition digital identifier, the value being Among the input digital representation identifiers of the physical components of the fragrance or aroma, at least one relative ranking with respect to at least one determined human emotion or sensory perception, For at least one of the input digital representation identifiers of the physical components of the fragrance or aroma, a value representing a perception with respect to at least one determined human emotion or sensory perception, and / or A value representing a class of the fragrance or aroma component in the classification of the fragrance or aroma component by the perception of human emotion or sensory perception with respect to at least one determined human emotion or sensory reaction, Representing, step 515, Step 525 of providing on the computer interface at least one value obtained during the step of operating with respect to the input physical composition digital identifier and Including.

[0135] Specific embodiments of the means, interface, and computing device are disclosed with respect to FIG. 3.

[0136] FIG. 6 schematically shows a particular series of steps for method 700 of the present invention. Such a particular series of steps uses exemplary data, and the exemplary data is, for example, - At least one digital representation identifier of at least one physical component of a fragrance or aroma, wherein the at least one digital representation identifier of the at least one physical component of the fragrance or aroma forms a physical composition digital identifier, and the physical composition digital identifier is optionally associated with the physical component digital representation identifier, - human reactions detected with respect to at least one physical composition digital identifier of an exemplary set and may be included.

[0137] In a more advanced embodiment, the exemplary set of data may also - at least one solvent digital identifier associated with the physical composition digital identifier, - at least one dosage digital identifier associated with the physical composition digital identifier, - at least one pH of the composition digital identifier associated with the physical composition digital identifier, - at least one citral concentration digital identifier associated with the physical composition digital identifier, - at least one expiration date of the product digital identifier associated with the physical composition digital identifier, and / or - at least one complete sensory data of the composition digital identifier associated with the physical composition digital identifier may be included.

[0138] Such an exemplary set of data may be obtained by assembling the physical composition digital identifiers of the components, presenting the physical composition digital identifiers to human users, and recording the reactions of those users in a database. These records may be classified, for example, by the type of reaction ("like", "dislike", "smile", "leave"). This classification may be performed manually by associating tags with each reaction, or may be performed automatically by a computing system configured to process text, audio, photographic, or video recordings to determine the type of reaction from a list of possible types of reactions. These types of reactions may then be associated with the types of reaction digital identifiers that can be used in the exemplary set of data.

[0139] Next, this exemplary data may be input 705 into a gradient boosting tree device, and then the device may be operative 710 to train a trained gradient boosting tree model 305 obtained as an output 715 of step 710 of training the gradient boosting tree model.

[0140] Such examples have as an object to predict a value representing a reaction in relation to a particular association of components forming a physical composition digital identifier, while the examples of FIGS. 1 - 5 predict a relative value of a reaction in relation to a particular association of components forming a physical composition digital identifier (i.e., a particular physical composition digital identifier triggers more particular emotions or sensations in a user than another physical composition digital identifier).

[0141] In certain embodiments, the method subject of the present invention includes, downstream of the step of operating a trained ensemble of gradient boosting tree devices, determining, by a computing device, a numerical value representing the effect of at least one input fragrance or aroma component on a reaction, and providing, on a computer interface, the determined numerical value of the effect of the reaction.

[0142] Such an embodiment is shown in relation to FIG. 7 schematically showing an embodiment of the method 800 of the present invention. The effect of this fragrance or aroma prediction on the human reaction prediction system 800 is - at least two inputs 805 including at least two physical component digital representation identifiers of a fragrance or aroma, wherein the resulting inputs 805 correspond to physical composition digital identifiers of the fragrance or aroma components, and each input 805 includes a different set of physical component digital representation identifiers of the fragrance or aroma, at least two inputs 805 -At least one ensemble of gradient boosting tree devices 810 trained according to the method shown in FIG. 6, wherein the physical composition digital identifier of the input 805 is input into a single ensemble of the gradient boosting tree devices 810, and the ensemble of the gradient boosting tree devices 810 is configured to provide a prediction for each physical composition digital identifier of the input 805, at least one ensemble; -The predicted output 815 of the influence of each flavor or aroma physical component digital representation identifier as a function of the prediction of the decision tree neural network device 810 and.

[0143] Such predicted output 815 is obtained by relative comparison of the human reaction predicted output for each physical composition digital identifier, and an influence score may be associated with each component based on the components of the physical composition digital identifier.

[0144] FIG. 8 represents a block diagram showing an exemplary computer system 900 that may implement an embodiment of the present invention. In the example of FIG. 8, the computer system 905 and the instructions for implementing the disclosed technology in hardware, software, or a combination of hardware and software are generally used by those skilled in the art to which this disclosure pertains for communicating about computer architectures and embodiments of computer systems. It is schematically represented, for example, as boxes and circles, at the same level of detail.

[0145] The computer system 905 includes an input / output (IO) subsystem 920 that may include a bus and / or other communication mechanism for communicating information and / or instructions between components of the computer system 905 via an electronic signal path. The I / O subsystem 920 may include an I / O controller, a memory controller, and at least one I / O port. The electronic signal path is schematically represented in the drawings, for example, as lines, one-way arrows, or two-way arrows.

[0146] At least one hardware processor 910 is coupled to an I / O subsystem 920 for processing information and instructions. The hardware processor 910 may include, for example, a general-purpose microprocessor or microcontroller, and / or a dedicated microprocessor such as an embedded system or a graphics processing unit (GPU) or a digital signal processor or an ARM processor. The processor 910 may include an integrated arithmetic logic unit (ALU) or may be coupled to a separate ALU.

[0147] The computer system 905 includes one or more units of memory 925, such as main memory, coupled to the I / O subsystem 920 for electronically and digitally storing data and instructions to be executed by the processor 910. The memory 925 may include volatile memory such as various forms of random access memory (RAM) or other dynamic storage devices. The memory 925 may also be used to store temporary variables or other intermediate information during the execution of instructions to be executed by the processor 910. When such instructions are stored on a non-transitory computer-readable storage medium accessible to the processor 910, the computer system 905 can be made into a dedicated machine customized to execute the operations specified by the instructions.

[0148] The computer system 905 further includes non-volatile memory, such as a read-only memory (ROM) 930 or other static storage device, coupled to the I / O subsystem 920 to store information and instructions for the processor 910. The ROM 930 may include various forms of programmable ROM (PROM), such as erasable PROM (EPROM) or electrically erasable PROM (EEPROM). The unit of persistent storage 915 may include various forms of non-volatile RAM (NVRAM), such as FLASH memory, or solid-state storage, magnetic disk, or optical disk such as CD-ROM or DVD-ROM, and may be coupled to the I / O subsystem 920 to store information and instructions. The storage 915 is an example of a non-transitory computer-readable medium that can be used to store instructions and data that, when executed by the processor 910, execute a computer-implemented method to carry out the techniques herein.

[0149] Instructions in the memory 925, ROM 930, or storage 915 may include one or more sets of instructions organized as a module, method, object, function, routine, or call. The instructions may be organized as an application program including one or more computer programs, operating system services, or mobile apps. The instructions may include an operating system and / or system software; one or more libraries supporting multimedia, programming, or other functions; data protocol instructions or stacks for implementing TCP / IP, HTTP, or other communication protocols; file format processing instructions for parsing or rendering files encoded using HTML, XML, JPEG, MPEG, or PNG; user interface instructions for rendering or interpreting commands for a graphical user interface (GUI), command line interface, or text user interface; application software such as an office suite, internet access application, design and manufacturing application, graphic application, audio application, software engineering application, educational application, game, or other application. The instructions may implement a web server, web application server, or web client. The instructions may be organized as a presentation layer, application layer, and data storage layer such as a relational database system using Structured Query Language (SQL) or without SQL, object store, graph database, flat file system, or other data storage.

[0150] Computer system 905 may be coupled to at least one output device 935 via I / O subsystem 920. In one embodiment, output device 935 is a digital computer display. Examples of displays that may be used in various embodiments include touch screen displays or light emitting diode (LED) displays or liquid crystal displays (LCD) or electronic paper displays. Computer system 905 may include other types of output devices 935 instead of or in addition to a display device. Examples of other output devices 935 include printers, ticket printers, plotters, projectors, sound cards or video cards, speakers, buzzers or piezoelectric devices or other audible devices, lamps or LEDs or LCD indicators, tactile devices, actuators, or servos.

[0151] At least one input device 940 is coupled to I / O subsystem 920 to communicate signals, data, command selections or gestures to processor 910. Examples of input devices 940 include touch screens, microphones, still and video digital cameras, alphanumeric and other keys, keypads, keyboards, graphic tablets, image scanners, joysticks, clocks, switches, buttons, dials, slides.

[0152] Another type of input device is the control device 945, which may perform cursor control or other automated control functions such as navigation within a graphical interface on a display screen, instead of or in addition to an input function. The control device 945 may communicate direction information and command selections to the processor 910 and may be a touchpad, mouse, trackball, or cursor direction keys for controlling cursor movement on the display 935. The input device may have at least two degrees of freedom in two axes, a first axis (e.g., x) and a second axis (e.g., y), that enable the device to specify a position within a plane. Another type of input device is a wired, wireless, or optical control device such as a joystick, wand, console, steering wheel, pedal, gear shift mechanism, or other type of control device. The input device 940 may include a combination of multiple different input devices such as a video camera and depth sensor.

[0153] In another embodiment, the computer system 905 may comprise a mono Internet of Things (IoT) device in which one or more of the output device 935, input device 940, and control device 945 are omitted. Or, in such an embodiment, the input device 940 may comprise one or more cameras, motion detectors, thermometers, microphones, seismometers, other sensors or detectors, measurement devices or encoders, and the output device 935 may comprise a dedicated display such as a single-line LED or LCD display, one or more indicators, a display panel, a meter, a valve, a solenoid, an actuator, or a servo.

[0154] Computer system 905 may implement the techniques described herein using customized hardwired logic, at least one ASIC or FPGA, firmware, and / or program instructions or logic, which, when loaded and used or executed in combination with the computer system, cause the computer system to operate or program as a dedicated machine. According to one embodiment, the techniques herein are performed by computer system 905 in response to a processor 910 executing at least one sequence of at least one instruction included in main memory 925. Such instructions may be read into main memory 925 from another storage medium such as storage 915. Execution of the instruction sequence included in main memory 925 causes processor 910 to perform the process steps described herein. In an alternative embodiment, hardwired circuitry may be used in place of, or in combination with, software instructions.

[0155] As used herein, the term "storage medium" refers to any non-transitory medium that stores data and / or instructions that cause a machine to operate in a particular manner. Such storage media may comprise non-volatile media and / or volatile media. Non-volatile media includes, for example, optical disks or magnetic disks such as storage 915. Volatile media includes dynamic memory such as memory 925. Common forms of storage media include, for example, hard disks, solid state drives, flash drives, magnetic data storage media, any optical or physical data storage media, memory chips, and the like.

[0156] Storage media is distinct from but may be used in conjunction with transmission media. Transmission media participates in the transfer of information between storage media. For example, transmission media includes coaxial cables, copper wire, and fiber optics, including the wires that comprise a bus of I / O subsystem 920. Transmission media can also take the form of acoustic or light waves, such as those generated during radio wave and infrared data communications.

[0157] Various forms of media may be involved in carrying at least one sequence of at least one instruction to the processor 910 for execution. For example, the instruction may first be carried on a magnetic disk or solid state drive of a remote computer. The remote computer can load the instruction into its dynamic memory and use a modem to transmit the instruction via a communication link such as an optical fiber or coaxial cable or telephone line. A modem or router local to the computer system 905 can receive data on the communication link and convert the data into a format readable by the computer system 905. For example, a receiver such as a radio frequency antenna or infrared detector can receive data carried by a radio signal or optical signal, and appropriate circuitry can provide the data to the I / O subsystem 920, such as placing the data on a bus. The I / O subsystem 920 carries the data to the memory 925, and the processor 910 fetches and executes the instruction from the memory. The instruction received by the memory 925 may optionally be stored in the storage 915 either before or after execution by the processor 910.

[0158] Computer system 905 also includes a communication interface 960 coupled to bus 920. Communication interface 960 provides bi-directional data communication coupled to a network link 965 directly or indirectly connected to at least one communication network such as network 970 or a public or private cloud on the Internet. For example, communication interface 960 may be an Ethernet networking interface, an Integrated Services Digital Network (ISDN) card, a cable modem, a satellite modem, or a corresponding type of communication line, such as an Ethernet cable or any type of metal cable, or a fiber optic line or a telephone line, and may be a modem for providing a data communication connection. Network 970 broadly represents a Local Area Network (LAN), a Wide Area Network (WAN), a campus network, an Internetwork, or any combination thereof. Communication interface 960 may include a LAN card for providing a data communication connection to a compatible LAN, or a cellular telephone interface wired to transmit or receive cellular data according to a cellular radio network standard, or a satellite radio interface wired to transmit or receive digital data according to a satellite radio network standard. In any such implementation, communication interface 960 transmits and receives electrical, electromagnetic, or optical signals via a signal path that carries a digital data stream representing various types of information.

[0159] Network link 965 typically provides electrical, electromagnetic, or optical data communication to other data devices directly or via at least one network using, for example, satellite, cellular, Wi-Fi, or BLUETOOTH technology. For example, network link 965 may provide a connection to host computer 950 via network 970.

[0160] Furthermore, network link 965 may provide a connection to other computing devices via network 970 or via Internet network devices and / or computers operated by an Internet service provider (ISP) 975. ISP 975 provides data communication services via a worldwide packet data communication network represented as the Internet 980. Server computer 955 may be coupled to the Internet 980. Server 955 broadly represents any computer, data center, virtual machine, or virtual computing instance, or a computer that runs a containerized program system such as DOCKER or KUBERNETES, regardless of the presence of a hypervisor. Server 955 may be implemented using multiple computers or instances and represents an electronic digital service that is accessed and used by sending web service requests, uniform resource locator (URL) strings with parameters in an HTTP payload, API calls, application service calls, or other service calls. Computer system 905 and server 955 may form elements of a distributed computing system that includes other computers, processing clusters, server farms, or other organizations of computers that cooperate to perform tasks or run applications or services. Server 955 may comprise one or more sets of instructions organized as a module, method, object, function, routine, or call. The instructions may be organized as an application program that includes one or more computer programs, operating system services, or mobile apps.The commands may include operating system and / or system software; one or more libraries supporting multimedia, programming, or other functions; data protocol instructions or stacks for implementing TCP / IP, HTTP, or other communication protocols; file format processing instructions for parsing or rendering files encoded using HTML, XML, JPEG, MPEG, or PNG; user interface instructions for rendering or interpreting commands for a graphical user interface (GUI), command line interface, or text user interface; application software such as office suites, Internet access applications, design and manufacturing applications, graphic applications, audio applications, software engineering applications, educational applications, games, or other applications. The server 955 may comprise a web application server hosting a presentation layer, an application layer, and a data storage layer such as a structured query language (SQL) or non-SQL relational database system, an object store, a graph database, a flat file system, or other data storage.

[0161] The computer system 905 can send messages and receive data and instructions including program code via a network, network link 965, and communication interface 960. In the example of the Internet, the server 955 can send the requested code of an application program via the Internet 980, ISP 975, local network 970, and communication interface 960. The received code may be executed by the processor 910 when received and / or stored in the storage 915 or other non-volatile storage for later execution.

[0162] The execution of instructions as described in this section may implement a process in the form of an instance of a computer program consisting of program code and its current activity. Depending on the operating system (OS), the process may consist of multiple execution threads that execute instructions simultaneously. In this regard, a computer program is a passive set of instructions, and a process may be the actual execution of those instructions. Multiple processes may be associated with the same program; for example, opening multiple instances of the same program often means that multiple processes are being executed. Multitasking may be implemented to allow multiple processes to share the processor 910. Each processor 910 or core of a processor executes a single task at a time, but the computer system 905 may be programmed to implement multitasking to allow the processor to switch the task being executed without having to wait for each task to complete. In one embodiment, the switch may occur when a task is performing an input / output operation, when a task indicates that it is switchable, or when executed by a hardware interrupt. Time slicing may be implemented to allow for a fast response to interactive user applications by quickly performing context switches to provide the appearance of simultaneous execution of multiple processes. In one embodiment, for security and reliability, the operating system may prevent direct communication between independent processes and provide a strictly mediated and controlled inter-process communication function.

Claims

1. 1. A method (400) for determining emotions or sensory perceptions related to exposure to a flavor or aroma component, comprising: inputting (405) at least two flavor or aroma physical component digital representation identifiers on a computer interface, the inputs corresponding to physical composition digital identifiers representing physical compositions of the physical flavor or aroma components; Executing (410) instructions by a computing device configured to associate at least one value with the input physical composition digital identifier, the value being: a relative ranking of the input flavor or aroma physical component digital representation identifiers to at least one determined perceived emotion or sensation; a value representing said perception of at least one determined perceived emotion or sensation for at least one input flavor or aroma physical component digital representation identifier; and / or a value representing a class of flavor or aroma components in a classification of flavor or aroma components by perceived emotional or sensory perception for at least one determined emotional or sensory perception; Step (410) representing providing (415) on a computer interface at least one value obtained during said operating step for said input physical composition digital identifier; A method (400) comprising:

2. 2. The method of claim 1, wherein the operating step comprises operating a trained gradient boosting decision tree device on the input physical composition digital identifier.

3. Providing (105) a set of example data to the gradient boosting decision tree device or to the neural network device, the set of data comprising: at least two flavor or aroma physical composition digital identifiers formed by the at least two flavor or aroma physical component digital representation identifiers; at least one emotional or sensory perceptual value for the at least two aroma or flavor components, the emotional or sensory perceptual value being associated with at least one emotional or sensory perceptual digital identifier, the emotional or sensory perceptual digital identifier representing a category of a human emotional or sensory response to the embodied physical composition digital identifier from a finite list of emotions or sensory responses; Step (105) including: operating (110) the gradient boosting decision tree device or neural network device based on the set of example data; Obtaining a trained gradient boosting decision tree model or a trained neural network model (115); The method (400) of claim 2, comprising:

4. The exemplary data further includes at least one digital identifier associated with the at least one physical composition digital identifier, the at least one digital identifier comprising: the gender of the human being exposed to the embodied physical composition (103); the country of origin of the human being exposed to the embodied physical composition (104); the type of use of the embodied physical composition (106); a composition chemical base (107) used to support the embodied physical composition; and / or a dose (108) of at least one physical component flavor or aroma physical component represented by said corresponding digital representation identifier; The method (400) of claim 3, wherein:

5. 3. The method (400) of claim 2, further comprising, downstream of the operating step (410), determining (420) by a computing device a numerical value representing an emotional or sensory response impact for at least one input flavor or aroma component, and providing (415) the determined numerical value of the response impact on a computer interface.

6. 6. The method (400) of claim 5, further comprising, downstream of the step (420) of determining a numerical value representing an emotional or sensory response impact, a step (435) of providing at least one alternative flavor or aroma physical component digital representation identifier for at least one input flavor or aroma physical component digital representation identifier, wherein the alternative physical composition digital identifier is formed as a function of a value representing an emotional or sensory response impact associated with the input and alternative flavor or aroma physical component digital representation identifiers.

7. 7. The method of claim 6, wherein the step of providing at least one alternative flavor or aroma physical component digital representation identifier is further configured to provide at least one value representing a concentration of the at least one alternative flavor or aroma physical component digital representation identifier.

8. 7. The method of claim 6, wherein the step of providing at least one alternative flavor or aroma physical component digital representation identifier is further configured to provide minimum and / or maximum values ​​representing concentrations of the at least one alternative flavor or aroma physical component digital representation identifier.

9. 10. The method (400) of claim 1, comprising assembling a physical composition corresponding to the input physical composition digital identifier or providing the input physical composition digital identifier to a system configured to assemble a physical composition (430).

10. Assembling (401) a database associating values ​​representing measured human emotions or sensory perceptions for at least one group of at least two flavor or aroma physical component digital representation identifiers; a step (407) of determining by the computing device a set of instructions to be operated during said step (410) of operating; The method (400) of claim 1, comprising:

11. The step of assembling a database (401) comprises: exposing (402) at least one human to a physical composition of flavor or aroma physical components; measuring (403) the emotion or sensory perception of the at least one human exposed to the physical composition; recording (404) a value representing the measured perceived emotion or sensation in a database in association with the group of flavor or aroma physical component digital representation identifiers representing the physical components used for the composition used during the exposing step; The method (400) of claim 10, comprising:

12. 10. The method of claim 1, further comprising: replacing an input flavor or aroma physical component digital representation identifier with a different, equivalent flavor or aroma physical component digital representation identifier (406), said equivalence being defined in a database of equivalent flavor or aroma physical component digital representation identifiers.

13. The step of operating (410) comprises: a first step (411) of associating, by a computing device, at least one olfactory or gustatory descriptor with at least one input flavor or aroma physical component digital representation identifier; a second step (412) of associating, by the computing device, at least one perceived emotion or sensation as a function of at least one olfactory or gustatory descriptor associated with at least one input flavor or aroma physical component digital representation identifier; The method (400) of claim 1, comprising:

14. 14. The method (400) of any one of claims 1 to 13, wherein at least one emotional or sensory perception represents a perception of a health or hygiene benefit associated with said exposure to said physical composition.

15. 1. A system (500) for determining emotions or sensory perceptions related to exposure to a flavor or aroma component, comprising: inputting (505) at least two flavor or aroma physical component digital representation identifiers on a computer interface (510), the inputs corresponding to physical composition digital identifiers representing physical compositions of the physical flavor or aroma components; Executing (515) instructions by a computing device (520) configured to associate at least one value with the input physical composition digital identifier, the value being: a relative ranking of the input flavor or aroma physical component digital representation identifiers to at least one determined perceived emotion or sensation; a value representing said perception of at least one determined perceived emotion or sensation for at least one input flavor or aroma physical component digital representation identifier; and / or a value representing a class of flavor or aroma components in a classification of flavor or aroma components by perceived emotional or sensory perception for at least one determined emotional or sensory perception; Step (515) representing providing (525) on a computer interface at least one value obtained during said operating step for said input physical composition digital identifier; A system (500) comprising:

Images