Systems and methods for optimizing recipes of food products using machine learning

Abstract

This disclosure relates to using machine learning, more specifically a trained generative adversarial network (GAN) for developing a flavour blend for a food product, such as chewing gum or other confectionary. The method comprises receiving, by one or more processors, a target profile for a food flavour, such as chewing gum and flavour data; predicting, by the one or more processors, one or more flavour blends based on the flavour data and the target profile; comparing, by the one or more processors, the one or more flavour blends to the target profile to identify at least one best flavour blend; and displaying, by the one or more processors, the at least one best flavour blend on at least one user interface of at least one user device.

Description

SYSTEMS AND METHODS FOR OPTIMIZING RECIPES OF FOOD PRODUCTS USING MACHINE LEARNINGCROSS-REFERENCE TO RELATED APPLICATION(S)

[0001] This patent application claims the benefit of priority to U.S. Application No. 63 / 579,953, filed on August 31 , 2023, the entirety of which is incorporated herein by reference.TECHNICAL FIELD

[0002] Various embodiments of this disclosure relate generally to systems and methods for using a machine-learning model to predict a flavor blend for food products, including confectionary products, such as chewing gum.BACKGROUND

[0003] Flavors are an incredibly important component of food products. The process of developing flavor recipes can be time consuming and costly. For example, conventional methods of developing flavors may include multiple rounds of experiments and a flavor recipe design process involving significant trial and error by experts, and subjective feedback from numerous experts and other individuals. Additionally, flavor designers and other experts typically analyze an extensive amount of data, including different ingredients, different ingredient interactions, and potential ingredient substitutes, when creating flavors. It can be difficult to leverage the significant amounts of flavor data from previous recipes when making new flavor recipes. As a result, a suitable supplement or alternative to the existing process is desired, so as to improve efficiency, reduce costs, and improve the overall value and results of the flavor blending process.

[0004] This disclosure is directed to addressing above-referenced challenges. The background description provided herein is for the purpose of generally presenting the context of the disclosure. Unless otherwise indicated herein, the materials described in this section are not prior art to the claims in this application and are not admitted to be prior art, or suggestions of the prior art, by inclusion in this section.SUMMARY OF THE DISCLOSURE

[0005] According to certain aspects of the disclosure, embodiments are disclosed for predicting a flavor blend for a food product such as chewing gum or other confectionary.

[0006] In one aspect, an exemplary embodiment of a method for predicting a flavor blend for a food product is disclosed. The method may include receiving, by one or more processors, a target profile for a food product and flavor data. The method may include predicting, by the one or more processors, one or more sensory profiles based on the flavor data and the target profile. The method may include comparing, by the one or more processors, the one or more sensory profiles to the target profile to identify at least one best sensory profile. The method may include displaying, by the one or more processors, the at least one best sensory profile on at least one user interface of at least one user device.

[0007] In one aspect, a computer system for predicting a flavor blend for a food product is disclosed. The computer system may comprise at least one memory storing instructions and at least one processor configured to execute the instructions to perform operations. The operations may comprise receiving a target profile for a food product and flavor data. The operations may comprise predicting one or more sensory profiles based on the flavor data and the target profile. The operations may comprise comparing the one or more sensory profiles to the target profile to identify at least one best sensory profile. The operations may comprise displaying the at least one best sensory profile on at least one user interface of at least one user device.

[0008] In one aspect, a non-transitory computer-readable medium storing instructions that, when executed by at least one processor, cause the at least one processor to perform operations for predicting a flavor blend for a food product. The operations may comprise receiving a target profile for a food product and flavor data and the target profile. The operations may comprise predicting one or more sensory profiles based on the flavor data. The operations may comprise comparing the one or more sensory profiles to the target profile to identify at least one best sensory profile. The operations may comprise displaying the at least one best sensory profile on at least one user interface of at least one user device.

[0009] It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the disclosed embodiments, as claimed.BRIEF DESCRIPTION OF THE DRAWINGS

[0010] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate various exemplary embodiments and together with the description, serve to explain the principles of the disclosed embodiments.

[0011] FIG. 1 depicts a flowchart of an exemplary method for predicting a flavor blend for a food product, such as a chewing gum or other confectionary product, according to one or more embodiments.

[0012] FIG. 2 depicts a flow chart of an exemplary method for training a machine-learning model to predict a flavor blend, according to one or more embodiments.

[0013] FIG. 3 depicts an exemplary networked computing environment that may be utilized with techniques presented herein, according to one or more embodiments.

[0014] FIG. 4 depicts an example of a computing device that may execute the techniques described herein, according to one or more embodiments.DETAILED DESCRIPTION OF EMBODIMENTS

[0015] According to certain aspects of the disclosure, methods and systems for predicting flavor blends of food products including confectionary products, such as chewing gum, are disclosed.

[0016] Determining the proper ingredients for flavor blends for food products can be a time consuming and costly process. For example, conventional methods of developing a flavor blend may include a series of experimental prototypes by human experts. Such methods may include multiple rounds of experiments and an extensive flavor recipe design process. Moreover, it may also be difficult to leverage previous experimental data and results when creating different new flavor recipes. As a result, a recent desire exists for a suitable supplement or alternative to the existing experimental process of developing flavor blends, so as to improve efficiency, reduce costs, and improve the overall value and results of the flavor blending process.

[0017] The present disclosure describes an optimal flavor system and methods for designing and blending flavors for food products, including confectionaries such as chewing gum through experimental and computational methods. It should be appreciated that the present disclosure describes such a flavor system and methods with reference to chewing gum solely as an example, and that the presently disclosed systems and methods are applicable to the flavor design of any other confectionary, or even any other food product in general. However, the present disclosure describes the use of these systems and methods as applied to chewing gum because of its unique sensitivity to flavor profile, and duration of flavor profile, particularly as it relates to texture and duration of sweetness.

[0018] In particular, when applied to the design of confectionary products, such as chewing gum, the presently disclosed flavor system and methods may leverage both human knowledge and machine-learning algorithms to optimize the interaction of flavor volatile compounds, flavor, flavor duration, sweetness, sweetness duration, and texture. The systematic approach to designing the flavor blend for food products, such as chewing gum, may include analyzing the flavor chemical property data, historical sensory results, and a domain expert’s knowledge, as well as applying a machine-learning model to optimize multiple response variables simultaneously. Such response variables may include, for example: sensory descriptions, ingredient costs, product stability, and any user specified responses.

[0019] As will be discussed in more detail below, in various embodiments, systems and methods are described for predicting a flavor blend for food products, such as chewing gum. The systems and methods may include receiving a target profile for a food product and flavor data. The systems and methods may include predicting one or more sensory profiles based on the flavor data and the target profile. The systems and methods may include comparing the one or more sensory profiles to the target profile to identify at least one best sensory profile. The systems and methods may include displaying the at least one best sensory profile on at least one user interface of at least one user device.

[0020] The terminology used below may be interpreted in its broadest reasonable manner, even though it is being used in conjunction with a detailed description of certain specific examples of the present disclosure. Indeed, certainterms may even be emphasized below; however, any terminology intended to be interpreted in any restricted manner will be overtly and specifically defined as such in this Detailed Description section. Both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the features, as claimed.

[0021] In the detailed description herein, references to “embodiment,” “an embodiment,” “one non-limiting embodiment,” “in various embodiments,” etc., indicate that the embodiment(s) described can include a particular feature, structure, or characteristic, but every embodiment might not necessarily include the particular feature, structure, or characteristic. Moreover, such phrases are not necessarily referring to the same embodiment. Further, when a particular feature, structure, or characteristic is described in connection with an embodiment, it is submitted that it is within the knowledge of one skilled in the art to affect such feature, structure, or characteristic in connection with other embodiments whether or not explicitly described. After reading the description, it will be apparent to one skilled in the relevant art(s) how to implement the disclosure in alternative embodiments.

[0022] In general, terminology can be understood at least in part from usage in context. For example, terms, such as “and”, “or”, or “and / or,” as used herein can include a variety of meanings that may depend at least in part upon the context in which such terms are used. Typically, “or” if used to associate a list, such as A, B or C, is intended to mean A, B, and C, here used in the inclusive sense, as well as A, B or C, here used in the exclusive sense. In addition, the term “one or more” as used herein, depending at least in part upon context, can be used to describe any feature, structure, or characteristic in a singular sense or can be used to describe combinations of features, structures or characteristics in a plural sense. Similarly, terms, such as “a,” “an,” or “the,” again, can be understood to convey a singular usage or to convey a plural usage, depending at least in part upon context. In addition, the term “based on” can be understood as not necessarily intended to convey an exclusive set of factors and can, instead, allow for existence of additional factors not necessarily expressly described, again, depending at least in part on context.

[0023] As used herein, the terms “comprises,” “comprising,” or any other variation thereof are intended to cover a non-exclusive inclusion, such that a process, method, composition, article, or apparatus that comprises a list of elementsdoes not include only those elements, but may include other elements not expressly listed or inherent to such process, method, composition, article, or apparatus. The term “exemplary” is used in the sense of “example” rather than “ideal.” As used herein, the singular forms “a,” “an,” and “the” include plural reference unless the context dictates otherwise. Relative terms such as “about,” “substantially,” and “approximately” refer to being nearly the same as a referenced number or value, and should be understood to encompass a variation of ±5% of a specified amount or value.

[0024] As used herein, “taste” refers to a sensation caused by activation or inhibition of receptor cells in a subject's oral cavity. Different types of taste may include sweet, sour, salt, bitter, kokumi, umami, and any combination thereof.

[0025] The terms “flavor,” “flavor blend,” and “flavor compound” may refer to a ligand or a compound that is capable of imparting a flavor.

[0026] As used herein, a “model” or “machine-learning model” generally encompasses instructions, data, and / or a model configured to receive input, and apply one or more of a weight, bias, classification, or analysis on the input to generate an output. The output may include, for example, a classification of the input, an analysis based on the input, a design, process, prediction, or recommendation associated with the input, or any other suitable type of output. A machine-learning model is generally trained using training data, e.g., experiential data and / or samples of input data, which are fed into the model in order to establish, tune, or modify one or more aspects of the model, e.g., the weights, biases, criteria for forming classifications or clusters, or the like. Aspects of a machine-learning model may operate on an input linearly, in parallel, via a network (e.g., a neural network), or via any suitable configuration.

[0027] The execution of the machine-learning model may include deployment of one or more machine learning techniques, such as linear regression, logistical regression, random forest, gradient boosted machine (GBM), deep learning, and / or a deep neural network. Supervised and / or unsupervised training may be employed. For example, supervised learning may include providing training data and labels corresponding to the training data, e.g., as ground truth. Unsupervised approaches may include clustering, classification or the like. Any suitable type of training may be used, e.g., stochastic, gradient boosted, random seeded, recursive, epoch or batch-based, etc.

[0028] All percentages and proportions used herein are by weight of the total composition.

[0029] Certain non-limiting embodiments are described below with reference to block diagrams and operational illustrations of methods, processes, devices, and apparatus. It is understood that each block of the block diagrams or operational illustrations, and combinations of blocks in the block diagrams or operational illustrations, can be implemented by means of analog or digital hardware and computer program instructions. These computer program instructions can be provided to a processor of a general purpose computer to alter its function as detailed herein, a special purpose computer, ASIC, or other programmable data processing apparatus, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, implement the functions / acts specified in the block diagrams or operational block or blocks. In some alternate implementations, the functions / acts noted in the blocks can occur out of the order noted in the operational illustrations. For example, two blocks shown in succession can in fact be executed substantially concurrently or the blocks can sometimes be executed in the reverse order, depending upon the functionality / acts involved.Predicting a Flavor for a Food Product

[0030] FIG. 1 depicts a flowchart of an exemplary method 100 for predicting a flavor blend for a food product, such as chewing gum, according to one or more embodiments. Again, as described above, even though FIG. 1 describes method 100 in relation to chewing gum as an example, it should be appreciated that the disclosed flavor systems and methods are applicable to the flavor design of any other confectionary, or even any other food product in general. In addition, method 100 may be performed by one or more processors of a server that is in communication with one or more user devices and other external system(s) via a network. However, it should be noted that method 100 may be performed by any one or more of the server, one or more user devices, or other external systems.

[0031] Turning again to FIG. 1 , method 100 may include receiving, by one or more processors, a target profile for a food product (e.g., chewing gum) and flavor data (Step 102). A user may create and / or input the target profile into the system. Alternatively, the target profile may have been created and input by an internal orexternal system. For example, a machine-learning model may have created the target profile. In some embodiments, the target profile may include at least one best sensory profile (e.g., at least one preferred sensory profile), as described below.

[0032] The target profile may describe a target flavor blend for a food product, such as chewing gum. For example, the target profile may include one or more flavor characteristics, spectral data, or sensory data. The target profile may include analytical data (GC-MS spectral data) or sensory data (target description), such as spicy and sweet blueberry flavor. The one or more flavor characteristics may include characteristics of the flavor blend in the target profile. The one or more flavor characteristics may include at least one of: ethereal, fruity, sweet, buttery, creamy, sharp, banana, almond, cherry, floral, citrus, orange, green apple, cinnamyl, balsamic, spicy, woody, peach, berry, drying, or jammy. The sensory data may describe the smell, touch, taste, or visual description of the target profile.

[0033] The flavor data may describe details corresponding to one or more flavors that include one or more ingredients. The flavor data may have been previously created and stored in the system (e.g., data storage). For example, the flavor data may have been stored in an in-house database (server). The flavor data may include domain knowledge, one or more material descriptors, and / or one or more formula descriptors. The domain knowledge may describe specific details about the flavors and the ingredients. The domain knowledge may include at least one ingredient interaction mechanism or at least one historical insight. The ingredient interaction mechanism may include details regarding how the one or more ingredients interact with other ingredients and / or other components of the gum (e.g., gum base, sweetener, coolant, and the like). The historical insight may describe previous attributes, structure, and / or other characteristics of the one or more ingredients. For example, the historical insight may describe how an ingredient previously reacted with a gum base. The historical insight may provide information related to an understanding of key chemosensates of a specific targeted flavor type and other flavor chemical compounds to modulate tonality. The one or more material descriptors may include one or more chemical structures, one or more chemical interactions, one or more partition coefficients, or one or more water solubility and dissolution rates. The chemical structures may describe one or more chemical structures of the one or more ingredients. The one or more chemical interactions may describe chemical interactions between the ingredients. Forexample, the one or more chemical interactions may include interactions between one or more flavors and one or more other product variables, such as a gum base, sweetener, and / or modifiers which may impact a flavor sensory perception. The one or more formula descriptors may describe the characteristics of the formula. For example, the one or more formula descriptors may include one or more qualitative formula characteristics or one or more quantitative formula characteristics. The qualitative formula characteristics may include formula descriptors, such as sensory or product descriptors (e.g., oral refreshing, cognitive alert). The quantitative formula characteristics may include regulatory limits, standard of identification, and the like.

[0034] The method may further include predicting, by the one or more processors, one or more sensory profiles based on the flavor data (Step 104). The sensory profiles may include data describing one or more flavor blends. For example, the flavor blends may include one or more ingredients, one or more ingredient amounts to achieve a particular flavor, and / or one or more descriptors. When combined, the ingredients may create a flavor blend, where the ingredient amounts may indicate an amount of each ingredient that should be included in the flavor blend. The ingredient amount may be expressed as an exact amount, a percentage, and / or an amount range. In some embodiments, the sensory profiles may include one or more descriptors, such as ethereal, fruity, sweet, buttery, creamy, sharp, banana, almond, cherry, floral, citrus, orange, green apple, cinnamyl, balsamic, spicy, woody, peach, berry, drying, or jammy. In some embodiments, the sensory profiles may include chemical structures of the one or more flavor blends.

[0035] In some embodiments, the predicting may be performed by a trained machine-learning model. The machine-learning model may receive the flavor data and the target profile as input, where the machine-learning model may analyze the flavor data to determine one or more sensory profiles. For example, the machinelearning model may analyze the chemical structures included in the flavor data and the target profile, where the machine-learning model may determine sensory profiles that include chemical structures that are analogous to the chemical structure in the target profile. Additionally, or alternatively, the machine-learning model may analyze the flavor data to determine the sensory profiles that have common release profiles to the target profile.

[0036] The machine-learning model may optimize the one or more sensory profiles. For example, the machine-learning model may have been trained tooptimize the one or more sensory profiles. Optimizing the one or more sensory profiles may include analyzing the ingredients in the sensory profiles and determining at least one optimized ingredient replacement. The optimized ingredient replacement may correspond to at least one ingredient that optimizes the sensory profile. For example, the at least one optimized ingredient replacement may substitute one or more ingredients in a sensory profile. The optimized ingredient replacement may include a cheaper ingredient than the ingredient(s) originally included in the sensory profile. Additionally, or alternatively, the optimized ingredient replacement may include an ingredient that provides a substitute for at least two ingredients originally included in the sensory profile. Additionally, or alternatively, optimizing the one or more sensory profiles may include modifying one or more of the ingredient amounts. For example, an amount for one ingredient in the sensory profile may be decreased, and an optimized ingredient replacement may substitute two ingredients in the sensory profile.

[0037] The method may further include comparing, by the one or more processors, the one or more sensory profiles to the target profile to identify at least one best sensory profile (Step 106). In some embodiments, a trained machinelearning model may perform the comparing of the sensory profiles to the target profile. In some embodiments, the comparing may further include determining, by the trained machine-learning model, that the at least one best sensory profile is a closest match to the target profile. For example, the trained machine-learning model may compare each sensory profile to the target profile. Based on the comparison, the trained machine-learning model may identify at least one sensory profile that is a closest match to the target profile. For example, the sensory profile may have the most similar chemical structure to the target profile. Additionally, or alternatively, the at least one best sensory profile may meet an odor threshold. For example, the odor threshold may indicate a certain level of odor and / or a certain type of odor for the flavor blend of the sensory profile. The system may identify such a sensory profile as the at least one best sensory profile. For example, the system may flag the sensory profile as the best sensory profile (e.g., the preferred sensory profile).

[0038] The method may further include displaying, by the one or more processors, the at least one best sensory profile on at least one user interface of at least one user device (Step 108). The user interface may display one or more graphical representations of the at least one best sensory profile. For example, thegraphical representation may include a list of the ingredients in the at least one best sensory profile. In some embodiments, the user interface may also display additional food product components for the food product. For example, the user interface may display additional chewing gum components for the chewing gum (e.g., gum base, flavor, sweetener, coolant, and the like). The user interface may also display one or more descriptors of the best sensory profile (e.g., fruity). The user interface may also display corresponding ratios for each of the additional gum components. The user interface may also display interactive elements that allow a user to provide feedback. The interactive elements may also allow a user to indicate whether one or more ingredients of the best sensory profile are / are not a good fit. The interactive elements may also receive updates to the ingredients and / or ingredient amounts. For example, upon receiving the updated ingredients and / or ingredient amounts, the system may store such updated information for future use. In some embodiments, the best sensory profile may be used as flavor data or a target profile for determining a future best profile.

[0039] Although FIG. 1 shows example blocks of exemplary method 100, in some implementations, the exemplary method 100 may include additional blocks, fewer blocks, different blocks, or differently arranged blocks than those depicted in FIG. 1 . Additionally, or alternatively, two or more of the blocks of the exemplary method 100 may be performed in parallel.

[0040] The chart below displays an exemplary best profile for a fruit flavor. The chart displays a list of ingredients (e.g., compound), an ingredient amount (e.g., %), and one or more descriptors. The descriptors for the best profile may include sweet, fruity, jammy, floral, woody, and drying.Training a Machine-Learning Mode to Predict a Flavor for a Food Product

[0041] FIG. 2 depicts a flowchart of an exemplary method 200 for training a machine-learning model for predicting a flavor for a food product, such as a chewing gum or any other confectionary, according to one or more embodiments. Notably, method 200 may be performed by one or more processors of a server that is in communication with one or more user devices and other external system(s) via a network. However, it should be noted that method 200 may be performed by any one or more of the server, one or more user devices, or other external systems.

[0042] In some embodiments, the trained machine-learning model may include a generative adversarial network (GAN). A flavor formula generator may receive flavor data (e.g., random flavor compounds) (Step 202). Based on the flavor data, the flavor formula generator may create one or more synthetic or “fake” flavor samples (Step 204) and feed such fake flavor samples into a discriminator (Step 206). The discriminator may receive the fake flavor samples from the generator (Step 210). Additionally, an internal or external system may feed real flavor samples into the discriminator (Step 208). The discriminator may also receive the real flavor samples. The discriminator may then predict a response profile based on the fake and real flavor samples (Step 210). The response profile may include analytical data (e.g., GC-MS) and / or sensory data (e.g., descriptive analysis). For example, the discriminator may classify the received flavor samples as “real” or “fake” (Steps 212 and 214). Based on the classification, the discriminator loss may penalize the discriminator for misclassifying the flavor samples as “real” or “fake” (Steps 216 and 218). The fake flavor sample data may be adjusted and fed into the generator for further training (Steps 220 and 222).

[0043] For example, in a GAN model, there may be two main components: the generator and the discriminator. The generator may include a neural network that may take a random noise vector as input and produce a synthetic data samplethat may look like it came from the real data distribution. The discriminator may also include a neural network that may take a data sample (either real or synthetic) as input and output a probability that the sample may be real. The goal of the generator may be to produce synthetic data samples that are indistinguishable from real data samples. To do this, the generator may learn the underlying distribution of the real data. The goal of the discriminator may be to correctly classify data samples as real or fake. The discriminator may be trained to distinguish between real and synthetic data samples, providing feedback to the generator about how well it is doing at generating realistic data. During training, the generator and discriminator may be updated iteratively. The generator may try to generate data samples that may fool the discriminator into thinking they are real, while the discriminator tries to correctly classify the samples. The training process may be thought of as a game between the generator and discriminator, where the generator may try to generate more realistic samples, and the discriminator may try to become better at distinguishing between real and fake samples. The training loss for the generator may be calculated using the following equation:Lg= -E[log(D(x))] + log(l - D(G(z))) where x is the generated sample, D is the discriminator, and E is the expected value operator. This equation may mean that the generator may be trying to maximize the log-likelihood of the discriminator outputting a high probability that the generated sample may be real. In other words, the generator may want the discriminator to think that the generated sample may be real. The training loss for the discriminator may be calculated using the following equation:Ld= -E[log(D(x))] + log(l - D(G(z))) where G is the generator, z is a random noise vector used as input to the generator, and x is the real data sample. This equation may mean that the discriminator is trying to minimize the log-likelihood of misclassifying a real data sample as fake, while also minimizing the log-likelihood of mistaking a generated sample for a real one. During training, the generator and discriminator may be updated alternately. The generator may updates its parameters to increase the likelihood of the discriminator outputting a high probability that the generated sample may be real. The discriminator may update its parameters to improve its ability to distinguish between real and fake samples. Overall, the training process in a GAN may involvesan adversarial game between the generator and discriminator, where the generator may try to generate more realistic samples, and the discriminator may try to become better at distinguishing between real and fake samples. Through this process, both the generator and discriminator improve in performance, and the generator may eventually learn to produce highly realistic data samples that may be similar to the real data distribution.

[0044] Although FIG. 2 shows example blocks of exemplary method 200, in some implementations, the exemplary method 200 may include additional blocks, fewer blocks, different blocks, or differently arranged blocks than those depicted in FIG. 2. Additionally, or alternatively, two or more of the blocks of the exemplary method 200 may be performed in parallel.Exemplary Environment and Exemplary Device

[0045] FIG. 3 depicts an exemplary environment 300 that may be utilized with techniques presented herein. One or more user device(s) 305, one or more external system(s) 310, and one or more server system(s) 315 may communicate across a network 301 . As will be discussed in further detail below, one or more server system(s) 315 may communicate with one or more of the other components of the environment 300 across network 301 . The one or more user device(s) 305 may be associated with a user, e.g., a user associated with one or more of generating, training, or tuning a machine-learning model for predicting a flavor blend for a food product, such as chewing gum.

[0046] In some embodiments, the components of the environment 300 are associated with a common entity, e.g., a food producer, research center, “flavor house,” or the like. In some embodiments, one or more of the components of the environment is associated with a different entity than another. The systems and devices of the environment 300 may communicate in any arrangement. As will be discussed herein, systems and / or devices of the environment 300 may communicate in order to one or more of generate, train, and / or use a machine-learning model to predict a flavor blend for a food product, such as chewing gum, among other activities.

[0047] The user device 305 may be configured to enable the user to access and / or interact with other systems in the environment 300. For example, the user device 305 may be a computer system such as, for example, a desktop computer, amobile device, a tablet, etc. In some embodiments, the user device 305 may include one or more electronic application(s), e.g., a program, plugin, browser extension, etc., installed on a memory of the user device 305.

[0048] The user device 305 may include a display / user interface (III) 305A, a processor 305B, a memory 305C, and / or a network interface 305D. The user device 305 may execute, by the processor 305B, an operating system (O / S) and at least one electronic application (each stored in memory 305C). The electronic application may be a desktop program, a browser program, a web client, or a mobile application program (which may also be a browser program in a mobile O / S), an applicant specific program, system control software, system monitoring software, software development tools, or the like. For example, environment 300 may extend information on a web client that may be accessed through a web browser. In some embodiments, the electronic application(s) may be associated with one or more of the other components in the environment 300. The application may manage the memory 305C, such as a database, to transmit streaming data to network 301 . The display / U I 305A may be a touch screen or a display with other input systems (e.g., mouse, keyboard, etc.) so that the user(s) may interact with the application and / or the O / S. The network interface 305D may be a TCP / IP network interface for, e.g., Ethernet or wireless communications with the network 301. The processor 305B, while executing the application, may generate data and / or receive user inputs from the display / UI 305A and / or receive / transmit messages to the server system 315, and may further perform one or more operations prior to providing an output to the network 301.

[0049] External systems 310 may be, for example, one or more third party and / or auxiliary systems that integrate and / or communicate with the server system 315 in performing various flavor blend prediction tasks. External systems 310 may be in communication with other device(s) or system(s) in the environment 300 over the one or more networks 301 . For example, external systems 310 may communicate with the server system 315 via API (application programming interface) access over the one or more networks 301 , and also communicate with the user device(s) 305 via web browser access over the one or more networks 301 .

[0050] In various embodiments, the network 301 may be a wide area network (“WAN”), a local area network (“LAN”), a personal area network (“PAN”), or the like. In some embodiments, network 301 includes the Internet, and informationand data provided between various systems occurs online. “Online” may mean connecting to or accessing source data or information from a location remote from other devices or networks coupled to the Internet. Alternatively, “online” may refer to connecting or accessing a network (wired or wireless) via a mobile communications network or device. The Internet is a worldwide system of computer networks — a network of networks in which a party at one computer or other device connected to the network can obtain information from any other computer and communicate with parties of other computers or devices. The most widely used part of the Internet is the World Wide Web (often-abbreviated “WWW’ or called “the Web”). A “website page” generally encompasses a location, data store, or the like that is, for example, hosted and / or operated by a computer system so as to be accessible online, and that may include data configured to cause a program such as a web browser to perform operations such as send, receive, or process data, generate a visual display and / or an interactive interface, or the like.

[0051] The server system 315 may include an electronic data system, e.g., a computer-readable memory such as a hard drive, flash drive, disk, etc. In some embodiments, the server system 315 includes and / or interacts with an application programming interface for exchanging data to other systems, e.g., one or more of the other components of the environment.

[0052] The server system 315 may include a database 315A and at least one server 315B. The server system 315 may be a computer, system of computers (e.g., rack server(s)), and / or or a cloud service computer system. The server system may store or have access to database 315A (e.g., hosted on a third party server or in memory 315E). The server(s) may include a display / UI 315C, a processor 315D, a memory 315E, and / or a network interface 315F. The display / UI 315C may be a touch screen or a display with other input systems (e.g., mouse, keyboard, etc.) for an operator of the server 315B to control the functions of the server 315B. The server system 315 may execute, by the processor 315D, an operating system (O / S) and at least one instance of a servlet program (each stored in memory 315E).

[0053] The server system 315 may generate, store, train, or use a machinelearning model, configured to predict a flavor blend for a food product, such as chewing gum. The server system 315 may include a machine-learning model and / or instructions associated with the machine-learning model, e.g., instructions for generating a machine-learning model, training the machine-learning model, using themachine-learning model, etc. The server system 315 may include instructions for predicting a flavor blend for a food product, such as chewing gum, e.g., based on the output of the machine-learning model, and / or operating the display 315C to output flavor data, e.g., as adjusted based on the machine-learning model. The server system 315 may include training data, e.g., at least one sensory profile, at least one target profile, and at least one flavor.

[0054] In some embodiments, a system or device other than the server system 315 is used to generate and / or train the machine-learning model. For example, such a system may include instructions for generating the machinelearning model, the training data and ground truth, and / or instructions for training the machine-learning model. A resulting trained machine-learning model may then be provided to the server system 315.

[0055] Generally, a machine-learning model includes a set of variables, e.g., nodes, neurons, filters, etc., that are tuned, e.g., weighted or biased, to different values via the application of training data. In supervised learning, e.g., where a ground truth is known for the training data provided, training may proceed by feeding a sample of training data into a model with variables set at initialized values, e.g., at random, based on Gaussian noise, a pre-trained model, or the like. The output may be compared with the ground truth to determine an error, which may then be back- propagated through the model to adjust the values of the variable.

[0056] Training may be conducted in any suitable manner, e.g., in batches, and may include any suitable training methodology, e.g., stochastic or non-stochastic gradient descent, gradient boosting, random forest, etc. In some embodiments, a portion of the training data may be withheld during training and / or used to validate the trained machine-learning model, e.g., compare the output of the trained model with the ground truth for that portion of the training data to evaluate an accuracy of the trained model. The training of the machine-learning model may be configured to cause the machine-learning model to learn associations between the target profile, flavor data, and the best sensory profile, such that the trained machine-learning model is configured to determine a best sensory profile in response to the input target profile and flavor data based on the learned associations.

[0057] In various embodiments, the variables of a machine-learning model may be interrelated in any suitable arrangement in order to generate the output. For example, the machine-learning model may include one or more convolutional neuralnetwork (“CNN”) configured to identify flavor blends, and may include further architecture, e.g., a connected layer, neural network, etc., configured to determine a relationship between the identified features in order to determine a predicted flavor blend for a food product, such as chewing gum.

[0058] Although depicted as separate components in FIG. 3, it should be understood that a component or portion of a component in the environment 300 may, in some embodiments, be integrated with or incorporated into one or more other components. For example, a portion of the display 315C may be integrated into the user device 305 or the like. In some embodiments, operations or aspects of one or more of the components discussed above may be distributed amongst one or more other components. Any suitable arrangement and / or integration of the various systems and devices of the environment 300 may be used.

[0059] Further aspects of the machine-learning model and / or how it may be utilized to predict a sweetness intensity are discussed in further detail in the methods above. In the following methods, various acts may be described as performed or executed by a component from FIG. 3, such as the server system 315, the user device 305, or components thereof. However, it should be understood that in various embodiments, various components of the environment 300 discussed above may execute instructions or perform acts including the acts discussed below. An act performed by a device may be considered to be performed by a processor, actuator, or the like associated with that device. Further, it should be understood that in various embodiments, various steps may be added, omitted, and / or rearranged in any suitable manner.

[0060] In general, any process or operation discussed in this disclosure that is understood to be computer-implementable, such as the processes illustrated in FIGS. 1 and 2, may be performed by one or more processors of a computer system, such any of the systems or devices in the environment 300 of FIG. 3, as described above. A process or process step performed by one or more processors may also be referred to as an operation. The one or more processors may be configured to perform such processes by having access to instructions (e.g., software or computer-readable code) that, when executed by the one or more processors, cause the one or more processors to perform the processes. The instructions may be stored in a memory of the computer system. A processor may be a centralprocessing unit (CPU), a graphics processing unit (GPU), or any suitable types of processing unit.

[0061] A computer system, such as a system or device implementing a process or operation in the examples above, may include one or more computing devices, such as one or more of the systems or devices in FIG. 3. One or more processors of a computer system may be included in a single computing device or distributed among a plurality of computing devices. A memory of the computer system may include the respective memory of each computing device of the plurality of computing devices.

[0062] FIG. 4 is a simplified functional block diagram of a computer 400 that may be configured as a device for executing the methods of FIGS. 1 and 2, according to exemplary embodiments of the present disclosure. For example, device 400 may include a central processing unit (CPU) 420. CPU 420 may be any type of processor device including, for example, any type of special purpose or a general-purpose microprocessor device. As will be appreciated by persons skilled in the relevant art, CPU 420 also may be a single processor in a multi- core / multiprocessor system, such system operating alone, or in a cluster of computing devices operating in a cluster or server farm. CPU 420 may be connected to a data communication infrastructure 410, for example, a bus, message queue, network, or multi-core message-passing scheme.

[0063] Device 400 also may include a main memory 440, for example, random access memory (RAM), and also may include a secondary memory 430. Secondary memory 430, e.g., a read-only memory (ROM), may be, for example, a hard disk drive or a removable storage drive. Such a removable storage drive may comprise, for example, a floppy disk drive, a magnetic tape drive, an optical disk drive, a flash memory, or the like. The removable storage drive in this example reads from and / or writes to a removable storage unit in a well-known manner. The removable storage unit may comprise a floppy disk, magnetic tape, optical disk, etc., which is read by and written to by the removable storage drive. As will be appreciated by persons skilled in the relevant art, such a removable storage unit generally includes a computer usable storage medium having stored therein computer software and / or data.

[0064] In alternative implementations, secondary memory 430 may include other similar means for allowing computer programs or other instructions to beloaded into device 400. Examples of such means may include a program cartridge and cartridge interface (such as that found in video game devices), a removable memory chip (such as an EPROM, or PROM) and associated socket, and other removable storage units and interfaces, which allow software and data to be transferred from a removable storage unit to device 400.

[0065] Device 400 also may include a communications interface (“COM”) 460. Communications interface 460 allows software and data to be transferred between device 400 and external devices. Communications interface 460 may include a modem, a network interface (such as an Ethernet card), a communications port, a PCMCIA slot and card, or the like. Software and data transferred via communications interface 460 may be in the form of signals, which may be electronic, electromagnetic, optical, or other signals capable of being received by communications interface 460. These signals may be provided to communications interface 460 via a communications path of device 400, which may be implemented using, for example, wire or cable, fiber optics, a phone line, a cellular phone link, an RF link or other communications channels.

[0066] The hardware elements, operating systems and programming languages of such equipment are conventional in nature, and it is presumed that those skilled in the art are adequately familiar therewith. Device 400 also may include input and output ports 450 to connect with input and output devices such as keyboards, mice, touchscreens, monitors, displays, etc. Of course, the various server functions may be implemented in a distributed fashion on a number of similar platforms, to distribute the processing load. Alternatively, the servers may be implemented by appropriate programming of one computer hardware platform.

[0067] Program aspects of the technology may be thought of as “products” or “articles of manufacture” typically in the form of executable code and / or associated data that is carried on or embodied in a type of machine-readable medium. “Storage” type media include any or all of the tangible memory of the computers, processors or the like, or associated modules thereof, such as various semiconductor memories, tape drives, disk drives and the like, which may provide non-transitory storage at any time for the software programming. All or portions of the software may at times be communicated through the Internet or various other telecommunication networks. Such communications, for example, may enable loading of the software from one computer or processor into another, for example,from a management server or host computer of the mobile communication network into the computer platform of a server and / or from a server to the mobile device. Thus, another type of media that may bear the software elements includes optical, electrical and electromagnetic waves, such as used across physical interfaces between local devices, through wired and optical landline networks and over various air-links. The physical elements that carry such waves, such as wired or wireless links, optical links, or the like, also may be considered as media bearing the software. As used herein, unless restricted to non-transitory, tangible “storage” media, terms such as computer or machine “readable medium” refer to any medium that participates in providing instructions to a processor for execution.

[0068] Reference to any particular activity is provided in this disclosure only for convenience and not intended to limit the disclosure. A person of ordinary skill in the art would recognize that the concepts underlying the disclosed devices and methods may be utilized in any suitable activity. The disclosure may be understood with reference to the following description and the appended drawings, wherein like elements are referred to with the same reference numerals.

[0069] The terminology used above may be interpreted in its broadest reasonable manner, even though it is being used in conjunction with a detailed description of certain specific examples of the present disclosure. Indeed, certain terms may even be emphasized above; however, any terminology intended to be interpreted in any restricted manner will be overtly and specifically defined as such in this Detailed Description section. Both the general description and the detailed description are exemplary and explanatory only and are not restrictive of the features, as claimed.

[0070] It should be appreciated that in the above description of exemplary embodiments of the invention, various features of the invention are sometimes grouped together in a single embodiment, figure, or description thereof for the purpose of streamlining the disclosure and aiding in the understanding of one or more of the various inventive aspects. This method of disclosure, however, is not to be interpreted as reflecting an intention that the claimed invention requires more features than are expressly recited in each claim. Rather, as the following claims reflect, inventive aspects lie in less than all features of a single foregoing disclosed embodiment. Thus, the claims following the Detailed Description are herebyexpressly incorporated into this Detailed Description, with each claim standing on its own as a separate embodiment of this invention.

[0071] Furthermore, while some embodiments described herein include some but not other features included in other embodiments, combinations of features of different embodiments are meant to be within the scope of the invention, and form different embodiments, as would be understood by those skilled in the art. For example, in the following claims, any of the claimed embodiments can be used in any combination.

[0072] Thus, while certain embodiments have been described, those skilled in the art will recognize that other and further modifications may be made thereto without departing from the spirit of the invention, and it is intended to claim all such changes and modifications as falling within the scope of the invention. For example, functionality may be added or deleted from the block diagrams and operations may be interchanged among functional blocks. Steps may be added or deleted to methods described within the scope of the present invention.

[0073] The above disclosed subject matter is to be considered illustrative, and not restrictive, and the appended claims are intended to cover all such modifications, enhancements, and other implementations, which fall within the true spirit and scope of the present disclosure. Thus, to the maximum extent allowed by law, the scope of the present disclosure is to be determined by the broadest permissible interpretation of the following claims and their equivalents, and shall not be restricted or limited by the foregoing detailed description. While various implementations of the disclosure have been described, it will be apparent to those of ordinary skill in the art that many more implementations are possible within the scope of the disclosure. Accordingly, the disclosure is not to be restricted except in light of the attached claims and their equivalents.

Claims

What is claimed is:1 . A computer-implemented method for predicting a flavor blend for a food product, the method comprising: receiving, by one or more processors, a target profile for a food product and flavor data, wherein the target profile includes one or more flavor characteristics, spectral data, or sensory data; predicting, by the one or more processors, via a trained machine-learning model, one or more sensory profiles based on the flavor data and the target profile, wherein the one or more sensory profiles include one or more flavor blends that include one or more ingredients and one or more ingredient amounts; comparing, by the one or more processors, via the trained machine-learning model, the one or more sensory profiles to the target profile to identify at least one best sensory profile; and displaying, by the one or more processors, one or more graphical representations of the at least one best sensory profile on at least one user interface of at least one user device.

2. The computer-implemented method of claim 1 , the computer-implemented method comprising: creating, by the trained machine-learning model, the target profile, wherein the target profile includes a preferred sensory profile.

3. The computer-implemented method of claim 1 , wherein the computer- implemented method further comprises: analyzing, by the one or more processors, via the trained machine-learning model, the one or more ingredients of the one or more sensory profiles; and based on the analyzing, determining, by the one or more processors, via the trained machine-learning model, at least one optimized ingredient replacement.

4. The computer-implemented method of claim 1 , wherein the computer- implemented method further comprises: analyzing, by the one or more processors, via the trained machine-learning model, the one or more ingredient amounts of the one or more sensory profiles; andbased on the analyzing, determining, by the one or more processors, via the trained machine-learning model, at least one optimized ingredient amount.

5. The computer-implemented method of claim 1 , wherein the trained machinelearning model includes a generative adversarial network (GAN).

6. The computer-implemented method of claim 1 , wherein the comparing further comprises: determining, by the trained machine-learning model, that the at least one best sensory profile is a closest match to the target profile.

7. The computer-implemented method of claim 1 , wherein the one or more flavor characteristics includes at least one of: ethereal, fruity, sweet, buttery, creamy, sharp, banana, almond, cherry, floral, citrus, orange, green apple, cinnamyl, balsamic, spicy, woody, peach, berry, drying, or jammy.

8. The computer-implemented method of claim 1 , wherein the flavor data includes domain knowledge, one or more material descriptors, or one or more formula descriptors.

9. The computer-implemented method of claim 8, wherein the domain knowledge includes at least one ingredient interaction mechanism or at least one historical insight, wherein the at least one historical insight describes one or more previous attributes, a structure, or at least one ingredient characteristic.

10. The computer-implemented method of claim 8, wherein the one or more material descriptors include one or more chemical structures, one or more chemical interactions, one or more partition coefficients, or one or more water solubility and dissolution rates.11 . The computer-implemented method of claim 8, wherein the one or more formula descriptors include one or more qualitative formula characteristics or one or more quantitative formula characteristics.

12. The computer-implemented method of claim 1 , wherein the at least one best sensory profile meets an odor threshold.

13. A computer system for predicting a flavor blend for a food product, the computer system comprising: at least one memory storing instructions; and at least one processor configured to execute the instructions to perform operations comprising: receiving a target profile for a food product and flavor data, wherein the target profile includes one or more flavor characteristics, spectral data, or sensory data; predicting, via a trained machine-learning model, one or more sensory profiles based on the flavor data and the target profile, wherein the one or more sensory profiles include one or more flavor blends that include one or more ingredients and one or more ingredient amounts; comparing, via the trained machine-learning model, the one or more sensory profiles to the target profile to identify at least one best sensory profile; and displaying, one or more graphical representations of the at least one best sensory profile on at least one user interface of at least one user device.

14. The computer system of claim 13, the operations further comprising: creating, by the trained machine-learning model, the target profile, wherein the target profile includes a preferred sensory profile.

15. The computer system of claim 13, the operations further comprising: analyzing, via the trained machine-learning model, the one or more ingredients of the one or more sensory profiles; and based on the analyzing, determining, via the trained machine-learning model, at least one optimized ingredient replacement.

16. The computer system of claim 13, wherein the flavor data includes domain knowledge, one or more material descriptors, or one or more formula descriptors.

17. The computer system of claim 16, wherein the domain knowledge includes at least one ingredient interaction mechanism or at least one historical insight.

18. The computer system of claim 16, wherein the one or more material descriptors include one or more chemical structures, one or more chemical interactions, one or more partition coefficients, or one or more water solubility and dissolution rates.

19. A non-transitory computer-readable medium storing instructions that, when executed by at least one processor, cause the at least one processor to perform operations for predicting a flavor blend for a food product, the operations comprising: receiving a target profile for a food product and flavor data, wherein the target profile includes one or more flavor characteristics, spectral data, or sensory data; predicting, via a trained machine-learning model, one or more sensory profiles based on the flavor data and the target profile, wherein the one or more sensory profiles include one or more flavor blends that include one or more ingredients and one or more ingredient amounts; comparing, via the trained machine-learning model, the one or more sensory profiles to the target profile to identify at least one best sensory profile; and displaying one or more graphical representations of the at least one best sensory profile on at least one user interface of at least one user device.

20. The non-transitory computer-readable medium of claim 19, wherein the at least one best sensory profile meets an odor threshold.

Images