System and method for predicting product sensory shelf life progression using machine learning

Abstract

Systems and methods for predicting sensory shelf life progression are provided. The method may include receiving data associated with a product. The data may include one or more sensory attributes of the product. The method may also include providing the one or more sensory attributes of the product to a predictive machine learning model trained to identify associations between the one or more sensory attributes of the product and output a sensory shelf life progression of the product. The method may also include transmitting the sensory shelf life progression to a display of a user device.

Description

Cross Reference to Related Applications

[0001] This application claims priority to U.S. Provisional Application No. 63 / 466,090, filed May 12, 2023, U.S. Provisional Application No. 63 / 466,122, filed May 12, 2023, and U.S. Provisional Application No. 63 / 466,107, filed May 12, 2023, all of which are incorporated by reference herein. TECHNICAL FIELD

[0002] Various embodiments of the present disclosure generally relate to machine learning based techniques for predicting sensory panel results, and more particularly to systems and methods for predicting sensory shelf life progression using machine learning based regression models to approximate sensory attributes of a product. In some embodiments, the present disclosure relates to systems and methods for training machine learning based models to predict sensory shelf life progression. BACKGROUND

[0003] Current processes for evaluating food products (e.g., confectionery products) involve conducting sensory tests throughout the shelf life of the product. In some cases, a full study can take up to 52 weeks to decide which product to launch. In many cases, a product can be evaluated using a minimum of three time points (e.g., 4 weeks, 12 weeks, and 20 weeks) to determine whether to continue evaluating the product formulation for launch. Further, materials (e.g., finished goods, ingredients, packaging) can be placed in environmental test chambers to determine a respective shelf life by collecting sample data at specific time points (e.g., predetermined times) to confirm that the materials meet specific laboratory measurements to confirm how long the materials’ shelf life is. Specifically, food products and packaging can be tested for their shelf life by placing them within an environmental test chamber using environmental conditions (e.g., constant temperature and humidity). Depending on the materials being tested, it can remain in the test chamber for up to 52 weeks. Samples can be collected at predetermined intervals until enough data has been collected to confirm that the materials do not meet the constraints required for shelf life, or until the end of the time range to confirm the materials’ consistency has been reached. The time required to run the tests can have an impact on the speed of innovation (e.g., creation of new or improved products) and the speed of getting products to market.

[0004] 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 virtue of their inclusion in this section. SUMMARY

[0005] According to aspects of the present disclosure, systems and computer-implemented methods for predicting sensory shelf life progression of a product are disclosed. The systems and methods for predicting sensory shelf life progression can utilize a machine learning model. Specifically, the machine learning model can approximate sensory attributes of a product.

[0006] In one aspect, a computer-implemented method for predicting sensory shelf life progression can include receiving data associated with a product. The data can include one or more sensory attributes of the product. The method can also include providing the one or more sensory attributes of the product to a predictive machine learning model trained to identify associations between the one or more sensory attributes of the product and output a sensory shelf life progression of the product. The method can also include transmitting the sensory shelf life progression to a display of a user device.

[0007] In another aspect, a system for predicting sensory shelf life progression can include a memory storing instructions and a processor configured to execute the instructions to perform operations. The operations can include receiving data associated with a product. The data can include one or more sensory attributes of the product. The operations can also include providing the one or more sensory attributes of the product to a predictive machine learning model trained to identify associations between the one or more sensory attributes of the product and output a sensory shelf life progression of the product. The operations can also include transmitting the sensory shelf life progression to a display of a user device.

[0008] In another aspect, a non-transitory machine-readable medium storing instructions that, when executed by one or more processors, cause a computing system to perform a method for predicting sensory shelf life progression. The method can include receiving data associated with a product. The data can include one or more sensory attributes of the product. The method can also include providing the one or more sensory attributes of the product to a predictive machine learning model trained to identify associations between the one or more sensory attributes of the product and output a sensory shelf life progression of the product. The method can also include transmitting the sensory shelf life progression to a display of a user device.

[0009] Additional objects and advantages of the disclosed embodiments will be set forth in part in the description that follows, and in part will be obvious from the description, or can be learned by practice of the disclosed embodiments. The objects and advantages of the disclosed embodiments will be realized and attained by means of the elements and combinations particularly pointed out in the appended claims.

[0010] 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 DRAWINGS

[0011] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate various example implementations and together with the description, explain the principles of the disclosed implementations.

[0012] Figure 1 An example environment that can be used in connection with the technology presented herein is depicted in accordance with one or more implementations.

[0013] Figure 2 An example distribution of product attributes assessed by a sensory panel is depicted.

[0014] Figure 3 An example plot illustrating the predicted profile average response at 12 weeks versus the actual average response at 12 weeks derived from data collected at 4 weeks is depicted.

[0015] Figure 4 An example data flow diagram for testing the shelf life of a product using machine learning is depicted.

[0016] Figure 5A An example test condition for testing the shelf life of a product is depicted.

[0017] Figure 5B An example plot illustrating the environmental shelf life test versus the accelerated shelf life test is depicted.

[0018] Figure 6 An example data flow diagram for generating a recipe and a set of custom manufacturing requirements using machine learning is depicted.

[0019] Figure 7 An example processing environment for formulating a recipe and manufacturing requirements is depicted.

[0020] Figure 8 A flow diagram illustrating an example method for predicting sensory shelf life progression is depicted.

[0021] Figure 9 A flow diagram illustrating an example method for generating a recipe and a set of custom manufacturing requirements using machine learning is depicted.

[0022] Figure 10 An example computer system that can execute the technology presented herein is depicted. DETAILED DESCRIPTION

[0023] Various embodiments of the present disclosure generally relate to machine learning based techniques for predicting sensory shelf life progression of a product (e.g., a food product, a confectionery product, a material, a package, etc.). More specifically, various embodiments of the present disclosure relate to systems and methods for predicting sensory shelf life progression using machine learning based regression models to approximate sensory attributes of a product.

[0024] Sensory attributes of a food product (e.g., a chocolate confectionery), such as stickiness, caramel notes, chocolate notes, and roasted peanut, can impact the shelf life of the food product. Thus, being able to easily determine the progression of such attributes is critical for product development and testing. Existing methods for assessing or measuring sensory shelf life progression can require data to be collected at various time intervals over a long period of time.

[0025] Accordingly, to shorten the time required to evaluate a product and make determinations during such evaluations, one or more machine learning models can be used to estimate future sensory panel results for a product. In such implementations, the one or more machine learning models can use data collected from a first sensory evaluation (e.g., at 4 weeks) to predict sensory shelf life progression.

[0026] Embodiments of the present disclosure relate to addressing, mitigating, or correcting the above-referenced problems by determining a machine learning model configured to predict sensory shelf life progression. The machine learning model can use an initial data set collected at an initial time point. Systems and methods of the present disclosure can be used to predict sensory shelf life progression, particularly with respect to attributes of a food product, such as stickiness, caramel notes, chocolate notes, and roasted peanut. In examples, other such attributes associated with a food product can include chocolate notes, deep roast, alkaline, vanilla, caramel notes, complex fruit notes, citrus notes, wine notes, woody notes, nutty notes, cinnamon, sweetness, sourness, saltiness, bitterness, astringency, heat, crunch, friability, moistness, cohesiveness, stickiness, roughness, melt rate, roasted peanut, complex milky notes, steamed milky notes, fermented milky notes, skimmed milk powder, butter, springiness, chewy texture contrast, crunching sound, snap texture contrast, roasted almond, fruity notes, brown spice, density, steamed cereal notes, and toasted cereal.

[0027] Additionally, attributes associated with other consumable confectioneries, such as chewing gum, can include hardness, overall flavor, toughness, overall flavor, sweetness, bitterness, peppermint flavor, menthol flavor, creamy flavor, mouth cooling, overall heat, juiciness, smoothness, stickiness, springiness, crunching sound, spearmint flavor, thymol flavor, pepper-like pungency, vanilla flavor, nasal cooling, mouth heat, throat heat, tongue heat, astringency, and springiness.

[0028] Further, attributes associated with food, materials, packaging, etc. can include moisture, water activity, preservatives, surface microorganisms, pH, chemical migration (e.g., from packaging to food), texture properties, protein or lipid oxidation markers (e.g., ketones, peroxide values, aldehydes, fatty acids, etc.), vitamin depletion, flavor and aroma, sensory evaluation (e.g., visual, taste, feel), etc.

[0029] In various implementations, the machine learning model of the present disclosure can include a regression model. In various implementations, any suitable regression model can be utilized, including but not limited to linear, stepwise linear, partial least squares, principal components, support vector (SVR), neural networks, multivariate adaptive, multivariate linear, random forest, polynomial, generalized additive, Bayesian Additive Regression Tree (BART), Classification and Regression Tree (CART), or neural network models such as Multilayer Perceptron (MLP) and Recurrent Neural Network (RNN), Convolutional Neural Network (CNN). Based on the prediction, a user can determine whether a product is suitable based on the predicted shelf life of the product.

[0030] Although the models and implementations described herein relate to food and materials, the models of the present disclosure are applicable to a variety of products that can degrade over time (e.g., pharmaceuticals, biological materials, chemicals, etc.).

[0031] The terminology used below can be interpreted in its broadest reasonable manner, even if it is used in connection with a detailed description of certain specific examples of the disclosure. In fact, certain terms can even be emphasized below; however, any term that is intended to be interpreted in any limiting manner will be disclosed and specifically defined in the DETAILED DESCRIPTION section. Both the general description above and the detailed description below are exemplary and explanatory only and are not restrictive of the claimed features.

[0032] In the detailed description of the application herein, references to“an implementation,”“one implementation,”“one non-limiting implementation,”“in various implementations,” etc. indicate that the described implementation can include a particular feature, structure, or characteristic, but every implementation need not necessarily include the particular feature, structure, or characteristic. Further, such phrases are not necessarily referring to the same implementation. Further, where a particular feature, structure, or characteristic is described in connection with an implementation, it is submitted that it is within the knowledge of those in the art to effect such feature, structure, or characteristic in connection with other implementations whether or not explicitly described. After reading this specification, skilled artisans will be able to employ the disclosure in variety of circumstances some of which are not described herein.

[0033] Generally, the terminology can be understood at least in part from usage of the terms in the context in which they are used. For example, the terms such as "and", "or", and "and / or" as used herein can include a variety of meanings that can 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, at least in part, depending on the context, can be taken to mean any feature, structure or characteristic of a singular implementation or can be taken to mean a combination of features, structures or characteristics of more than one implementation. Similarly, terms such as "one", "a", or "the" again can be understood to convey a singular usage or to convey a plural usage, at least in part, depending on the context in which such terms are used. In addition, the term "based on" can be understood as not necessarily requiring explicit

[0034] 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 elements does not include only those elements but can 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, unless the context dictates otherwise, the singular forms "a", "an", and "the" include plural referents. Relative terms such as "about", "substantially", and "approximately" mean nearly the same as a reference number or value, and should be understood to encompass variations of ±5% of the specified amount or value.

[0035] As used herein, a "machine learning model" generally encompasses instructions, data, and / or models configured to receive an input and apply one or more of a weight, a bias, a classification, or an analysis to the input to generate an output. The output can include, for example, a classification of the input, an analysis based on the input, a design, a process, a prediction, or a recommendation associated with the input, or any other suitable type of output. Machine learning models are typically trained using training data (e.g., empirical data and / or input data samples) that is fed into the model in order to establish, tune, or modify one or more aspects of the model, such as weights, biases, criteria for forming a classification or clustering, etc. Aspects of a machine learning model can operate on an input linearly, in parallel, via a network (e.g., a neural network), or via any suitable configuration.

[0036] Execution of the machine learning model can include deploying one or more machine learning techniques, such as linear regression, logistic regression, random forest, gradient boosting machine (GBM), deep learning, deep neural networks, etc. Supervised and / or unsupervised training can be employed. For example, supervised learning can include providing training data and labels corresponding to the training data (e.g., as ground truth). Unsupervised methods can include clustering, classification, etc. Any suitable type of training can be used, e.g., stochastic, gradient boosting, random seed, recursive, epoch or batch based, etc.

[0037] Certain non-limiting embodiments are described below with reference to block diagrams and operational examples of methods, processes, apparatus, and devices. It should be understood that each block of the block diagrams and operational examples, and combinations of blocks in the block diagrams and operational examples, can be implemented by analog hardware or digital hardware and computer program instructions. Such computer program instructions can be provided to a processor of general purpose computer, special purpose computer, ASIC, or other programmable data processing apparatus to produce a machine, 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 examples. In some alternative implementations, the functions / acts noted in the blocks can occur out of the order noted in the operational examples. 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.

[0038] Referring now to the drawings, Figure 1 An example environment 100 that can be used with the technology presented herein is depicted. User devices 110, sensory shelf life progression prediction platform 120, and data storage devices 140 can communicate across network 150. User devices 110 can be associated with a user (e.g., a user associated with one or more of: generating, training, tuning, or using a machine learning model for predicting sensory shelf life progression). A user can use a user device 110 to interact with a user interface module 129 of the sensory shelf life progression prediction platform 120.

[0039] Sensory shelf life progression prediction platform 120 can be a platform having a plurality of interconnected components. Sensory shelf life progression prediction platform 120 can include one or more servers, smart networked devices, computing devices, components, and corresponding software for determining sensory attributes, analyzing relationships, determining machine learning models configured to predict sensory shelf life progression, and making sensory shelf life predictions using trained machine learning models. Sensory shelf life progression prediction platform 120 includes data collection module 121, data processing module 123, training module 125, machine learning models 127, and user interface module 129.

[0040] The data collection module 121 can receive input from the user device 110. In some examples, the input can include data associated with a product. In examples, the product can be a food product, such as a chocolate or a gum. In other examples, the product can be a material, such as a packaging, among others. The data associated with the product can include analytical data obtained from analytical testing of the product and sensory attributes of the product, such as stickiness, caramel flavor, chocolate flavor, roasted peanut, moisture, water content, among others. In certain examples, a user can have the option to input data associated with a product for which they desire to predict the progression of its sensory shelf life via the user device 110, and the data collection module 121 can receive the input. Additionally or alternatively, the data collection module 121 can receive (e.g., retrieve) data associated with the product from the data storage device 140. In further implementations, the data collection module 121 can receive data associated with the product from the testing device 160 (e.g., a test chamber).

[0041] The data processing module 123 can process data collected by the data collection module 121. Processing the collected data can include determining average values of sensory attributes, including actual and predicted shelf life progression. In examples, actual shelf life progression can be inputted into a machine learning model to train the machine learning model to predict shelf life progression of a product. In some examples, the data processing module 123 can access data associated with a product from a database or another information source (e.g., the data storage device 140). Further details are provided below with reference to Figure 2 Further details are provided regarding the computation of the predicted profile response.

[0042] In some implementations, the data processing module 123 generates a data structure 130. The data structure 130 can include each sensory attribute of a product. In at least one example, the data structure 130 can be a table.

[0043] The training module 125 can provide learning or training to the machine learning model 127 by providing training data (e.g., from data from other modules containing inputs (e.g., features) and correct outputs (e.g., labels)) to allow the machine learning model 127 to learn over time. For example, the training module 125 can receive data from the data structure 130 generated by the data processing module 123. When input is fed into the machine learning model 127, training can be performed based on a deviation of the processing results from the recorded results (e.g., an algorithm measures its accuracy through a loss function and makes adjustments until errors are sufficiently minimized). The training module 125 can train in any suitable manner (e.g., in batches) and can include any suitable training method. Training can be performed periodically and / or continuously, such as in real-time or near real-time. Further details are provided below regarding training a machine learning model.

[0044] The machine learning model 127 can receive training data from the training module 125 to learn the relationship between sensory attributes and shelf life progression. The ordering of the training data can be randomized during training. The machine learning model 127 can visualize the training data to identify correlations between different variables and to identify any data imbalance. The training data can be split into two parts, with one part used to train the model and the other part used to validate the trained model, de-duplicate, normalize, correct errors in the training data, etc. In some examples, the machine learning model 127 can receive data directly from the data structure 130. The machine learning model 127 can implement various machine learning techniques discussed in the present disclosure (e.g., random forest, k-nearest neighbors, partial least squares regression, principal component regression, etc.).

[0045] The user interface module 129 can enable presentation of a graphical user interface (GUI) in the user device 110. The user interface module 129 can include various interfaces, e.g., interfaces for data input and output devices (referred to as I / O devices, storage devices, etc.).

[0046] The data storage 140 can store and manage data associated with products. The data storage 140 can store the data structure 130 generated by the data processing module 123. The data storage 140 can also store any information provided by a user via the user device 110 and / or the testing device 160. Additionally, the data storage 140 can store the data structure 130 generated by the sensory shelf life progression prediction platform 120. In some examples, the data storage 140 can include a machine learning based training database with predefined mappings defining relationships between various input parameters and output parameters based on various statistical methods. The training database can include machine learning algorithms to learn the mapping between sensory attributes and shelf life progression. In some examples herein, the training database is periodically updated and / or supplemented based on machine learning methods.

[0047] Figure 2A distribution of attributes 200 of a food product (e.g., chocolate confection) assessed by a sensory panel is depicted. As illustrated, a distribution of actual sensory profiles versus predicted sensory profiles is compared at 12 weeks. A set of predicted profile responses 202 (12 weeks vs. 4 weeks) is illustrated. A set of control results 204 shows sensory shelf life progression of the product based on four sensory attributes (adhesion 206a, caramel flavor 206b, chocolate flavor 206c, and roasted peanut 206d). A set of test results 208 also shows sensory shelf life progression of the product based on the four sensory attributes. As illustrated, actual sensory shelf life progression 210 is compared to predicted sensory shelf life progression 212. In various implementations, actual sensory shelf life progression 210 is determined by measuring sensory attributes 206a-206d at various time intervals (e.g., 2 weeks, 4 weeks, 6 weeks, 12 weeks, etc.). In contrast, the present systems and methods enable predicted sensory shelf life progression 212 to be determined by measuring sensory attributes at only one initial time (e.g., 2 weeks, 4 weeks, 6 weeks, etc.).

[0048] Machine learning models (such as the machine learning model utilized by the machine learning module 127 as depicted in FIG. 1) can relate sensory attributes 206a-206d to sensory shelf life progression. For example, a machine learning model can be trained to determine a correlation between sensory attributes and sensory shelf life progression. Once determined, the machine learning model 208 can be used to predict sensory shelf life progression of a product whose sensory attributes have not been measured more than once at an initial stage. Figure 1

[0049] Figure 3 A plot 300 illustrating predicted profile average responses at 12 weeks versus actual average responses at 12 weeks derived from data collected at 4 weeks is depicted. A set of average values for attributes 302 is illustrated. A set of control results 304 shows sensory shelf life progression of the product based on four sensory attributes (adhesion 306a, caramel flavor 306b, chocolate flavor 306c, and roasted peanut 306d). A set of test results 308 also shows sensory shelf life progression of the product based on the four sensory attributes. As illustrated, actual sensory shelf life progression 310 is compared to predicted sensory shelf life progression 312. In various implementations, actual sensory shelf life progression 310 is determined by measuring sensory attributes 306a-306d at various time intervals (e.g., 2 weeks, 4 weeks, 6 weeks, 12 weeks, etc.). In contrast, the present systems and methods enable predicted sensory shelf life progression 312 to be determined by measuring sensory attributes at only one initial time (e.g., 2 weeks, 4 weeks, 6 weeks, etc.).

[0050] Machine learning models (such as the machine learning model utilized by the machine learning module 127 as depicted in FIG. 1) can relate sensory attributes 306a-306d to sensory shelf life progression. For example, a machine learning model can be trained to determine a correlation between sensory attributes and sensory shelf life progression. Once determined, the machine learning model 308 can be used to predict sensory shelf life progression of a product whose sensory attributes have not been measured more than once at an initial stage. Figure 1 ​The machine learning model utilized by the machine learning module 127 depicted in the middle can correlate sensory attributes 306a-d with sensory shelf life progression. For example, the machine learning model can be trained to determine a correlation between sensory attributes and sensory shelf life progression. Once determined, the machine learning model 308 can be used to predict the sensory shelf life progression of a product whose sensory attributes have not been measured more than once at the initial stage.

[0051] Figure 4 An example data flow diagram for testing shelf life of a product using machine learning is depicted. As illustrated, a material 402 (e.g., a product) can be placed in a shelf life test chamber 404. Settings of the shelf life test chamber 404 can be adjusted 406 based on the material 402. A sample of the material 402 can be extracted 408 from the shelf life test chamber at a predetermined time. The sample results and the predetermined time point can be provided to a machine learning model to predict 410 a mean time to failure (MTTF). In an example, the MTTF can be used to measure the average time before a coating fails and the contents of the package are exposed. Factors such as coating thickness, environmental conditions, and manufacturing defects affect the MTTF of a packaging material. The predicted MTTF can be analyzed to determine whether the material 402 will meet a threshold of a critical shelf life specification 412. If it is determined that the threshold of certainty is met, the test can stop 414. If it is determined that the threshold of certainty is not met, the test can continue to extract another sample of the material 402 at another predetermined time point.

[0052] Figure 5A An example test condition for testing shelf life of a product is depicted. In various implementations, accelerated shelf life conditions generally require higher heat and humidity to affect the product. Thus, setting the shelf life test chamber to these conditions will accelerate degradation, but can not shorten the shelf life. By cycling the shelf life conditions in the shelf life test chamber from high temperature high humidity 502 to low temperature low humidity 504 (e.g., from ambient to accelerated), the degradation curve of the product can be shortened as one adjustment cycle can equal one predetermined time point, as shown. Figure 5B

[0053] Figure 6 An example data flow diagram for generating a recipe and a set of custom manufacturing requirements using machine learning is depicted. As illustrated, data associated with a product (such as sensory attributes 602, consumer preferences 604, existing recipes 606, existing ingredients 608, and existing manufacturing processes / requirements 610) can be collected and input into one or more machine learning models 612. The one or more machine learning models 612 can then determine and output a formulated recipe 614, formulated ingredients 616, and formulated processes 618. In an example, the formulated recipe and manufacturing requirements (e.g., formulated processes 618) are predicted to meet consumer preferences.

[0054] ​Machine learning models (such as those derived from...) Figure 1 The machine learning module 127 described herein utilizes a machine learning model that can correlate product-related data with one or more machine learning models. For example, a machine learning model can be trained to determine the correlation between sensory attribute 602 and consumer preference 604. Once determined, the machine learning model 612 can be used to formulate predicted (e.g., new) product recipes and manufacturing requirements.

[0055] Figure 7 An exemplary processing environment for formulating recipes and manufacturing requirements is depicted. As illustrated, data (e.g., sensory data, ingredient data, processing / manufacturing data) can be received from multiple data sources 702. Data can be aggregated in a temporary storage layer 704 and processed in a processing layer 706. A consumption layer 708 can serve as an intermediary between the processing layer 706 and a web application 710. The web application 710 may include an application programming interface (API) layer 712 and a front-end 714 (e.g., a user interface). In various specific implementations, the processing environment 700 may... Figure 1 Implemented in one or more modules (such as data processing module 123 and / or user interface module 129) depicted in the text.

[0056] Figure 8 A flowchart illustrating an exemplary method 800 for predicting the progression of sensory shelf life according to this disclosure is depicted. Method 800 may be derived from... Figure 1 The sensory shelf life progression prediction platform 120 is implemented. Specifically, machine learning models can be used to predict the sensory shelf life progression of a product.

[0057] In step 805, product-related data may be received. In an example, the product may be a food, such as chocolate candy or other food items. The data may include one or more sensory attributes. In an example, one or more sensory attributes may include one or more of adhesiveness, caramel flavor, chocolate flavor, and roasted peanuts. As described in this system and method, product-related data may be collected at a single specific time point (e.g., 2 weeks, 4 weeks, 6 weeks, etc.). In some examples, a user may input product-related data into the user interface module 129 of the sensory shelf-life progression prediction platform 120 via user device 110. In other examples, product-related data may be retrieved from data storage device 140. The data may identify the sensory attributes of the product. The data may be received by data collection module 121.

[0058] In step 810, data associated with the product can be provided to a machine learning model. In step 815, the trained machine learning model can be used to determine the sensory shelf life progression of the product based on the sensory attributes received in step 805. The trained machine learning model used in step 810 (e.g., machine learning model 127) can be trained in any suitable manner. The trained machine learning model can use one or more regression models to determine the sensory shelf life progression. In examples, such regression models can include linear, stepwise linear, partial least squares, principal components, support vector (SVR), neural networks, multivariate adaptive, multivariate linear, random forest, polynomial, generalized additive, Bayesian Additive Regression Tree (BART), Classification and Regression Tree (CART), or neural network models such as Multilayer Perceptron (MLP) and Recurrent Neural Network (RNN), Convolutional Neural Network (CNN). In various implementations, the prediction for the sensory shelf life progression can be displayed on a graphical user interface (GUI) of a user device (such as user device 110) depicted in FIG. 1. In step 815, the sensory shelf life progression can be transmitted to a display of the user device. Figure 1

[0059] Figure 9 A flowchart illustrating an example method for generating a recipe and a set of custom manufacturing requirements using machine learning is depicted. Method 900 can be performed by sensory shelf life progression prediction platform 120 of FIG. 1. Figure 1 Specifically, a machine learning model can be used to generate a custom recipe and a set of custom manufacturing requirements for each cluster of consumer preferences in a set or plurality of clusters of consumer preferences.

[0060] In step 905, data associated with a product can be received. In examples, the product can be a food product, such as a chocolate confection or other food product. The data can include one or more sensory attributes and consumer preferences. In examples, the one or more sensory attributes can include one or more of a tactile, olfactory, gustatory, etc. In some examples, a user can input the data associated with the product to a user interface module 129 of product formulation platform 120 via a user device 110. In other examples, the data associated with the product can be retrieved from data storage 140. The data can identify the sensory attributes and consumer preferences of the product. The data can be received by data collection module 121.

[0061] ​At step 910, the one or more sensory attributes and the one or more sets of consumer preference data can be provided to a deterministic machine learning model trained to identify associations between the one or more sensory attributes and the one or more sets of consumer preference data and output a target ideal sensory profile. At step 915, the target ideal sensory profile can be provided to a generative machine learning model trained to identify associations within the target ideal sensory profile and output a recipe and a set of manufacturing requirements. In an example, the recipe can include an exact list of ingredients to manufacture the consumable confectionery. In other examples, the manufacturing requirements can include an exact list of ordered manufacturing steps and processing conditions for manufacturing the consumable confectionery.

[0062] At step 920, the deterministic machine learning model can separate the plurality of consumer preference clusters into one or more sets of consumer preference clusters based on the one or more sensory attributes of the product and the one or more sets of consumer preference data. At step 925, the deterministic machine learning model can generate a customized recipe and a set of customized manufacturing requirements for each of the one or more sets of consumer preference clusters.

[0063] The trained machine learning models used in steps 910, 915, and 920 (e.g., machine learning model 127) can be trained in any suitable manner. The trained machine learning models can use one or more regression models to ultimately formulate the recipe and the manufacturing requirements. In an example, such regression models can include linear, stepwise linear, partial least squares, principal components, support vector (SVR), neural networks, multivariate adaptive, multivariate linear, random forest, polynomial, generalized additive, Bayesian additive regression trees (BART), classification and regression trees (CART), or neural network models such as multilayer perceptron (MLP) and recurrent neural networks (RNN), convolutional neural networks (CNN). In various implementations, the formulated recipe and the manufacturing requirements can be displayed on a graphical user interface (GUI) of a user device, such as user device 110 depicted in FIG. 1. Figure 1 The formulated recipe and the manufacturing requirements can be displayed on a graphical user interface (GUI) of a user device, such as user device 110 depicted in FIG. 1.

[0064] Figure 10 Embodiments of computer systems that perform the techniques presented herein are illustrated. The computer system 1000 can include a set of instructions that can be executed to cause the computer system 1000 to perform any one or more of the methods or computer-based functions disclosed herein. The computer system 1000 can operate as a standalone device or can be connected, e.g., using a network, to other computer systems or peripheral devices.

[0065] Unless specifically stated otherwise, as apparent from the following discussions, it is appreciated that throughout the specification discussions utilizing terms such as "processing," "computing," "calculating," "determining," "analyzing" or the like, can refer to the action and / or processes of a computer or computing system or similar electronic computing device, that manipulates and / or transforms data represented as physical, such as electronic, quantities within the computer's registers and / or memories into other data similarly represented as physical quantities within the computer's memories, registers or other such information storage, transmission or display devices.

[0066] In similar manner, the term "processor" can refer to any device or portion of a device that manipulates electronic data, such as digital data, using logic or other hardware that changes the form of the electronic data. A "computer," "computing machine," "computing platform," "computing device" or "server" can include one or more processors.

[0067] In networked deployments, computer system 1000 can operate in the capacity of a server or as a client user computer in server-client user network environments, or as a peer computer system in peer-to-peer (or distributed) network environments. Computer system 1000 can also be implemented as or incorporated into various devices, such as a personal computer (PC), a tablet PC, a set-top box (STB), a personal digital assistant (PDA), a mobile device, a palmtop computer, a laptop computer, a desktop computer, a communications device, a wireless telephone, a fixed telephone, a control system, a camera, a scanner, a facsimile machine, a printer, a pager, a personal trusted device, a web appliance, a network router, switch or bridge, or any other machine capable of executing a set of instructions (sequential or otherwise) that specify actions to be taken by that machine. In particular embodiments, computer system 1000 can be implemented using electronic devices that provide voice, video or data communication. Further, while a single computer system 1000 is illustrated, the term "system" shall also be taken to include any collection of systems or sub-systems that individually or jointly execute a set, or multiple sets, of instructions to perform one or more computer functions.

[0068] As Figure 10 As illustrated, computer system 1000 can include a processor 1002, such as central processing unit (CPU), a graphics processing unit (GPU), or both. Processor 1002 can be a component in a variety of systems. For example, processor 1002 can be part of a standard personal computer, a workstation, a server, a server farm, a personal digital assistant, a television, a medical device, a set-top box, a game console, an appliance, a network router, switch or bridge, or any other machine capable of executing a set of instructions (sequential or otherwise) that specify actions to be taken by that machine. Further, processor 1002 can be a component of an electronic system, such as a wearable electronic system, a mobile electronic system, or any other electronic system. Processor 1002 can be a component of a wearable electronic system, such as a watch, a bracelet, a ring, a pair of glasses, a pair of shoes, a pair of socks, a pair of pants, a shirt, a hat, a belt, a necklace, a pair of earrings, a pair of rings, a pair of shoes, a pair of gloves, a pair of boots, a pair of shorts, a pair of swim trunks, a pair of swim goggles, a pair of swim fins, a pair of swim caps, a pair of swim suits, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers, a pair of swim diapers

[0069] The computer system 1000 can include a memory 1004 that can communicate with the bus 1008. The memory 1004 can be a main memory, a static memory, or a dynamic memory. The memory 1004 can include, but is not limited to computer-readable storage media such as various types of volatile and non-volatile storage media, including but not limited to random access memory, read-only memory, programmable read-only memory, electrically programmable read-only memory, electrically erasable read-only memory, flash memory, magnetic tape or disk, optical media and the like. In one implementation, the memory 1004 includes a cache or random access memory for the processor 1002. In alternative implementations, the memory 1004 is a separate component from the processor 1002, such as a cache memory of a processor or a system memory or other memory. The memory 1004 can be an external storage device or database accessible by the processor 1002. Examples include a hard disk drive, a Compact Disc (“CD”), a Digital Video Disc (“DVD”), a memory stick, a floppy disk, a Universal Serial Bus (“USB”) memory device, or any other device operational with to store data. The memory 1004 is operable to store instructions executable by the processor 1002. The functions, acts or tasks illustrated in the figures or described herein can be performed by the programmed processor 1002 operating the instructions stored in the memory 1004. The functions, acts or tasks are independent of the particular type of instructions set, storage media, processor or processing strategy and can be performed by software, hardware, integrated circuits, firm ware, micro-code and the like, operating alone or in combination. Likewise, processing strategies can include multiprocessing, multitasking, parallel processing and the like.

[0070] As shown, the computer system 1000 can further include a display unit 1010, such as a liquid crystal display (“LCD”), an organic light emitting diode (“OLED”), a flat panel display, a solid state display, a cathode ray tube (“CRT”), a projector, a printer, or other now known or later developed display device for outputting determined information. The display 1010 can serve as an interface for a user to view the functions of the processor 1002, or specifically, to interface with software stored in the memory 1004 or drive unit 1006.

[0071] Additionally or alternatively, the computer system 1000 can include an input device 1012 configured to allow a user to interact with any of the components of system 1000. The input device 1012 can be a number keypad, a keyboard, or a cursor control device, such as a mouse, or a joystick, touch screen display, remote control, or any other device operative to interact with the computer system 1000.

[0072] The computer system 1000 can also or alternatively include a disk or optical drive unit 1006. The disk drive unit 1006 can include a computer-readable medium 1022 in which one or more sets of instructions 1024 (e.g., software) can be embedded. Further, the instructions 1024 can embody one or more of the methods or logic as described herein. The instructions 1024 can reside completely, though not necessarily entirely, within the memory 1004 and / or the processor 1002 during execution by the computer system 1000. The memory 1004 and the processor 1002 also can include the computer-readable medium, as discussed above.

[0073] In some systems, the computer-readable medium 1022 includes instructions 1024 or receives and executes instructions 1024 responsive to the propagation of a signal, such as a carrier wave or other propagated signal that is generated by a Figure 1 transmission of voice, video, audio, images, or any other data over the network 150 (e.g., as depicted in FIG. 1). Further, the instructions 1024 can be transmitted or received over the network 150 via the communication port or interface 1020 and / or using the bus 1008. The communication port or interface 1020 can be part of the processor 1002 or can be a separate component. The communication port 1020 can be created in software or can be a physical connection in hardware. The communication port 1020 can be configured to connect with a network 150, external media, the display 1010, or any other components in the system 1000, or combinations thereof. The connection with the network 150 can be a physical connection, such as a wired Ethernet connection, or can be established wirelessly as discussed below. Likewise, the additional connections with other components of the system 1000 can be physical or wireless connections. Alternatively, the network 150 can be directly connected to the bus 1008.

[0074] While the computer-readable medium 1022 is illustrated as a single medium, the term "computer-readable medium" can include a single medium or multiple media, such as a centralized or distributed database, and / or associated caches and servers that store one or more sets of instructions. The term "computer-readable medium" can also include any medium that is capable of storing, encoding or carrying a set of instructions for execution by a processor or that cause a computer system to perform any one or more of the methods or operations disclosed herein. The computer-readable medium 1022 can be non-transitory and tangible.

[0075] The computer-readable medium 1022 can include solid-state memory such as a memory card or other package that houses one or more non-volatile read-only memories. The computer-readable medium 1022 can be random access memory (RAM) or other volatile re-writable memory. Additionally, or alternatively, the computer-readable medium 1022 can include a magnetic, optical, or other form of storage that is external to the processor 1001. For example, the computer-readable medium 1022 can include a magnetic or optical disk that is read by disk drives or other storage devices. The digital files of email or other self-contained information files or file sets can be considered as distribution media of tangible storage media. Accordingly, the disclosure is considered to include any one or more of a computer-readable medium or distribution media and other equivalents and successor media in which data or instructions can be stored.

[0076] In alternative implementations, dedicated hardware implementations such as application specific integrated circuits, programmable logic arrays and other hardware devices can be constructed to implement any one or more of the methods described herein. Applications that can include the various implementations of the devices and systems can broadly include a variety of electronic and computer systems. One or more implementations described herein can implement functions by using two or more specific interconnected hardware modules or devices with related control and data signals that can be transmitted between and through the modules, or as portions of an application-specific integrated circuit. Accordingly, the present system encompasses software, firmware, and hardware implementations.

[0077] The computer system 1000 can be connected to one or more networks 150 (e.g., as depicted in FIG. 1). The network(s) 150 can define one or more networks including wired or wireless networks. The wireless network can be a cellular telephone network, an 802.11, 802.16, 802.20, or WiMax network. Further, such networks can include public networks such as the Internet, private networks such as an intranet, or combinations of them, and can utilize various network protocols now available or later developed including, but not limited to, TCP / IP based networking protocols. The network(s) 150 can include a wide area network (WAN), such as the Internet, a local area network (LAN), a campus area network, a metropolitan area network, a direct connection such as through a universal serial bus (USB) port, or any other network that can enable communication of data among computing devices. The network(s) 150 can be configured to couple one computing device to another computing device to Figure 1 communicate data between or among the devices. The network(s) 150 can generally be capable of employing any form of machine-readable media for communicating information from one device to another. The network(s) 150 can include methods by which information can be propagated between computing devices. The network(s) 150 can be divided into sub-networks. A sub-network can allow access to all other components connected to it, or the sub-network can limit access between components. The network(s) 150 can be considered public or private network connections and can include, for example, a virtual private network or encryption or other security mechanisms employed over a public Internet.

[0078] In accordance with various implementations of the present disclosure, the methods described herein can be implemented by software programs executable by a computer system. Further, in an exemplary, non-limited implementation, implementing the methods can include distributing processing, component / object distributed processing, and parallel processing. Alternatively, virtual computer system processing can be constructed to implement one or more of the methods or functionality as described herein.

[0079] Although this specification describes particular standards and protocols that can be implemented in particular implementations of components and functionality, the present disclosure is not limited to such standards and protocols. For example, standards for Internet and other packet-switched network transmissions (e.g., TCP / IP, UDP / IP, HTML, HTTP) represent examples of the state of the art. Such standards are periodically superseded by faster or more efficient equivalents having the same or similar functionality. Therefore, replacement standards and protocols having the same or similar functionalities are considered equivalents thereof.

[0080] It should be understood that in one embodiment, the steps of the methods discussed are performed by an appropriate processor (or processors) of a processing (i.e., computer) system executing instructions (computer readable code) stored in storage. It should also be understood that the disclosed embodiments are not limited to any particular implementation or programming technique and that the disclosed embodiments can be implemented using any appropriate techniques for implementing the functionality described herein. The disclosed embodiments are not limited to any particular programming language or operating system.

[0081] It should be understood that in the foregoing description of exemplary embodiments of the application, various features of the application are sometimes grouped together in a single embodiment, a figure, or a description thereof for the purposes of simplifying the disclosure and aiding in the understanding of one or more of the inventive aspects. However, this disclosure should not be interpreted as reflecting an intention that the application requires more features than are explicitly recited in each claim. On the contrary, as the claims below reflect, inventive aspects lie in fewer than all features of a single previously disclosed embodiment. Thus, any claims following the detailed description of specific embodiments are hereby expressly incorporated by reference into this detailed description, with each claim standing on its own as a separate embodiment of this application.

[0082] Furthermore, while some embodiments described herein include some, but not other, features of the embodiments described in other embodiments, it is contemplated that the features of the different embodiments can be combined with each other in any combination within the scope of the application and form different embodiments. For example, in the appended claims, any of the claimed embodiments can be used in any combination.

[0083] Thus, although certain embodiments have been described, a person of ordinary skill in the art will recognize that other and further modifications can be made thereto and it is expressly intended that all such modifications and variations are within the scope of the application. For example, functionality can be added or deleted from the diagramed embodiments, and operations can be interchanged among functional blocks. Steps can be added or deleted to the methods described.

[0084] 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 falling within the true spirit and scope of the 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 to the abo ve detailed description. While various embodiments of the present disclosure have been described above, it should be understood that they have been presented by way of example only, and not limitation. Numerous other embodiments will become apparent to those skilled in the art upon consideration of the foregoing description. Accordingly, the disclosure is not limited to any particular embodiment, but embraces all such embodiments that fall within the scope of the appended claims and their equivalents.

Claims

1. A computer-implemented method for predicting the progression of sensory shelf life, the method comprising: Data associated with the product is received by one or more processors, the data including one or more sensory attributes of the product; The one or more processors provide the one or more sensory attributes of the product to a predictive machine learning model, which is trained to identify the associations between the one or more sensory attributes of the product and output the sensory shelf life progression of the product. The sensory shelf-life progress is transmitted to the display of the user device by the one or more processors.

2. The computer-implemented method of claim 1, wherein the data associated with the product is collected at predetermined time points.

3. The computer-implemented method as described in claim 2, wherein the predetermined time point is four weeks.

4. The computer-implemented method of claim 1, wherein the one or more sensory properties include one or more of adhesiveness, caramel flavor, chocolate flavor, moisture, oxidation, vitamin loss, surface microorganisms, touch, smell, and taste.

5. The computer-implemented method of claim 1, wherein the product is an edible product.

6. The computer-implemented method of claim 1, further comprising: The one or more processors provide the one or more sensory attributes and one or more sets of consumer preference data to a deterministic machine learning model, which is trained to identify the association between the one or more sensory attributes and the one or more sets of consumer preference data and output a target ideal sensory profile. as well as The target ideal sensory profile is transmitted to the display of the user device by the one or more processors.

7. The computer-implemented method of claim 6, further comprising: The one or more processors provide the target ideal sensory profile to a generative machine learning model, which is trained to identify associations within the target ideal sensory profile and output a recipe and a set of manufacturing requirements. as well as The recipe and the set of manufacturing requirements are transmitted to the display of the user device by the one or more processors.

8. The computer-implemented method of claim 7, further comprising: The deterministic machine learning model uses one or more sensory attributes of the product and one or more sets of consumer preference data to separate multiple consumer preferences into one or more sets of consumer preference clusters. as well as The deterministic machine learning model generates a customized formula and a set of customized manufacturing requirements for each of the one or more consumer preference clusters.

9. A system for predicting the progression of sensory shelf life, the system comprising: The memory stores instructions; and A processor configured to execute the instructions to perform operations, the operations including: The processor receives data associated with the product, the data including one or more sensory attributes of the product; The processor provides one or more sensory attributes of the product to a predictive machine learning model, which is trained to identify the associations between the one or more sensory attributes of the product and output the sensory shelf life progress of the product. The processor transmits the sensory shelf-life progress to the display of the user device.

10. The system of claim 9, wherein the data associated with the product is collected at predetermined time points.

11. The system of claim 10, wherein the predetermined time point is four weeks.

12. The system of claim 9, wherein the one or more sensory properties include one or more of adhesiveness, caramel flavor, chocolate flavor, moisture, oxidation, vitamin loss, surface microorganisms, touch, smell, and taste.

13. The system of claim 9, wherein the product is an edible product.

14. The system of claim 9, wherein the operation further comprises: The processor provides one or more sensory attributes and one or more sets of consumer preference data to a deterministic machine learning model, which is trained to identify the association between the one or more sensory attributes and the one or more sets of consumer preference data and output a target ideal sensory profile. as well as The processor transmits the target ideal sensory profile to the display of the user device.

15. The system of claim 14, wherein the operation further comprises: The processor provides the target ideal sensory profile to a generative machine learning model, which is trained to identify associations within the target ideal sensory profile and output a recipe and a set of manufacturing requirements. as well as The processor transmits the recipe and the set of manufacturing requirements to the display of the user device.

16. The system of claim 15, wherein the operation further comprises: The deterministic machine learning model uses one or more sensory attributes of the product and one or more sets of consumer preference data to separate multiple consumer preferences into one or more sets of consumer preference clusters. as well as The deterministic machine learning model generates a customized formula and a set of customized manufacturing requirements for each of the one or more consumer preference clusters.

17. A non-transitory machine-readable medium storing instructions that, when executed by one or more processors, cause a computing system to perform a method for predicting the progression of sensory shelf life, the method comprising: Data associated with the product is received by one or more processors, the data including one or more sensory attributes of the product; The one or more processors provide the one or more sensory attributes of the product to a predictive machine learning model, which is trained to identify the associations between the one or more sensory attributes of the product and output the sensory shelf life progression of the product. The sensory shelf-life progress is transmitted to the display of the user device by the one or more processors.

18. The non-transitory machine-readable medium of claim 17, wherein the method further comprises: The one or more processors provide the one or more sensory attributes and one or more sets of consumer preference data to a deterministic machine learning model, which is trained to identify the association between the one or more sensory attributes and the one or more sets of consumer preference data and output a target ideal sensory profile. as well as The target ideal sensory profile is transmitted to the display of the user device by the one or more processors.

19. The non-transitory machine-readable medium of claim 18, wherein the method further comprises: The one or more processors provide the target ideal sensory profile to a generative machine learning model, which is trained to identify associations within the target ideal sensory profile and output a recipe and a set of manufacturing requirements. as well as The recipe and the set of manufacturing requirements are transmitted to the display of the user device by the one or more processors.

20. The non-transitory machine-readable medium of claim 19, wherein the method further comprises: The deterministic machine learning model uses one or more sensory attributes of the product and one or more sets of consumer preference data to separate multiple consumer preferences into one or more sets of consumer preference clusters. as well as The deterministic machine learning model generates a customized formula and a set of customized manufacturing requirements for each of the one or more consumer preference clusters.

Images