GastroNexus

All up in your flavor space
Successful New Product Development with Genetic Algorithms for Predictive Sensory

Predicting how consumers, from a variety of backgrounds and demographics, will respond to a new product is hard. Most food and beverage producers today use a mix of qualitative and quantitative consumer surveys to determine if a product is “good enough” to launch, only to find out later that good enough doesn’t make for a great product development strategy.

The problem is that data generated from qualitative and quantitative consumer surveys isn’t predictive – it’s only explanatory. Statistical analysis by an innovation team or product development team can show if the consumers included in the survey liked a product or even how much they liked a product, but that data is only useful if those consumers are a stratified and representative sample of every cohort in your target consumer group. Which is unlikely, so the data collected most likely will not generalize well.

White demographic beer panel

So after identifying which flavors are going to trend into long-term brand-making preferences with Deep Market Insights, how do you cut down on the iterative cycles and expensive consumer tests and use predictive data to create winning new products?

Background: The Problem

Gastrograph AI is the first AI platform for the food and beverage industry. The AI models the sensory perception of every consumer demographic across age, sex, race, socioeconomic status, and smoking habits, to predict consumer preference today and the evolution of preference into the future. Gastrograph AI uses hundreds of self-learning algorithms to various attributes of incoming reviews. Unfortunately, many of the highest performing algorithms are uninterpretable – we can prove that they work, but can’t see into the model to know why it works.

Thus, with models such as neural nets, autoencoders, and support vector machines, it is increasingly difficult to look inside and see which group of flavor attributes or somatosensations contributes to achieving a high perceived quality for a given style of product. Yet this is precisely what food and beverage producers care about: what flavor profile should I aim for in order to achieve a high perceived quality score among the general population?

How can the team at Analytical Flavor System help our clients solve this problem, while still using the highest accuracy models?

We could try to input every possible flavor profile, for every possible class and style of product, into an algorithm to predict perceived quality and consumer preference. If that worked, we could just choose the highest performing flavor profiles in each category. But there are over 4.7 x 1018 possible flavor profiles per class of product on the Gastrograph System, and that does not include specific reference flavors or account for demographic and geolocation biases.

Using brute force and trying out every flavor profile is not only inefficient, but also improbable: this would have to be rerun for every flavor-profile with every update of the model across every tasting demographic and preference archetype. Now, the problem becomes: what is a way of efficiently searching for the best flavor profile of a given style of product?

The Solution: Genetic Algorithms

Genetic algorithms are a way of efficiently searching through any range of values. They are the result of data scientists and mathematicians trying to model how evolution results in a population better fit to survive in a given environment. In biological evolution, when DNA from the parental generation makes the genetic information for their progeny, DNA crossover, mutation, and recombination to form the progeny. The best fit offspring in turn survive and reproduce, to produce offspring even better fit for the environment. In this way, the DNA of the population changes and converges over time to produce the best fit individuals for any given environment.

Genetic algorithms mimic this process. The algorithm first generates a random population of values for some given parameters. In subsequent 'generations', the algorithm chooses the high performing units of the population, chooses the parameter values of the subsequent child unit from either of the parental units, mutates the values with some low percentage, and evaluates the generated population with a fitness function to determine how fit it is to survive in this environment. The highest performing units are chosen to survive to the next generation, and the rest are either terminated or bred again, depending once again on how well they score according to the fitness function. After multiple iterations of this process, the population converges to a set of observations that result in a high value of fitness.

Genetic algorithms allow us to generate theoretical flavor profiles and test them against a prediction of consumer preference learned through reviews of existing products.

If we start with the flavor profile of the first iteration of a product under development, we can ask the AI to optimize it for a target demographic – say, millennial female new-premium consumers. The AI sets off mutating the product around inferred axis of variance (keeping clear of flavor combinations which can’t be produced) and tests each of its new hypothetical flavor profiles against the predicted preferences of the target consumer cohort. The AI can spawn hundreds of mutations at every generation, and learns which combination of changes has the greatest effect on preference. It then culls the worst performing hypothetical flavor profiles (like a flavor eugenics program!) and mutates a new generation of flavor profiles from the best of the last generation. After a few generations, the population converges to a set of flavor profiles that lead to the highest perceived quality.

Starting Flavor ProfileFigure 1. Starting Flavor Profile


Generation 2Figure 2. Generation 2
Generation 3Figure 3. Generation 3

This results in a few possible target flavor profiles for the product under development – with highly accurate predictions for its success and distribution of consumer assessment across demographics.

Predicted Market Preference Graph
Figure 4. Predicted Market Preference

Applications

Using Gastrograph AI’s genetic algorithms, your innovation and product development teams can create better, more successful, products in fewer iterative cycles using the power of data.

Since all of these parameters can be controlled, we can help producers that want to create specific categories or styles (for instance, an IPA) for their target consumers (i.e. female millennials).

You and your team can also limit the possible flavor profile values however they want. For example, if a brewery wants to make an IPA with a low level of herbaceous flavor (or aroma), they could specify that, and the AI will find the optimal flavor profiles with low herbaceous perception. Our clients can put as many restrictions on the output flavor profile as they desire. Linking the output flavor profile with the production process needed to achieve it provides an end-to-end new product development application, from conception to actualization.

Read more
NVIDIA podcast - Better Beer Through AI

Last week, we were featured on NVIDIA's The AI Podcast. Listen to our CEO, Jason Cohen, discuss the future of personalized products in the beer industry!

A huge thank you to NVIDIA for having us!

Read more
We're hiring domain experts, project managers, and R&D!

Analytical Flavor Systems is hiring a Domain SME, On-Site Technical Project Manager, and R&D Team.

Intern & Full-time Interview Project

Overview:

Gastrograph AI is an artificial intelligence platform for the food and beverage industry. Our technology enables food and beverage producers to update, formulate, and optimize new and existing products to meet changing consumer preferences in demographics around the world.

Roles:

  1. Domain SME: Our Domain Subject Matter Experts are production experts, usually hired from the industries in which AFS has clients. (i.e. a Brew Master for beer or Floor Manager for chocolate production)

  2. On-Site Technical Project Manager: Our on-site team integrates with our clients’ production, R&D, and/or quality staff to successfully implement our technology into their day-to-day operations. Individuals on the on-site team may be assigned to a rotation, an industry, a client, or a specific production center depending on their expertise and the scope of our work.

  3. R&D Team: Our R&D team works across a variety of functions to test, train, and improve our platform, technology, and domain expertise. The R&D team is responsible for the training and updating of our vertical process maps, the operations of our Food and Beverage Production R&D Lab, and creating the baseline datasets and implementation plans for the company to enter new verticals within the food and beverage industry.
    Interns, by nature of the work involved, are only eligible for the R&D Team.

Goals of this interview:

AFS uses a standardized technical project to gauge every candidate’s fit for a position on our team. As the same technical interview is used for interns and full-time applicants across multiple roles, we adjust our expectations and rubric accordingly. Our primary goal is to gauge your ability to understand the implementation, application, and value of our technology from conversations or publicly available information – and to think critically on how to build, apply, and improve our platform for our clients.

The Interview:

Analytical Flavor Systems frequently enters new verticals of food and beverage production, requiring the creation of a new expert-constructed process map detailing the production pathway of that verticals products.

You will be assigned a product vertical AFS is already active in. Your goal is to:

1) Map the process for producing your assigned product. Encode the knowledge of your production process into a computer readable format. For example, in the production of tea, all tea is first picked, withered, and then further processed. Include metrics, opportunities for data collection, and points of control accompanying each production step. For example, in the case of tea withering: time, temperature, ambient humidity, etc.

2) Select a single production step of interest:

a. Indicate the options for parameters at that production step and note how those options will change the resultant flavor profile

b. Extend the production map at the selected production step to include possible issues that can occur (in the example of tea withering, rainy weather after harvesting can make it difficult for the tea to properly lose ~80% of it’s moisture content) and the possible flaws that can occur such as specific chemical compounds or other sensory or physical defects (to continue this example, tea processed on a rainy day, often has to be heated more during withering to aide in the removal of excess moisture, which can sometimes give the tea a smoky taste)

c. Show how those issues and flaws can be predicted at any step earlier in the process (in the example of tea withering, predictive metrics such as weather reports and trends could have been used to optimize the harvest window).

d. Assuming that the algorithm we create to predict the issue or flaw has failed, or that a specific flaw or issue could not have been predicted, and that flaw has formed – show how that issue or flaw can be identified in real time (which combination of variables would be indicative of its presence?)

e. Further extend the production map at your production step of interest to include a mitigation strategy to negate the occurrence of the issue or flaw when predicted beforehand (preemptive mitigation).

f. Finally, extend the production map one last time to include a mitigation strategy to negate the effect of the issue or flaw when identified as present (reactionary mitigation).

We use graphical databases for our encoding in production environments.You may use excel or any other format that fits your idea for an optimally complete interview project. It’s definitely not required, but if you wanted to do something super impressive, do this in a graphical database (just not Neo4J).

The deliverables:

1) The process map and all extensions

a. Instructions to run or view the process map if hosted in a database

2) Any list of assumptions you make while developing this process map

a. This includes things such as limitations on scope, or the purposeful exclusion of a sub-class within the product vertical that follows a different production pathway

3) A brief write-up in PDF format on how you would test, update, and apply your production map to optimize your selected process in a client’s production environment.

4) Should you pass the technical interview on steps 1 – 3, we will fly/conjure/taxi you to our office to meet the team, and present your findings to both data scientists and chemistry team members.

Deadlines:

There is no deadline for full-time applications. We accept submissions on a rolling basis, and are almost always hiring for our growing team. Summer interns should apply by the last day in April.

Cool things and fun ideas:

There are no rules - feel free (or only slightly constrained) to expand or modify this interview to fit your special skill set. If it showcases your skills and we find it impressive, it would greatly benefit your application.

Ideas that we like:

An interactive query app to see the process options and pathway changes, a math-heavy version of the process control system, or predicted flavor outcomes – use one of these or come up with your own.

Collaboration and communication:

At AFS, entering a new vertical is a team effort. The AFS team is happy to collaborate with, answer questions, and point you in the right direction. Feel free to hit us up.

Please email this interview to:

JasonCEO@gastrograph.com

Best of luck! - Jason (CEO) & Ryan (R&D Manager)

Read more
Analytical Flavor Systems is hiring a Full-Stack Engineer!

Analytical Flavor Systems | Manhattan - NYC | Full-Time | Onsite |

Our Services: Analytical Flavor Systems uses machine learning and artificial intelligence to build tools for the food & beverage industry. Our Quality, Process, and Market Intelligence services create real-time predictive decisions metrics at each stage of a products life-cycle. We leverage our predictive models across products & industries for flavor profile optimization, production process optimization, demographic targeting & cognitive marketing - helping companies create and sell the best product to their highest value consumers with every batch.

-Quality Intelligence: Real-time predictive quality control, assurance, and improvement from human sensory data.
-Process Intelligence: Real-time predictive process control and optimization from human sensory data + manufacturing & LIMS data.
-Market Intelligence: Linking flavor-profile, demographics, and sales data to find the highest value consumer demographics for a product's flavor-profile.

Position: Full-Stack Engineer

Role: Web-application or Streaming Infrastructure focused full-stack engineer capable of integrating the data pipeline and outputs of machine learning models into an easy to use management platform.

Must have: -Professional experience or -Experience with Go, Javascript, Docker, R (OpenCPU), MySQL, AWS

Bonus: -Experience designing and maintaining streaming infrastructures -Experience with TensorFlow

Team: We're a diverse 6-person company (across data, engineering, chemistry, design and biz) that is passionate about high-quality food and beverage products.

Next Steps: Please submit something awesome to JasonCEO@Gastrograph.com to apply.

Read more

Looking for a specific post?