Using AI to Fight Food Fraud

Somewhat unexpected application, good details embedded here.  Note the automated tasting sensors involved, which we investigated for coffee taste applications.

Using AI to Fight Food Fraud  By Sandrine Ceurstemont, Commissioned by CACM Staff, July 7, 2022

Food fraud is a global problem that typically involves the dilution or mislabeling of food products, or ingredient substitution. In 2013, horse meat was found in many supermarket meals in Europe that claimed to contain beef, for example, while milk has often been found to be waterd down in India to increase profits.

A 2021 study by the United Nations Food and Agriculture Organization cited a 2018 European Commission finding that estimated " the cost of food fraud for the global food industry is approximately EUR 30 billion" (about $30.5 billion) each year.

While chemical analyses can be carried out in a lab to authenticate food, traditional methods are often expensive, time-consuming, and require technical expertise. That is why researchers are aiming to develop new tools that harness artificial intelligence (AI) to enable rapid, inexpensive screening of food and beverages.

"It would be a very exciting scenario to have AI help us expand the reach and impact of chemical analyses," says Patrick Ruch, a research staff member at IBM Research in Zurich, Switzerland. "All of the intelligence can be on a smartphone or in the cloud."

Ruch and his colleagues have been working on a system to authenticate beverages called HyperTaste that uses a small, portable device called an electronic tongue (e-tongue), combined with machine learning. The e-tongue contains 16 sensors made of conductive polymers that can be thought of as taste buds; when dipped into a drink, the sensors pick up chemical information in the liquid that can be converted into a unique digital fingerprint measured as a time series of voltages. "We know that the signal that we're measuring is a unique indicator of what's inside the liquid because these polymers are interacting with all of the small molecules inside," says Ruch.

Machine learning then is used to make sense of the complex signal detected, for example to identify a specific brand of wine or its origin. In recent work, Ruch and his colleagues focused on wines and juices, training three different machine learning models to perform various recognition tasks using data collected with a sensor-equipped robotic device. The automated system dipped its sensors into nine different types of fruit juices several times, collecting 72 voltage time-series measurements. The process was repeated using 11 different types of Italian red wines to generate 110 measurements. "Nowadays, you can really quickly obtain the data needed for training with automation," says Ruch. "Within half a day to maximum a day, you have all the training data you need."  ... ' 

Robot 'Chef' Represents Leap Forward in Automated Cooking

Practical application still behind, will it ultimately take over?  Ultimately in  what contexts?   Big interest of mine. 

Robot 'Chef' Represents Leap Forward in Automated Cooking

Bloomberg,Nina Massey, May 4, 2022

The designers of a robot chef at the U.K.’s University of Cambridge trained it to "taste" foods at different stages of the chewing process to determine their saltiness, which they suggest could help in developing automated or semi-automated food preparation. "We want the robots to understand the concept of taste, which will make them better cooks," said Cambridge’s Arsen Abdulali. People notice changes in taste and texture as they chew, and the robot, which had been trained to make omelets, tasted nine variations of scrambled eggs and tomatoes at three different stages of the chewing process, then produced taste maps of the different meals. The robot used a saltiness sensor attached to a robot arm to mimic the chewing/tasting process, and “tasted” dishes in a grid-like manner.  ....' 

Hypertaste for Chemical Sensing

We worked on automating taste and smell, but it was still not possible at the time.  Could have used this to analyse coffee, green and roasted, to determine its use in blends.  Counterfeiting too would be an application, also mentioned

IBM’s new Hypertaste system is an AI-assisted, reprogrammable chemical sensor  By Maria Deutscher in SiliconAngle.

Researchers at IBM Corp.’s Zurich lab today unveiled their latest project: a chemical detection system called Hypertaste that uses artificial intelligence to classify liquids.

Distinguishing different beverages and food items is usually a trivial task for humans, but the same can’t be said for the machines that experts use when the taste test doesn’t cut it.

Systems capable of performing chemical analysis on liquids generally fall into one of two categories. They’re either portable instruments that provide only limited information such as the pH level of a lake or bulky, expensive systems built for use in labs.

IBM developed Hypertaste for scenarios where a detailed chemical analysis is necessary but it’s not possible or economically feasible to send a sample to a lab. One application the company envisions for the system is evaluating the water quality of rivers. Another is finding counterfeit goods, such as low-quality perishable items shipped to a grocery chain by an untrustworthy supplier.  .... " 

Complexity of Taste Expertise

Learned the complexity of this in the favor space for roasted coffee.  Some people just cannot do it.  How do you make machines do it?

Learning to Become a Taste Expert
by Kathryn A. Latour and John A. Deighton in HBS Working knowledge

OVERVIEW — How should we learn to discriminate a fine wine or chocolate? Tradition says use a flavor wheel and map the taste into vocabulary. We find that works for novices, but, beyond a point, it is counterproductive. Enthusiasts perform more like experts when they abandon language and just “draw the shape” of the taste.

AUTHOR ABSTRACT
Evidence suggests that consumers seek to become more expert about hedonic products to enhance their enjoyment of future consumption occasions. Current approaches to becoming an expert center on cultivating an analytic mindset. In the present research the authors explore the benefit to enthusiasts of moving beyond analytics to cultivate a holistic style of processing. In the taste context the authors define holistic processing as non-verbal, imagery based, and involving narrative processing. The authors conduct qualitative interviews with taste experts (Master Sommeliers) to operationalize the holistic approach to hedonic learning, and then test it against traditional analytic methods in a series of experiments across a range of hedonic products. The results suggest that hedonic learning follows a sequence of stages whose order matters and that the holistic stage is facilitated by attending to experience as a narrative event and by employing visual imagery. The results of this multi-method investigation have implications for both managers and academics interested in how consumers learn to become expert in hedonic product categories.   ... "

AI and Smell Digitization Classification, and Health

We actively worked on digitally reproducing the senses of smell and taste.   Both in their classification and their production.   In one case for taste of products like coffee and also the analysis of aromas produced by perfumes in context.   In either case it can convert such recipes into digital form, which are more manipulable.  Both of these challenges depend on very low amounts in the atmosphere.  And now, Since Machine Learning AI can detect subtle patterns,  there is cause to believe we can solve these problems.   Further, the sense of smell is being used to detect the state of human health.

Pringles are Getting Louder

Kellogg bought Pringles canned chips from P&G in 2012 for $2.7 Billion.  Was slightly involved in the brand.   The new brand is pushing new and spicy tastes, and is selling well.  Strikes me that spicier foods continue to grow.  What is causing that?   See Pringles tag below for more.

Electronic Tongue Tastes Wine

An area we explored on the coffee blending business.  Was still very primitive in capabilities at the time.  Some advances are seen here.  How can resulting data be connected to blending or manufacturing data?   In New Scientist. 

NeuroWine Claimed

A more direct sensory link than the usual examples.  

SA neuroscientists develop first NeuroWine
By Admire Moyo, ITWeb's senior news journalist.
Johannesburg, 20 Apr 2016
.....
A South African team of neuromarketers and neuroscientists created the world's first NeuroWine, a wine that was developed by taking the tools and technologies that are traditionally used in neuroscience and applying them to the art of wine-making.   ... " 

Hot Sauce and Taste

In the past I worked in the food category of CPG, and learned a little about taste trends and how they influence purchases.  Also a devoted amateur chef.  The considerable change in the acceptance of  spiciness in US tastes is remarkable.  More on that.    See also work by McCormick, in the tag below.

Sharing Cokes as Advertising

Sharing Coke and user generated ads.

" ... Ogilvy's #shareacoke campaign for Coca-Cola has generated 250,000 tweets and almost 1 billion Twitter impressions, but that buzz may not have much long-term impact on consumers who aren't already Coke drinkers, writes Stephen Carter. "[W]ith younger consumers so fickle, the ability to turn marketing into magic will depend on part on whether the rising generation actually tries the product and likes the taste," he writes ... " 

" ... This is where Coke's challenges arise. Even if fans are tweeting like mad, recent work suggests that user-generated ads have little effect on the preferences of borderline customers, and even arouse their skepticism. True, this effect seems to be lessened when the audience thinks the consumers who created the messages are people much like themselves. But with younger consumers so fickle, the ability to turn marketing into magic will depend on part on whether the rising generation actually tries the product and likes the taste. ... " 

Cisco Managing Video

I reported on this some time ago, as it related to Video Content Analysis. (VCA) and its use for ad management in store.

Cisco introduces ‘videoscape unity’ to manage the future of video everywhere
Cisco is the world’s networking giant, and today it showed what it is doing to make sure that it can provide the plumbing for the explosion of video on live TV, paid services, and the wide-open Internet.  ... 

For the consumer, this means that Cisco will make it easier to manage video content, regardless of where it is coming from. Cisco is creating apps that allow a user to search through videos in a visual way, organized based on the user’s own tastes. You can login and your own profile will come up. The user interface will show you the videos that are relevant to you. Cisco refers to its unified approaching to managing content as Videoscape Unity.

Cognitive Cooking Book: Creativity?

I note that this book is now available, on order, I am a cook and was a Beta user and found the system useful, but not necessarily better than  using one of the many recipe systems out there.   It does provide an interesting template for other knowledge based systems. It also inserts the idea of creativity, not commonly included in AI.   What is creativity?  Knowledge and experience filtering an element of randomness?  I was hoping to use Chef Watson for a recent group dinner to get real feedback, but that did not work out.   Like to get other input as to its value.

Cognitive Cooking with Chef Watson: Recipes for Innovation  from IBM & the Institute of Culinary Education Hardcover 

Take Your Cooking to a Whole New Level with Chef Watson and the Institute of Culinary Education!

You don't have to be a culinary genius to be a great cook. But when it comes to thinking outside the box, even the best chefs can be limited by their personal experiences, the tastes and flavor combinations they already know. That's why IBM and the Institute of Culinary Education teamed up to develop a groundbreaking cognitive cooking technology that helps cooks everywhere discover and create delicious recipes, utilizing unusual ingredient combinations that man alone might never imagine.

In Cognitive Cooking with Chef Watson, IBM's unprecedented technology and ICE's culinary experts ... "

3D Printing Cuisine

Have covered some of the work by Hershey.   But the technology is still slow and in prototype.  Printers are still uncommon in the home:

" ... But there are signs of progress. Debrauwer says that his researchers have been playing with parameters such as the amount of air trapped in the food, the density of the food and the amount and length of fibre it contains to modify the texture. And Roovers says that he has had good experiences in printing cheese and strained yoghurt, which he praises for its firmness.

In the meantime, 3D-printing aficionados are already trading recipes on the Internet. Some are experimenting with unusual ingredients. Debrauwer says that his lab mates once gave him a wafer printed with protein harvested from ground-up insects. “It tasted quite good.” ... " 

Balancing Shopper Desire for Personalization

In Supermarket News: 

FMI CEO: Balancing shopper demand for personalization
Food retailers can fulfill consumer demand for more personalized service without becoming overwhelmed if they focus on the demographics of their customer base, identify the groups to which the majority of their shoppers belong and then customize the shopping experience based on the preferences and tastes of those groups, FMI President and CEO Leslie Sarasin writes. "If a food retailer can become sufficiently knowledgeable of the communities that comprise their shoppers and create a shopping experience contoured to each group's tastes, they will have made tremendous strides toward offering customized service," she writes. ... " 

Juji: Behavior-based Individuality Analysis

Michelle Zhou brought to my attention their new personality analysis startup that creates people profiles from Twitter and Facebook streams.  I  gave it a try and was impressed by its ease of use.     The idea of this kind of user modeling has been around for a long time, we tested related ideas with text analyses of 800 line calls.  This example is one of the first that uses new sources of social data effectively.  Consider if you combined results of this system with streams of IOT data, you could provide an analysis of people in context.

Juji has similarities to and in part derives from work done at IBM, that I have reported on previously, but it is now a unconnected startup.  It provides both profiles and alternative use case scenarios.  Give it a try, they need a critical mass of testers, and are providing incentives for students to participate.

Sign up and test the system.

They write about their goals:

Our Vision
Juji's mission is to discover the world’s people by their individuality, and to enable individualization at scale. Our notion of individualization includes two aspects. One is self individuation, a concept proposed by the renowned Psychologist Carl Jung: one should be aware of and realize one's true self. The other is for individuals to obtain hyper-personalized experiences that suit their true nature, including their personality, physical and emotional needs, values and beliefs, and taste and style.

As the first step toward fulfilling our mission, we empower our users, individuals and businesses alike, to automatically derive deep insights about people of interest from opt-in or public data (e.g., public blogs and web pages). Such insights characterize a person's or a group's individuality from multiple aspects, encompassing their professional and personal lives. Our users can then use the derived insights to develop and improve their own brand personality, and create individualized engagements at scale (e.g., interacting with hundreds or thousands of customers in an personalized way).  ....    "

Multi Sensory Computer

In Forbes, on the IBM Chip:

Meet The Brain-Inspired Computer Chip That Can Smell, Feel And Hear
By Dharmendra S. Modha, Ph.D., IBM Research

Building a computer that could match the power of the human brain has long been a goal of scientists. In August, we made a breakthrough, published in Science in collaboration with Cornell Tech, which is a significant step toward bringing cognitive computers to society. We announced that we’ve built a computer chip that functions like a brain does with the ability to sense, taste, feel, smell, hear and understand its surroundings.

The SyNAPSE chip is a neurosynaptic computer chip that is a size of a postage stamp and runs on the energy equivalent of a hearing-aid battery. This technology could transform science, technology, business, government, and society by enabling vision, audition, and multi-sensory applications ... " 

Fake Tongues and Taste in Virtual Reality

An article in the New Yorker looks at an extension of the 'Rubber Hand Illusion", which realisticaly fools the brain into thinking that a fake hand is their own.    Called the 'Butcher's Tongue Illusion'  this extends the same kind of sensory illusion idea to the tongue. Good article provides the background.   The implications and practical applications?

" ... The results of the butcher’s-tongue illusion suggest that, with further testing, it might be possible to one day enjoy out-of-body flavor experiences. To a person wearing Michel’s box, a Slim-Fast shake could be made to taste like a chocolate-fudge sundae, delivering a fraction of the calories and a great deal more pleasure. It’s the holy grail of dieters: a virtual-reality headset that makes those undressed salad leaves you’re dutifully eating taste like French fries or dark chocolate. At last, we could have nutritional restraint combined with sensory decadence, with no sacrifice  ... " required.

Transforming Messaging with Data

Foodies Transforming grocery messaging.  Not really a technical idea, but an understanding that the form of the message needs to adapt to change in consumer expectations.  Have seen this same form of change occurring in consumer taste for spicier foods.   How can we predict this trending more accurately?  Its in the data we have.

P&G, Pepsi Try Foursquare Swarm

In Adage: Looked at Foursquare in its early days.  Interesting that some giants in marketing space are trying the new functionality. " ... Now the company is calling them "place-based ads," a rebranding that mirrors its current efforts to distance itself from check-ins. The new Foursquare -- which executives say will launch this summer -- will dispense with check-ins altogether and focus on discovery of nearby venues tailored to users' tastes and will take  ... " 

McCormick Flavorprint

In Consumer Goods Technology:

Exclusive Video: Learn How McCormick Leveraged Marketing/IT Collaboration to Develop FlavorPrint
As marketers look to leverage digital innovation to reach consumers, collaboration often plays a role as partners work together to achieve success. In this video from the 2013 Consumer Goods Business & Technology Leadership Conference, CIO Jerry Wolfe and Digital Marketing Manager Colleen McClellan of McCormick discuss one such collaborative initiative – FlavorPrint, a new technology that helps consumers discover flavorful foods they'll love based on their flavor profile .... " 

During the session, Wolfe and McLellan explore how multiple internal cross functional teams at McCormick, from marketing to IT to sensory scientists, joined forces on the FlavorPrint project. Additionally, they discuss how McCormick worked with its external agency and retail partners to deliver this unprecedented service. ... " 

This has similarities to work we did to determine and model taste preference profiles for coffee blends.

Previously in this blog about delivering Flavorprint.

View Video in Consumer Goods Technology.

Also see McCormick's Flavorprint Site.