Research Revealed at Think 2019

cool stuff going on. And this is my particular point of view of what I saw. I'm really excited about this particular topic. If you listened to the episodes I recorded last year at IBM Think two thousand and eighteen, which was in Las Vegas, Nevada, you heard my two part series about the five and five presentation, which is a session in which IBM researchers presents some of the cool things they're working on and the results of some

high tech bleeding edge research. This year, IBM changed things up a little bit by focusing the presentation on one broad topic, and it's one of my favorites. Food. I love food, particularly if there's hot sauce involved, but the IBM research was a bit more ambitious than finding the

right condiment to make my burrito zing. Rather, the presentations I saw brought the audience on a journey for the entire ecosystem of our food, from growing it to distributing it, to figuring out what to do with plastic waste that's

generated afterward. The subject is a really important one, so when you combine the threats of climate change with the growing population, you quickly come to the conclusion that feeding the planet is just going to get more challenging than it already is, and that managing resources and making food available will be absolutely critical. Each presenter focused on a particular topic, which led pretty smoothly into the next one. So I'm going to give you a rundown on what

those presentations were and the related technologies. First up was Juliet Mutahi, a software engineer from Nairobi, Kenya. Her main focus was on the food chain, the food supply chain. She personalized her story by telling the audience of her background as the daughter of a coffee farmer in Kenya. Her father's coffee farm is part of a cooperative or a co op, and that's an association of business owners who worked together for the common benefit of the members.

They can coordinate to negotiate the best prices for their products, for example, and make sure that no one is failing to get his or her fair share. And co ops help establish best practices like fair pricing and labor and they can negotiate long term contracts with buyers on a level that an individual farm owner might not be able to manage. In Kenya, there are twelve thousand members who are part of these co ops and they contribute more

and half of Kenya's coffee production. Cooperatives work because their members share information across the value chain. They do it through spoken word. If one farm produces high quality coffee, the cooperative would negotiate to get a fitting price, a premium price for that coffee. So how can we use technology to sort of achieve the same sort of things

that have been going on in cooperatives. Well, technology is allowing for the next evolution of this model, and like pretty much every digital solution you can think of, it all revolves around data. How can farmers collect more information about their land, their soil and make reliable predictions of future harvests? To better anticipate contract negotiations or better manage their farms. She spoke of an approach in which a farmer would use a tractor with embedded sensors in it

to till the land. The tractor would be doing its normal tractory duties while simultaneously creating a digital map of the farmland itself, so that now there's a digital representation of the farm. Feeding the information to IBM S Watson for Agriculture Platform would allow for meaningful use of that data. Watson can take that information and combine it with other sources to make predictions about future yields and give farmers

ideas about the conditions of their land. It can also analyze the data to estimate what past yields might have been. And you might wonder, why would you ever need to know what has already happened. Why would you need to estimate a past yield? It becomes really important in cases where a farmer has to file an insurance claim. It

helps justify that insurance claim. If the farmer says that, due to whatever reason, the yield was a certain size and the data backs that up, it can help the farmer get that insurance claim and it also helps build out models that will increase efficiency further down the supply chain. Mutahi also talked about a cool analytical tool called the IBM Argo Pod. This is a great example of an Internet of Things device. It's about the size and shape

of a regular business card. Embedded in the card are sensors that can do soil analysis, and a farmer just needs to put a small sample of soil on the card and chemical reactions will tell the sensor everything it needs to know in just ten seconds, and then the farmer can take a picture of the card using a smartphone and use a related app to store the information

into a blockchain record. This doesn't just tell the farmer of the conditions on the farm, it can also help the farmer establish credit because a bank could extend credit to a farm against a predicted harvest that's based on this collected information. So for farmers all over the world, this could be an enormous help. Mutahi then handed the stage over to Shri ram rug Hoven. He is the Vice president and chief Technical Officer of IBM Research in India.

He shared a pretty staggering pair of facts. One third of all food produced and nearly half of all fruits and vegetables never get assumed. It just goes to waste. So imagine if half of your work was immediately dismissed. You still had to do the work, but you know that half of it wouldn't count. That would be frustrating for most jobs, but when it comes to food production, it can lead to waste management problems at best, and at worst you could be facing starvation issues. So what

is the challenge here? Is this just a case of some places having more food than the population can consume. Well, it's actually a lot more complicated than that. The food supply chain is a big issue. There are many points along a supply chain, and at every stage spoilage can and does occur, So there's a decent chance that a lot of food will be spoiled before it can ever

find its way onto a market shelf. From the fields to the storage facilities, to transportation vehicles to distribution centers, to processors to shops to the home, there are a lot of stops along the way, and presently this is largely a dumb system, meaning there's no real way to gather information and share it along the supply chain. So if you receive a creative apples in a distribution center that are maybe two days away from being at peak ripeness,

but you have no way of knowing that information. When you've got the crate, you might put that create on a truck that's going across the country, and by the time the apples get to their destination, a large number of them have already passed their prime and they just get thrown out. So what's the solution to this problem.

The proposal we heard is that you would take a combination of technologies, and that includes the Internet of things, blockchain, and artificial intelligence in order to keep track of everything and make decisions. The Internet of things and the blockchain combined could log each stage of the journey the food takes from field to the market and keep track on

the freshness of the food. There would be a record for each crate or palette or whatever unit of food that could record when it arrived and when it each point, and you could understand quickly how much time had passed since it was harvested. AI could help guide the decision making process when it comes to deciding where do you send this next. So the example we heard in the presentation involved oranges, and here's how it goes. Let's say that you run a distribution center in Florida near where

oranges are grown. So farmers oranges come into your distribution center, and it's your job to send those oranges out to UH stores in various cities. And the two main cities under your responsibility are Atlanta and Chicago. And as it turns out, in Atlanta, people are hog wild for these oranges. And I can confirm that because at least from this Atlanta's point of view, this is accurate. But in Chicago, oranges for some reason just aren't moving nearly so fast

from the produce section. With access to this information, the AI can make decisions. Oranges that still have a really long time to ripen could be sent to Chicago because they could sit in refrigeration a bit longer. They could remain fresh while the inhabitants of the Windy City decide they finally want to fight off scurvy. Oranges that are close to passing prime ripeness can make the shorter trip to Atlanta, where they are more likely to be purchased

and consumed quickly. The distributor can maximize efficiency and minimize waste. Now next we're going to hear about a proposal that is all about food safety. But first let's take a quick break. The third presenter in the five and five event was Gerroud Dubois, director of IBM research at Almondon. Dubois began his presentation by talking about a catastrophe in China in two thousand and eight. A shipment of baby

formula had been contaminated with a chemical called Melman. This is an industrial chemical that manufacturers used to treat stuff like ceramics and plastics. It is used in glue, in flame retardants, and laminates. It is not, as you might imagine, meant for consumption. Parents who fed their babies this formula quickly became alarmed when the babies began to get sick. Melman can cause kidney damage and Chinese hospitals had to

admit more than fifty thousand babies. And even darker part of the story is that the addition of the chemical was likely intentional. The chemical makes it appear that the formula, when combined with milk, is more protein rich than it actually is, so it was an attempt at deception. So that's even more disturbing. Well, du Bois proposed that we rely upon microbes as a sort of microscopic canary in a coal mine, or as he called it, a microscopic c I a agent, and it would detect when contaminants

are present in food products. And a microbe is a micro organism. Bacteria are a type of microbe, and you've likely used stuff like antimicrobial products to help sanitize stuff, and this could give you the implication that microbes are bad and some are definitely harmful to us, but some microbes are beneficial. And the human body is host to

almost actually slightly more microbial cells than human cells. The ratio appears to be somewhere around one point three bacterial cells to every human cell, so you could say that you're more bacteria than person, though that's not really the point. Our microbial biome or biota is actually part of us, so I would say we are collectively made of both bacteria and human cells. We are bored, I guess. Anyway,

Not all microbes are bad. Some are very helpful and they aid us in digesting certain materials, so some are good and we should rely upon those. Do Boa proposes we could use microbes to detect the presence of contaminants and food. So for a long time, scientists thought that the behaviors of microbes were too complex and unpredictable to

reveal meaningful information, that it was almost random. But by studying them at a genetic level, looking at the d n A and the RNA of microbes, we've gathered a lot more information that reveals microbes react in different ways to different conditions. To make that determination, we had to use a big data approach. Dubai used a helpful analogy to describe how challenging this process really was. So imagine

you have ten thousand boxes of jigsaw puzzles. They're different puzzles, and you dump all the pieces out into one enormous pile. You mix them up real good, then you throw the boxes away, so you don't even have the record of what the pictures look like, and now it's your job to put together those ten thousand puzzles. That's sort of the scale of the job we had to do to

suss out how microbes behave given different conditions. At IBM, the research team created a database containing all the microbial genomes the scientific world has mapped so far. The genomes and the analysis of microbial behavior represents around five hundred terabytes of data. But the process was successful and that there are microbial behaviors that can indicate the presence of

contaminants in our food. You have to know which microbes you're looking for and what behaviors indicates safe versus unsafe food. But it could add a new method for food safety inspectors to use to guarantee that the products that make it to consumers are actually safe to consume. Du Bois then introduced us to Donna Dillenberger and IBM Fellow at

the Thomas J. Watson Research Center. Now, I mentioned in an earlier episode that the title of IBM Fellow is the highest honor IBM bestows on distinguished researchers, scientists, engineers, and the like. Dillenberger also talked about food safety. She cited a CDC report that I tracked down. The report is the Emerging Infectious Diseases Report, and the section Dillenberger was specifically referencing is chapter five of that report titled

food Related Illness and Death in the United States. Now, I'm going to quote the abstract of that report verbatim. Here's part of that abstract. We estimate that food born diseases cause approximately seventy six million illnesses, three hundred twenty five thousand hospitalizations, and five thousand deaths in the United States each year. Known pathogens account for an estimated fourteen

million illnesses, sixty thousand hospitalizations, and eighteen hundred deaths. Three pathogens Salmonella, listeria, and toxoplasma, are responsible for fifteen hundred deaths each year, more than seventy of those caused by known pathogens, while unknown agents account for the remaining sixty two million illnesses, two hundred sixty five thousand hospitalizations, and

three thousand, two hundred deaths. Overall, food born diseases appear to cause more illnesses but fewer deaths than previously estimated. End quote. Now, fewer deaths is a good thing, but even fewer would be better, And the huge number of illnesses represents not just discomfort, stress, anxiety, at a or quality of life. There's a ripple effect. The hits society as a whole in ways like lost productivity or additional burden on the medical industry. So how can we tell

if the food we buy is safe to eat? The bacteria responsible are microscopic, As I mentioned, just a moment ago. We're not likely to have a lab set up in our kitchens. Dyllenberger introduced a hardware and software tool that could actually help. The hardware component is a smartphone peripheral. It's essentially a microscope that plugs into your smartphone and feeds a magnified image to the phone's camera sensor. The software side is an analysis tool using AI and image

recognition strategies to identify bacteria present in food. The results of the analysis pop up on the screen as shapes around the bacteria, and the type and color of shape indicates which kind of bacteria might be present. It can scan down to the micron scale and let you know if the food you have is safe to eat or if it is host to say, a colony of eke to lie. Dillenberger also revealed that the same technology could be used to help discover food counterfeits, and yes, that

is a thing. In fact, it's a big thing. Dillenberger referred to a study that found in the United States, olive oil producers weren't always being totally honest, or at least they weren't meeting the right standards. Because a study in two found that seventy of the bottles of labeled extra virgin olive oil actually failed to meet the standards

to qualify as extra virgin. You wouldn't be able to tell the casual glance the difference between regular olive oil and good old e v O O, but the scanner and paired software can analyze the reflected light from a sample of olive oil and determine whether or not it matches the real deal or not. The same tech can help you figure out if a label is legitimate or

if it's a clever fake. And it doesn't stop with food, though that was the primary focus of the five and five presentation, the scanner could also be used to sort out the real McCoy from fakers and all sorts of products,

from fashion accessories to prescription drugs. Moreover, Dylan Berger said that while the present implementation uses a smartphone, future versions could see embedded sensors integrated into common kitchen tools like cutting boards or knives or measuring cups or bowls, and by looking for not just authenticity, but for the presence of potential pathogens, we could improve our fine dining experience

and also prevent cases of food poisoning. Next, we're gonna learn a little bit about the conclusion of the five and five event and what we might do with some of the plastic waste we generate. But first let's take a quick break. The final presenter at five and five was Jeanette Garcia, who has one of the most kick ass titles I have ever seen. It's Master Inventor at IBM Research Almada. She helped put into perspective exactly how big a problem plastic waste has become, and it is

a huge problem. The world produces around three hundred million tons of plastic every year, and about half of that plastic is in the form of stuff that's intended for single use, meaning we just toss it out afterward. Lots of communities are now recycling, but sadly, only a small fraction, about ten per cent of plastic gets recycled. So why is that, Well, there are a few reasons. One is that there are different types of plastics and they can't all be processed in the same way, which means they

have to be sorted into different categories. That takes effort and time, which also means it costs money. There are contaminants that can be a problem particularly food waste, and Garcia mentioned that we often don't know what the rules are are we supposed to rinse off the plastic. First, some types of plastic aren't recyclable at all, And then there's the problem of economics. Pastic is pretty easy stuff

to manufacture. If it costs more money to recycle plastic than it does to just make new plastic, you have to come up with some other economic incentives to encourage recycling. At the same time, plastic is undeniably useful stuff. We rely on it for lots of things, not the least of which is food packaging. It's hard to dismiss how important plastic is in keeping our foods fresher longer, which goes back to cutting down food waste. So how do

we resolve the problem. The specific type of plastic Garcia talked about was pet or pet plastic, which is found not only in packaging but also stuff like polyester clothing. The seventies, this is the type of plastic used in stuff like water and soda bottles. If you look at a bottle made from pet, you'll see the chasing arrows sign with the number one at the center. It makes up a lot of plastic, but I should point out that it's also the most recycled type of plastic in

the world. PET is completely recy cyclable. The United States lags behind Europe when it comes to recycling PET plastic. We in the US recycle a little more than of PT plastic, while Europe is over the line, but we can always do better. One of the methods we rely upon to recycle this type of plastic is to wash the plastic and then melt it down. Another version uses chemicals to depolymerize the plastic. Polymer is a long chain of molecules, so this approach effectively breaks those chains down.

But the melting process has a zero tolerance for contamination, meaning the only stuff that can get melted down is the plastic itself. Any other stuff will foul the melted mixture and make it unusable. It also only works with clear bottles. The chemical approach has its own drawbacks, a big one being the cost of the process, which creates that economic barrier I mentioned earlier. IBM solution is called volcat which stands for volatile catalyst. So what the heck

does that mean? A catalyst is a substance that facility. It's a chemical reaction. You would typically use a catalyst to speed up a process that would otherwise take much longer to complete. Catalysts do this without undergoing permanent chemical changes themselves. They're incredibly useful in hundreds of different industrial applications. With the volecat approach, IBM could grind up pet plastic into flakes, and it wouldn't matter if the plastic was clear or had color added to it, or if it

was clean or if it was dirty. With volcat, IBM takes that plastic and puts it into a heating chamber. The chamber heats the plastic up to about a hundred nine degrees celsius or three seventy four degrees fahrenheit. At that point researchers add the catalyst. Then they cool the mixture down to get below one hundred degrees celsius or two hundred twelve degrees fahrenheit. That's the temperature which water

boils under one atmosphere of pressure. The plastic polymers break down into a more basic molecular form called B H E T, which is a monomer. And you can think of a monomer as a molecule that forms the basic component of a polymer identical monomers linked together to form polymers. So now you've got what is effectively the building blocks of PET. The volatile catalyst is completely recoverable and IBM engineers can filter the b ET from any other stuff

that was in the mix. The b h g T can go on to form new pet containers, So we could keep using the same plastic over and over again without the need to make more of the stuff, and should our demand for plastic exceed that amount, we could literally mine the Pacific garbage patch and landfills for more material. The process sounds incredibly promising and increasing recycling rates, particularly for plastics that are dirty or have color added to

the plastic material. Since the process removes the need to clean the plastic first, it could reduce the recycling workload. If it is scalable, it could be a viable alternative to the present approaches to plastic recycling. These research projects were all really interesting to me, and they are all trying to make big changes in the entire food life cycle.

It's a truly enormous challenge and there's still more to work on developing technologies to reduce water waste, improve transportation and distribution deal with other types of waste, and guaranteeing everyone in the world has access to healthy and safe food is a monumental effort. These research projects are just a few of the ways innovators are looking into making a big difference. Will we see each of these ideas

implemented in the real world. What's a bit early to say, but all of those five presenters made a prediction of what the world will be like in five years if we do further develop, scale and deploy those technologies, and it was a pretty good start to fixing some incredibly tough problems. Research will continue to play a huge part at IBM. It's ingrained in the company's culture. Much of it focuses on the company's core businesses, which is to be expected. You can't stay in business if you never

focus on your services and products. But a lot of that work has applications beyond the enterprise world, and it's this drive to leverage powerful technologies to make real changes out there that I find so interesting. I'd like to thank IBM for bringing me out here. It was great once again getting to explore and experience the Think Conference. I got to see a lot of cool presentations about technologies that I was aware of but didn't know very

much about. And I also really am thankful for the opportunity to talk to so many incredible people here at the conference and get their perspectives on technology and how we might apply it in the future. If you guys have any suggestions for future episodes of tech Stuff, contact me the email addresses tech Stuff at how stuff works dot com, or go to our website that's tech Stuff podcast dot com and you can find the ways to

get in touch with me through social media. There. You can also find the link to our merchandise store and I will talk to you again really soon for more on this and thousands of other topics. Is it how Stuff Work Stacom