Boundless Science Season 2: How Can AI Help to Reduce Food Waste?

What if you could buy your broccoli at the grocery store, knowing beforehand for how long it will remain good in your refrigerator? I am Shenara Ramadan, a master’s degree student in Agricultural and Natural Resources Communication. In this episode of Boundless Science, I had a great conversation with Dr. Tie Liu, an assistant professor of Horticultural Science at the University of Florida, about AI advances in reducing food waste worldwide and one of his recent research projects that enables producers and consumers to determine the shelf life of fruits and vegetables. #boundlessscience #season2 #BoundlessScienceseason2

Keywords: #FoodWaste #AIInAgriculture #FoodTech #HorticultureResearch #ScienceCommunication

Transcript

Shenara: Hi everyone. Welcome to this episode of our Boundless Science Podcast Series on Streaming Science. Streaming Science is a student-driven program that works to connect you with scientists is to learn how science impacts all of us in our everyday lives and interests. My name is Shenara Ramadan, I am a master's degree student in the Department of Agriculture, Education and Communication at University of Florida, and I am your host for this episode. In this interview, I am talking with Dr. Tie Liu, who is an Assistant Professor in Horticultural Sciences at the University of Florida. Dr. Liu started his studies in China. And then he did his graduate studies in both Peking University and Yale University, and then started his postdoctoral research at Stanford University. While there, one of his research projects was a comparative genetic and cell imaging approaches. Dr. Liu also has dedicated years of his research to food waste. According to USDA, about 40% of all food produced is wasted in the United States. And researchers also found out that post-harvest losses ranged from 12 to 46% after seven days in storage. Through this conversation, I hope you'll learn about how artificial intelligence is helping to solve food waste issue. Welcome, Dr. Liu. 

Dr. Tie Liu: Thank you for having me on your podcast and really glad to talk about food waste. 

Shenara: It's a pleasure to have you here with us. Dr. Liu, based on your trajectory as a researcher in this field, do you consider that science can have a determinant role in food waste reduction? 

Dr. Tie Liu: So, I do believe the sciences or technologies would be really great tools or, I mean, driver can reduce food waste and loss. The reason I'm talking is because there's a lot of research has been done, and especially in that a recent couple of years to apply different technologies, such as AI or more farming those technologies to see how they affect food productions or management and related food waste. So regarding to article published in Nature, basically, researchers looking at three major factors can really affect on food waste reductions, for instance, the food management system, the technology, such as the newer technologies, as well as personal diet is a kind of main factor effect on the food waste and loss. So in technology, such as those agricultural related technology, really can helping food management. Those including handling or storage, or processing. So think about if we can make the food management or transportation more efficient, that's really can reduce a lot of waste during the food supply chains. So in my opinion, the technology definitely can helping reduce food waste and loss. 

Shenara: In your opinion, what are the scientific breakthroughs in terms of food waste that come to your mind?

Dr. Tie Liu: In my opinion, there are two major technologies or two major fields such as gene editing technology and AI-associated technology are really a breakthrough, can helping reduce food waste and loss, gene editing, I think a lot of probably a lot of audience and heard those. It's a way to editing genomes and so then we can managing the, for instance, gens of vegetables and fruit to be able to develop more efficient cultivars or a variety of foods such as vegetables and fruit, they can either have prolonged shelf life or has better qualities and all these traits can really increase the customer likening, so then will lead to reduce food waste. The other is AI associated technology. This is also another broad field, AI’s can be very broad such as the we're talking about smart farming, for instance, during the vegetable and fruit grows stage, how to managing fertilizer applications, managing water efficiencies and water managing phosphorus or any other fertilizer application or pesticide applications really can improve the crop productions and also management. So I think either of those two areas can really advanced our understanding of crop or vegetable a system can, so then, we can managing those vegetable and fruit production in handling and transportation storage well, eventually can help us to reduce food waste and loss. 

Shenara: So when you talk about editing the genes of the vegetables, you are saying we can produce vegetables that the shelf life is longer.

Dr. Tie Liu: Right, exactly, so vegetables and fruit were, you know, currently purchase from grocery store. This is basically a lot of vegetable and fruit are, has been humans, selected through natural conditions. So that means the lot of vegetable during our selections mainly focus on either productions for instance, their high yield, or better taste, or size better or color have more bright for instance strawberry or blueberry. However, in the aspect of farmer they may not think a shelf life is a main factor to select those vegetables and fruit. However, to reduce food waste, and loss is really important to keep them have a longer shelf life and those traits are not being selected in in the current grocery store. So what we can do is, we have to go back to the original time, and then looking for trraits or looking for genetic coded information to helping us identify a new cultivar or new variety, which has a longer shelf life. So I think that's back to your question, it is important to find this quality or variety have longer shelf life. 

Shenara: Interesting. In terms of your research, what have you been researching about food waste in your scientific journey?

Dr. Tie Liu: So my lab is basically study postharvest biology. So, in other words, to understand how vegetable and fruit decay, or senescence or in other words aging, right? And then the goal, of course, is to develop a longer shelf life or extend their shelf life vegetable and fruit, at the same time maintain their better quality, for instance, their taste is still good, even your store and your longer time, their quality non reduced. So in the past few years, my lab is basically looking for different traits in vegetable and fruit that leading to a longer shelf life. So we're applying a different approach. For instance, use some molecular biology approach. Those process that I mentioned earlier, gene editing, it's a really cutting-edge technology. We're currently also using this technology to identify those traits.

The second, we're also very interested in to use AI-associated technology to understand those traits, and then developing technologies or developing devices to monitor those shelf-life-associated traits. 

Shenara: Your recent research is about artificial intelligence. Before diving into your project, can you explain what AI means? Is it a software or a group of machines? To what extent is it helpful to solve world issues?

Dr. Tie Liu: In my opinion, AI those are basically computational generates approach to deal with large datasets, either using modeling or using mathematic algorithms to predict any of the patterns for through those large data analyses. So that's how AI can be used, either use as model, modeling approach to predict any of the interesting traits or use or collect large datasets, and then identify algorithm to monitor those traits. 

Shenara: Can you share how AI has been applied in our research? 

Dr. Tie Liu: So in my researchers, we're not doing AI alone in my lab, is basically we're collaborating our colleagues into engineers or computational engineering department, though, there are collaborators of faculty in our department that are experts in terms of understanding those large datasets and use those datasets to predict any of the patterns by collaborating with our collaborators. We're collecting physiological or biochemistry data from the vegetable and fruit decay or senescence process. So then, when we have all this data collected, and we send it to our collaborators, and they were like, blindfolded to look at us large datasets and looking for is there any patterns among those datasets. Can they develop some methodology or algorithm to draw any interesting and conclusion from those datasets? So they are basically developing monitor tools to identify those traits.

Shenara: USDA aims to reduce food waste by 50% until 2030. Do you think your research, with your groups, can be a model for developing this type of analysis in other crops or even in the livestock industry?

Dr. Tie Liu: Yeah, I believe those kind of in the those cutting-edge technologies such as AI technologies are actually changing our crop productions or managing sort handling process. So, for long In conclusion, I think it really is important are can be used for applying industry for example, a lot of machine harvesting technique had or system has been used to crop the crop production system, for instance, to use the machine to harvesting strawberries, harvesting blueberries during their harvesting season. And you can imagine those technology probably going to significantly reduce the labor cost, going to reduced industry farms, you must have met in terms of harvesting dose, and then because the machine can harvest anytime in any any any conditions such as different than wheat or weathers and they're really can terms of reduce a lot of labor cost, and the same time can improve our crop production system. 

Shenara: In terms of your research, your project is called Fresh ID, right? What are the next step is offer research, but you can feel free like to give us a little bit of the historic about the project and then connect to the next steps.

Dr. Tie Liu: Sure. This is an interesting project that I collaborate with Dr. Alina Zare, who's a professor in Computational and Engineering Department. So in this collaborate project, we're basically use AI to characterizing shelf life we associate traits. For example, we were in a past years looking at broccoli senescence, we were looking same to avocado chilling injuries, we currently looking at lettuce decays or strawberry decays. So you can see that we're looking at different vegetables and fruit. The strategy for us is to focus on a particular post-harvest question first. For example, the broccoli yellowing: so this is an example when if you leave the broccoli in a refrigerator for too long, they are turning yellow. The main problem from this is because they are chlorophylls degraded during the storage conditions. So we’re using AI to examine those changes. In other words, during their broccoli storage, it is actually difficult to track when the broccoli turning yellow, right? So that's because when the broccoli is shifting those grocery store they may already stay in the field for a while and so then you have a several days before you can you you can cook the broccoli. So then what we want to know is what the actually physiological age of the broccoli so then we use to answer those questions we basically develop AI algorithm to testing how old of the broccoli when you're storing in the refrigerator. So in this project, my lab is characterizing different decay processes of broccoli and then, collect all the data's and Dr. Zare’s basically concludes those large datasets, identifies patterns that could associated with our physiological and biochemical indicators. So using this approach we can identify or use algorithm AI-based algorithms to characterize how old of those broccoli, right? So this is just an example we can we're characterizing the broccoli yellow. The same way we're also characterizing avocado chili injuries, this is also probably you notice when you keep the avocado in the fridge several days, especially often after they ripen and when you cut the avocado open they actually see a lot of dark spots and nobody really like those dark spots. So in this problem we are using AI to characterize when avocados actually turning those dark spots, forming those dark spots. So use AI algorithm or our AI-associated imaging tools were capture the whole process during the avocado chili injury process and that we use imaging to identify any of the early indicator to monitor those avocado decays. So that's you can see there two examples to show through we use fresh ID as a way to characterize the early senescence or decay process and to be able to monitor those decay process. 

Shenara: Is the project concluded or do you have other steps to continue with this project? 

Dr. Tie Liu: this is a USDA-funded project, it's from 2021, it will end it on 2025. So this four-year project is currently still ongoing. We have a look at different questions for instance, broccoli yellowing, avocado injuries, and now we're looking at the lettuce decays and passionfruit ripening, so we're keep looking at different post-harvest problem. And then the ultimate goal is to identify a one algorithm to characterize all different kinds of problems. So then we can really develop a robust stick and powerful tool to characterize those changes. 

Shenara: It's really interesting.

Dr. Tie Liu: Oh, thanks!

Shenara: Analyzing our research projects and those highlighted in your attention. Do you think they can be applied to a large-scale industry? 

Dr. Tie Liu: Yeah, that's a good question. It's definitely, it's our long-term goal, is we use the methods or system we developed in the lab can apply them in the industry and, as a matter of fact, a lot of. I'm not talking about my research. but there's a lot of other research has been already used in industries and one example is to use imaging-based AI technology to sort different colors of vegetables. I'm talking about a blueberry, for instance, in the blueberry industries when after harvesting, they’re actually all blue berries, going to all the large conveyors. In the conveyor, they actually have a sorter to sort blueberry based on the colors, right? So, this is a way you can identify size of the blueberries and identify different color of blueberries. So then you can sorting different blueberries into different categories, as now you can have a large size of the blueberries, you have ripen the blueberries or you have more ripened blueberries. So this way, you can actually have different increased efficiency in the packaging system. And those technologies associated is based on imaging or AI technology to monitor those color and sort the blueberry based on the color, based on the sides. And so, those technology has been already been applied in the industry. 

Shenara: Can you explain how AI technology is applied to the industry to reduce food waste? Do you have a few examples of AI scientific discoveries that can be applied nowdays? 

Dr. Tiu Lee: Yeah. So the example I provided the blueberry sortings. There are, of course, a lot of examples. There are two ways we can apply AI-associated technology in industry in terms of reducing food waste, and one is talking about pre-harvesting, which is how do we actually improve the production system? For example, how do we improve fertilizer usage? How do we improve the water use efficiencies or any of the pesticide applications? So then you can produce high-quality vegetables and fruit and, eventually, can reduce the waste of those vegetables, right? Because they either can improve their shelf life or improve their qualities and improve their likening. So then you can reduce the waste in the production side. So those are the AI-associate technologies that have been used in terms of managing pesticide applications, managing water use efficiency, and this is I'm talking about for instance, in drone associated those AI's basically phenotyping in the field and looking at how water usage has been applied in the field, and how can they associate technology monitor those pesticide applications or herbicide management, so then they can basically gather of those are and the reduce pesticide application so then you can improve the quality of vegetables. The other side is the postharvest levels. And so those are, for instance, more packagings those AI associated can identify newer technologies to increase their food packaging and transportation system and monitor those decay processes. For instance, one of the technologies called digital twins monitors those vegetables and or refrigeration conditions in how the quality change over time and use those technology or AI associate technology were able to quantify those decay processes and eventually can reduce food waste and loss. 

Shenara: what advice would you give for future researchers seeking AI to reduce food waste? 

Dr. Tie Liu: This is also a thing very few questions. I believe the collaboration is it is really the key. It is in a current world is multi-discipline and under research is can really helping team of different expert can can look at questions in different angles, and they're really solving the questions from a different areas. By collaboration we can take bomb was engineers can or by informaticians. And we design algorithms and design models to predict any of the biological questions. And those are collaborations really helping to solving those questions. For instance, in the pre-harvest field we're talking about we need expert in terms of understanding drones, understanding, ecologies, understanding how the computation approach can helping analyzing all the data sets. So then, people work in different fields to work together and really can forming a team to understanding those questions. And I think this is a really the current field in terms of people collaborate more and more weights expert in different expertise in to understand those questions. And this is probably you can see in the USDA project or NSF-funded projects, they really encourage multidisciplinary experts to working in the same questions. And then to understand the question from the from a micro levels to all the way a to macro levels, and then those can really helping from the local to the global kind of associated questions. 

Shenara: Oh, well, great. Thank you, Dr. Liu for sharing your knowledge with us today. 

Dr. Tie Liu: Thanks so much. Thanks for having me.

Shenara: And thank you for listen to the boundless science series on the Streaming Science Podcast. Check out our website and social media for more of our work. If you enjoyed this episode, we encourage you to tune in to the other episodes in our series. Once again, I am your host Shenara Ramadan. Thanks for listening. For more information about this episode, visit the links in our show notes.