Tel-Hai Magazine 2022

Can Artificial Intelligence Help Fight Food Loss?

Research by Prof. Ofer Shir and Dr. Dan Gamrasni

Artificial Intelligence (AI) technologies are increasingly recognized as game changers in almost every possible area in our lives – ranging from healthcare, secu rity, commerce, trade, to driving and gaming. However, AI remains very limit ed compared to human intelligence. Ma chines are still considered shortsighted and even stupid; the brightest AI systems to-date have less common sense than the typical house cat. In other words, ma chines may excel at learning certain problems and in conducting difficult and large-scale computations, and yet have no knowledge of how the world works. At the same time, AI constantly plays auxil iary roles in the discoveries and innova tions of scientists and engineers – primar ily in solving specific computational tasks, such as analyses by pattern recognition. A broadly accepted hypothesis is that AI will drive more decisions in future scien tific activities. Is this a scientist’s worst nightmare, or a dream come true? Prof. Ofer Shir has been investigating the ability of computers to drive scientific experimentation since his postdoctoral fellowship at Princeton University, USA. Back then, Prof. Shir’s algorithms en gaged with experiments in the Chemistry Department (teaching lasers how to con trol molecules, with Prof. Rabitz, and boosting the yield of organic syntheses, in Prof. MacMillan’s laboratory). Since joining Tel-Hai College and the MIGAL Institute in 2013, Prof. Shir has been striving to deploy his algorithms on experimental platforms in the research

laboratories at Tel-Hai/MIGAL. Prof. Shir’s first campaign took place in Dr. Dror Noy’s laboratory, where a computer attempted to learn laboratory protocols for producing a particular protein. That is, the computer proposed instructions concerning the sequence of operations, including their activation levels, in the targeted production process. These in structions were then implemented in the laboratory, and their outcomes were eval uated, and subsequently fed back into the computer. Consequently, the algorithm then proposed a new set of instructions, and the sequential experimentation pro ceeded in the same fashion. This cam paign attained a novel production proto col of an original nature, and obtained a production level comparable with the best human practice. This led to the real ization that such AI technology could break new ground, and would be a par ticularly good match for the Postharvest Innovation Center. Postharvest refers to the collection of practices for handling crops immediately following their harvest, with the explicit goal of maintaining their quality, while boosting their shelf-life. Postharvest tech nologies constitute a cornerstone of mod ern sustainability, and influence food se curity directly, with a potentially vast economic impact on the global food sup ply-chain. Nevertheless, they impose sig nificant scientific challenges concerning treatment protocols for fresh fruit and vegetables. The demand for affordable food supplies

is growing, in order to meet the increase in world population, as described by the Food and Agriculture Organization of the United Nations (FAO). A recent, or ganized effort to meet these food security demands is the Farm-to-Fork strategy of the European Union 1 . Sprang and co-workers reported 2 that postharvest losses might reach 20% to 30%. There fore, reduction of postharvest food loss would improve food security. Improved postharvest practices could increase food stability and availability, making food sys tems more efficient, and providing a solu tion to challenges the FAO has identified as significant. Dr. Dan Gamrasni, a Food Scientist and a Postharvest Researcher at Tel-Hai/MI GAL, captured the innovative opportuni ty in AI-driven experimentation, and decided to take a shot at it. He chose the cucumber fruit (Cucumis sativus L.) as the first experimental subject, to obtain proof-of-concept. The cucumber is a crop with a high economic value that constitutes a good source for antioxi dants, magnesium, vitamin C and dietary fiber. The cucumber fruit has a limited postharvest potential of less than 14 days in storage, due to weight loss, discolor ation of the peel, softening, fungal infec tions, and other known visual defects 3 . Cucumbers are very sensitive to chilling injuries, when stored at temperatures be low 7°C—10°C, and develop wa ter-soaked areas, pitting and accelerated decay. Hence, extending the postharvest shelf-life of cucumbers constitutes a sig nificant, real-world challenge, and it be

Tel-Hai Magazaine | 2022 12