Artificial Intelligence platform for food surplus

A solution by Calculus Energy Limited submitted to Food poverty - managing surplus food

We propose an artificial intelligence (AI) based marketplace for trading food surplus. Participants will be able to upload an image of the food, along with its description. Non profit organisations will be able to access the webapp and request the food item. To overcome the cumbersome process of entering the food description, an AI model will be deployed to identify food from the uploaded image.

(Pitched: 21/12/2017)

One Page Summary


According to the government’s waste advisory body Wrap, food waste at a supermarket level - any edible food that remains unsold – stands at just 2%, whereas 17% of edible food surplus found in manufacturers and on farms is lost. The problem is worser in Coventry city, where, Child Poverty Action Group have identified that (November 2016 study), in October-December 2015, 19.51% of Coventry children overall are in relative poverty before housing costs and 30.96% after housing costs. This problem needs to be immediately addressed, lifting people on the breadline by providing them access to the most basic necessity to survive - food. We believe that recent advances in information technology and big data analysis can make this possible more than ever, by providing an accessible marketplace for retailers and local charities to trade surplus food. Therefore, in this work, we propose an innovative and highly scalable artificial intelligence (AI) based marketplace for trading food excess from retail markets. Participants will be able to upload an image of the surplus food, along with its description. Local charities and non profit organisations will be able to freely access the web application and request available food items, either for delivery or collection. Although this model is fairly straightforward, one of the major bottlenecks is that the process of entering details such as food description, expiry date, ingredients, manufacturer, calorific value etc for the food surplus, takes an enormous amount of manual effort, thereby reducing the efficiency of the marketplace and rendering the surplus food identification process non-real time. To overcome this cumbersome problem of entering potentially 1000's of items into the food description database, an AI model will be deployed to identify food from the uploaded image. Similarly, the expiry date of the food and the corresponding ingredients will be extracted from the image using artificial intelligence models. This artificial intelligence based neural network approach, is the only possible solution to make food surplus identification real time - without the need for manual intervention. Further, the resulting web app will be able to make fast, real time search and match based on user input for either the food description (food based search/match) or any available surplus nearest to their location (location based search/match). To render the marketplace sustainable, farmers will be able to participate in the marketplace, to collect unused surplus food which is deemed unfit for human consumption to be used as animal feed. In addition to this, the marketplace will promote social inclusion, by creating jobs amongst the disadvantaged. Typically, these jobs will include delivering surplus food to farmers.