WPI Projects Net Two of Seven State Grants in Competition Two WPI projects that harness artificial intelligence to advance clean technology have been selected as winners of the Massachusetts AI Models Innovation Challenge. The competitive grant program, established in February, is an initiative of the Massachusetts AI Hub, supported by Massachusetts Technology Collaborative. The challenge supports the development of artificial intelligence models to unlock breakthroughs in the state’s key industry sectors, including robotics, climate technology, and advanced manufacturing. The Awardees A team led by Michael Timko, William B. Smith Professor of Chemical Engineering and head of the Department of Chemical Engineering, was awarded $381,931 for the project “Machine Learning Digital Twins to Transform Waste to Renewable Energy.” The project seeks to help Massachusetts meet its goals of reducing municipal solid waste, which is the trash generated by homes, businesses, and institutions that ends up in landfills.  Timko’s project will develop a digital twin that simulates a complex chemical process known as hydrothermal liquefaction, which turns waste into energy. By utilizing data from thousands of experiments and machine learning, the digital twin predicts the outcomes of using hydrothermal liquefaction in less time and at less cost than running laboratory experiments or through trial and error. Timko believes the digital simulation will allow waste processors and companies to be better informed about and lower the investment risk associated with adopting sustainable approaches to energy generation. Other Department of Chemical Engineering faculty working on the project include Andrew Teixeira, associate professor, Nikolaos Kazantzis, professor, and Geoffrey Tompsett, assistant research professor. A team led by Berk Calli, associate professor in the Robotics Engineering Department, received $279,731 for the project “Automated Dataset Generation for Training High-Performance Classification and Segmentation Models in Industrial Recycling Applications.” The project intends to reduce the amount of waste going to landfills by increasing recycling through the development of a less complex and arduous sorting process. Calli’s team will develop an AI-powered robotic system to identify and properly gather materials for recycling at recovery facilities so that fewer materials become waste.  The system would learn to identify materials from video footage of workers sorting manually in these facilities. The automated learning means workers would not be required to go through the tedious process of manual labelling, which involves looking at images of items in the waste stream and then marking and classifying millions of them one by one. To achieve the goal of more accurate and efficient sorting of materials, the system will rely on workers to help it learn how to properly sort new types of materials it hasn’t seen before, and workers will also maximize sorting accuracy by processing items that robotic machinery cannot handle. Calli believes this approach could revolutionize recycling and advance efforts to reuse and regenerate materials in support of the circular economy. WPI undergraduate and graduate students will participate in these projects to gain hands-on experience in the development and application of artificial intelligence. The research efforts exemplify the university’s commitment to technological innovation and addressing societal challenges by taking knowledge developed in the laboratory and bringing it into the world. Galen Brown, PhD candidate in computer science, presenting recycling research project on behalf of Berk Calli's team at Massachusetts AI Hub awards ceremony The WPI projects were two of seven that won the state challenge. The winners were named at an awards ceremony in Boston on October 16, hosted by the Massachusetts AI Hub. Read more about the Massachusetts AI Models Innovation Challenge in this press release. Michael Timko presenting digital twin research project at Massachusetts AI Hub awards ceremony