Hi everyone, I hope you had a good summer. We will resume the regular AI seminar<https://nam04.safelinks.protection.outlook.com/?url=https%3A%2F%2Fengineering.oregonstate.edu%2FEECS%2Fresearch%2FAI-seminars&data=05%7C01%7Cai%40engr.orst.edu%7C472fefc46deb46d4300d08dbb266bdfe%7Cce6d05e13c5e4d6287a84c4a2713c113%7C0%7C0%7C638299926567899001%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=sem97GcVmMRNnPXXDnh8ykNtWeTmu85NCrn8D0HCjBc%3D&reserved=0> on Fridays at 1 pm in KEC 1001 starting from September 29'th. It is open to public. Grads, please register for AI 507 to get 1 credit. We will also have a special virtual seminar this week on Thursday, August 14, at 4 pm PST in KEC 1001 by Jonathan Ferrer Mestres, a research scientist at CSIRO, Australia. I hope to see you there. Zoom link: https://oregonstate.zoom.us/j/98684050301?pwd=ZzhianQxUFBPUmdYVWJKOFhaVURCQT09<https://nam04.safelinks.protection.outlook.com/?url=https%3A%2F%2Foregonstate.zoom.us%2Fj%2F98684050301%3Fpwd%3DZzhianQxUFBPUmdYVWJKOFhaVURCQT09&data=05%7C01%7Cai%40engr.orst.edu%7C472fefc46deb46d4300d08dbb266bdfe%7Cce6d05e13c5e4d6287a84c4a2713c113%7C0%7C0%7C638299926567899001%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=pMJbS9GHcbPDJPdxquLf%2F0fGRnB7IGiLwbkNKjO1PYA%3D&reserved=0> Title: Towards more interpretable solutions for conservation problems Abstract: Markov Decision Processes (MDPs) provide a convenient model for representing sequential decision-making optimization problems when the decision maker has complete information about the current state of the system and dynamics are non-deterministic. MDPs have been applied to help recover populations of threatened species under limited resources, to control invasive species, to perform adaptive management of natural resources, and to test behavioral ecology theories. These domains are human-operated systems, where MDP policies provide recommendations. Solutions computed for MDPs with thousands of states are difficult to understand. In human-operated systems, it is crucial that solutions provided by artificial intelligence algorithms can be interpreted and explained in order to increase uptake of MDP solutions. Explainable artificial intelligence, also known as the interpretability problem, aims to generate decisions in which one of the criteria is how easily a human can understand these decisions. We propose to increase the interpretability of MDPs by providing explainable artificial intelligence algorithms that can be used to solve conservation decision problems. We define the problem of solving K-MDPs, i.e., given an original MDP and a number of states (K), generate a reduced state space MDP that minimizes the difference between the original and reduced optimal solutions. Abstracting states aims to reduce the size of large state spaces by aggregating states which are equivalent given a metric. We found that K-MDPs can achieve a substantial reduction of the number of states with a small loss of performance on a number of case studies of increasing complexity from the literature. Bio: Jonathan is a Research Scientist within the Conservation Decisions Team, where he focuses on developing trustworthy and explainable artificial intelligence solutions for environmental decision-making. His primary goal is to ensure that artificial intelligence systems are equipped with transparent capabilities, fostering trust among users and experts who rely on these solutions. Jonathan's research centers on making informed decisions in the face of uncertainty to provide effective environmental solutions. Prior to his current role, he earned his PhD from Universitat Pompeu Fabra, in the Artificial Intelligence and Machine Learning Group in Barcelona in 2018, under the supervision of Dr. Hector Geffner. His doctoral thesis explored the integration of task and motion planning. Jonathan's educational journey also includes a Master's degree in Intelligent Interactive Systems and a Bachelor's degree in Computer Science. ------------------------------------------------------- Prasad Tadepalli Director, AI Program School of Electrical Engineering and Computer Science Oregon State University Corvallis, OR 97330
