Sridhar Mahadevan
Research Professor
Director, Adobe Research
Co-Director, Autonomous Learning Laboratory
Research Interests
Artificial Intelligence
Applied Category Theory
Machine Learning
CONTACT INFO:
firstname.i.lastname AT gmail
Administrative Assistant: Sarah Byrne
I have published over 150 research articles in AI/ML conferences and journals, and elected Fellow of AAAI in 2014 for "significant contributions to ML". My current research involves the application of higher-order category theory to AI and ML. Category theory abstracts many mathematical frameworks used in AI, from calculus to logic to probability and statistical approaches into one unified approach. Higher-order categories, such as simplicial sets and objects, generalize directed graphs, relational models, and ordinary categories, and enable the application of powerful mathematical techniques in homotopy theory to AI and ML.
Recent Preprints and Invited Talks
Sridhar Mahadevan, A Layered Architecture for Universal Causality, Arxiv, December 2022
Sridhar Mahadevan, Universal Representations for AI, Invited Talk, Department of Computer Science, Oregon State University, September 23rd, 2022
Sridhar Mahadevan, Unifying Causal Inference and Reinforcement Learning using Higher-order Category Theory, Arxiv, September 2022
Sridhar Mahadevan, Categoroids: Universal Conditional Independence, Arxiv, July 2022
Sridhar Mahadevan, On the Universality of Causal Diagrams and the Causal Reproducing Property, Arxiv, July 2022
Sridhar Mahadevan, Universal Decision Models , Invited Talk, Indian Institute of Science, October 23rd, 2021
Sridhar Mahadevan, Universal Decision Models, Arxiv, 2021
Sridhar Mahadevan, Causal Homotopy, Arxiv, 2021.
