Utkarsh Soni
Papers
3
Total Citations
31
H-Index
3
About
Utkarsh Soni is a leading researcher in explainable artificial intelligence (XAI), with a specific focus on making reinforcement learning (RL) agents transparent and trustworthy. His work addresses the critical "black box" problem of RL systems used in autonomous driving, robotics, and stock trading, where understanding agent decisions is a legal and ethical necessity. Soni’s major contribution lies in developing visual and personalized explanations for sequential decision-making. His most cited work, "Why? Why not? When? Visual Explanations of Agent Behaviour in Reinforcement Learning" (2022, 21 citations), pioneers methods to answer causal and counterfactual questions about agent actions. He further advanced the field by recognizing that "Not all users are the same" (2021, 6 citations), creating tailored explanations for different human operators rather than one-size-fits-all solutions. By bridging the gap between complex RL algorithms and human understanding, Soni’s research is foundational for deploying AI in high-stakes environments where human-AI collaboration is essential. His work is particularly notable for its emphasis on user-centric design, ensuring that even non-experts can interpret and trust autonomous agent behavior.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3