Akansha Kalra

Akansha Kalra

I'm a PhD student in the CS program at University of Utah, advised by Dr. Daniel Brown. I graduated with an MS in Electrical and Computer Engineering from Carnegie Mellon University, where I was very fortunate to work under the guidance of Dr. Carlee Joe-Wong.

Education

Research Interests

  • Reinforcement Learning from Human Feedback (RLHF)
  • Human-AI Interaction and Alignment
  • Applied Reinforcement Learning
  • Interpretability and Explainability in Deep RL (XRL)

Selected Awards

  • 2023 - Won 1st place in Paradigm Challenge at Human-AI Teaming Hackathon organized by U.S. Army Research Lab.
  • 2022 - Awarded Graduate Fellowship from School of Computing, University of Utah.
  • 2018 - Awarded Honors degree in B.E. by Panjab University , given to top 5 students in the whole department for maintaining the highest GPA throughout 4 years consistently.
Basavasagar Patil, Akansha Kalra, , Daniel S Brown Under Review, ICLR 2025

Publications

How vulnerable is my learned policy? Adversarial attacks on modern behavioral cloning policies

Basavasagar Patil, Akansha Kalra, Guanhong Tao, Daniel S. Brown
Under Submission at ICLR 2025

Can Differentiable Decision Trees Enable Interpretable Reward Learning from Human Feedback?

Akansha Kalra, Daniel S. Brown
In Reinforcement Learning Journal (RLJ) | Presented at the Reinforcement Learning Conference (RLC), Amherst Massachusetts, August 9–12, 2024
PDF RLC Oral- Slides Website Code

Interpretable Reward Learning via Differentiable Decision Trees

Akansha Kalra, Daniel S. Brown
Oral Presentation at ML Safety Workshop, NeurIPS 2022
Refer to RLJ|RLC publication for extended version of this work.
PDF

Machine Learning on Volatile Instances: Convergence, Runtime, and Cost Tradeoffs

Xiaoxi Zhang, Jianyu Wang,Li-Feng LeeTom Yang, Akansha Kalra, Gauri Joshi, Carlee Joe-Wong
In IEEE/ACM Transactions on Networking 30.1 (2021): 215-228
PDF