Home - Chirag Agarwal

I am an Assistant Professor in the School of Data Science at UVA. I lead the Aikyam lab and am the founder of Agyeya Artificial IQ Foundation, which focuses on democratizing research by supporting open collaboration in ML research and training the next generation of AI students in India. What inspires me to do what we do is the feeling of helping the community through my work! Feel free to connect with me if you are excited about Trustworthy AI research or its applications.

I am recruiting students for Fall 2025. Please check Open Positions for more details and my mentoring philosophy. For more details about me and my research, please check out my CV.

Welcome to Aikyam Lab (Sanskrit: ऐक्यम्; meaning "oneness")! We focus on developing Trustworthy Machine Learning Frameworks that go beyond training models for specific downstream tasks and satisfy trustworthy properties, such as explainability, fairness, and robustness.

Latest News

07/16/2024: On the Trade-offs between Adversarial Robustness and Actionable Explanations is accepted to AIES, 2024!
07/10/2024: Our paper on Certifying Robustness for LLMs is accepted to COLM, 2024!
07/01/2024: Check out the new pre-print on On the Hardness of Faithful Chain-of-Thought Reasoning in Large Language Models.
05/14/2024: Check out the new pre-print on Towards Safe Large Language Models for Medicine.
05/01/2024: Our paper on Iterative Prompting for Truthfulness is accepted at ICML, 2024!
03/14/2024: Our position pre-print on Faithfulness vs. Plausibility: On the (Un)Reliability of Explanations from LLMs is now out!
01/23/2024: Tutorial on Explainability in Graph Deep Learning for Biomedicine accepted at ISMB 2024.
01/19/2024: Our paper on Quantifying Uncertainty in Natural Language Explanations of Large Language Models is accepted to AISTATS'24!
11/21/2023: Selected as the Top Reviewer for NeurIPS'2023!
11/20/2023: We won the Harvard Data Science Initiative (HDSI) Azure Credits Award!!
10/27/2023: Papers accepted at NeurIPS XAIA and Ro-FoMo (Spotlight!) workshops!
09/13/2023: We won the Adobe Data Science Research Award.
09/07/2023: Our preprint on Certified Robustness for LLMs is now out.
08/15/2023: Our workshop on "Regulatable ML" at NeurIPS'23 is accepting submissions!
07/30/2023: Tutorial on "Training the Next-Generation of AI Students" at Excellence School.
07/01/2023: Papers accepted at ICML'23 ES-FoMo, Counterfactuals in Minds and Machines, and SCIS workshops!
04/05/2023: Explain Like I am BM25 is accepted to SIGIR'2023.
03/18/2023: Evaluating Explainability for Graph Neural Networks is accepted in Nature Scientific Data, 2023.
02/27/2023: Super excited that DeAR (Debiasing Vision-Language Models using Additive Residuals) is accepted to CVPR'23!
01/21/2023: Our papers on Graph Unlearning and Explaining RL Decisions with Trajectories accepted at ICLR, 2023.
12/09/2022: EXPASS gets accepted at LOG'22.
10/21/2022: Trajectory-based Explainability Framework for Offline RL is accepted to Offline RL Workshop NeurIPS'22.
09/16/2022: OpenXAI gets accepted at NeurIPS'22 and is live on GitHub.
09/02/2022: Invited Talk in the TrustML Young Scientists Seminars at RIKEN-AIP.
07/15/2022: Invited Talk on OpenXAI at Auburn University
07/08/2022: Invited Talk on "XAI: Challenges, Solutions, and the Way Forward" at Adobe Research, Bangalore
06/24/2022: Selected as a Lecturer for AI Summer School (AISS)'22 at IIIT-Delhi.
04/05/2022: Papers accepted at ICLR'22 Pair2Struct workshop and WWW'22 workshop on Graph Learning Benchmarks
03/10/2022: Selected as a Mentor for the LOGML Summer School 2022
03/03/2022: VoG paper accepted to CVPR 2022.
01/24/2022: Joined Adobe as a Research Scientist.
01/18/2022: 2 papers accepted at AISTATS 2022.
07/27/2021: New preprint Towards a Rigorous Theoretical Analysis and Evaluation of GNN Explanations on arXiv
07/27/2021: Our paper Towards a Unified Framework for Fair & Stable Graph Representation Learning is accepted at UAI'21
07/21/2021: 3 papers accepted at ICML'21 workshop on Theoretic Foundation, Criticism, and Applications Trends of Explainable AI, 1 paper at the ICML'21 workshop on Socially Responsible Machine Learning, 1 paper at the ICML'21 workshop on Interpretable Machine Learning in Healthcare, and 1 paper at the ICML'21 workshop on Uncertainty Robustness in Deep Learning.
07/21/2021: Our paper Towards Unification & Robustness of Perturbation and Gradient Based Explanations is accepted at ICML'21
03/04/2021: Presented a tutorial at FAccT with Shalmali Joshi and Himabindu Lakkaraju on limitations of explanability methods in ML. Video: here & Slides: here.
03/02/2021: Guest Lecture in Topics in Machine Learning-Interpretability and Explainability course at Harvard
10/09/2020: Moved to Harvard as a HMS Postdoctoral Fellow
09/18/2020: Paper accepted to ACCV'20 on Explaining image classifiers by removing input features using generative models
07/08/2020: Two papers accepted at Workshop on Human Interpretability in Machine Learning (WHI) ICML 2020
06/22/2020: Defended my Ph.D.!
04/12/2020: Paper accepted to CVPR Workshop on Fair, Data Efficient and Trusted Computer Vision 2020 on Sensitivity of Attribution Methods
02/26/2020: Oral paper accepted to CVPR'20 on SAM: The Sensitivity of Attribution Methods to Hyperparameters