Hi, I'm Aniket Vashishtha — a first-year thesis-based Master's student in Computer Science at the University of Illinois Urbana-Champaign, where I'm fortunate to be advised by Prof. Hao Peng.

My research lies at the intersection of Large Language Models (LLMs), Causality, Reinforcement Learning, Fairness, and Cognitive Science. I'm particularly interested in understanding how LLM-based reasoning can be applied in real-world, high-stakes scenarios and I approach this problem through the lens of causality.

I explore how causal reasoning can both leverage and improve LLMs, enabling more robust and generalizable systems. My work focuses on understanding how to optimally utilize experts like LLMs (or humans) as domain experts for uncovering causal relationships for downstream tasks like effect estimation or counterfactual reasoning. I'm also drawn to the cognitive science perspective — investigating how concepts like human reasoning and inductive biases can shed light on the limitations of language models, and guide us toward building better ones. A major theme in my work has been training language models using synthetic data, with a focus on improving generalization to out-of-distribution settings.

Before UIUC, I was a Predoctoral Research Fellow at Microsoft Research India, where I was fortunate to be mentored by Amit Sharma and Prof. Vineeth N. Balasubramanian. I still continue to be mentored by them. At MSR, we worked on practical applications of LLMs for causal reasoning, including causal graph discovery, effect inference, and axiomatic causal reasoning of language models.

Earlier, I also had the opportunity to contribute to projects in Fairness and Responsible AI, working with Prof. Monojit Choudhury and Dr. Sunayana Sitaram. Our work focused on socio-cultural and multilingual challenges, aiming to extend RAI techniques to underrepresented languages, particularly in the Indian context.

Outside of research, I've been a passionate beatboxer for over 10 years, winning several national championships. I also enjoy working out, cooking, watching anime, movies, and stand-up comedy.

I graduated with a B.Tech (Hons.) in Information Technology from Guru Gobind Singh Indraprastha University, Delhi, India in 2022. For more details, check my CV or hit me up on my email.

Updates

Apr 2025: Presenting my work Causal Order: Leveraging Imperfect Experts for Causal Inference at ICLR'25 main conference.
Apr 2025: Invited to give a talk on my work on LLMs and Causality at CISPA, Germany.
Apr 2025: Amit Sharma discusses our work on Causal Axiomatic Training on Aleksander Molak's Podcast.
Mar 2025: Serving as an emergency reviewer at ACL'25.
Feb 2025: Reviewer for Reasoning and Planning and Re-Align workshops at ICLR 2025.
Jan 2025: Our paper Causal Order: Leveraging Imperfect Experts for Causal Inference accepted at ICLR'25.
Nov 2024: Gave a talk at Causal Data Science Summit on Teaching Transformers Causal Reasoning through Axiomatic Framework, alongside leading researchers from causality, economics, and language models.
Oct 2024: Paper accepted as Oral at CaLM workshop at NeurIPS 2024.
Sep 2024: Received the prestigious JN Tata Scholarship.
Aug 2024: Joined UIUC as a Thesis Masters student in Computer Science, and started working with Prof. Hao Peng.
Jul 2024: Presented our work on Axiomatic Training for Causal Reasoning at MSR reading group.
Jul 2024: Released our work Teaching Transformers Causal Reasoning through Axiomatic Training on arXiv. Work was well-received online (tweet) and covered by tech accounts (coverage).
Feb 2024: Gave an oral talk in-person at AAAI LLM-CP workshop in Vancouver, Canada.
Sep 2023: Invited to lead a 2-day workshop in Kerala.
May 2023: Paper on Evaluating Gender Biases in Multilingual Settings accepted at ACL'23 Findings.
Apr 2023: Started working with Dr. Amit Sharma and Prof. Vineeth N. Balasubramanian.
Jan 2023: Our paper Performance and Risk Trade-offs for Multi-word Text Prediction at Scale accepted at EACL 2023 Findings.
Oct 2022: Team won first place at Microsoft Turing Hackathon.
Jan 2022: Started interning in the NLP team at Microsoft Research India with Dr. Monojit Choudhury and Dr. Sunayana Sitaram, focusing on Responsible AI problems.

Publications

Causal Order: Leveraging Imperfect Experts for Causal Inference
Aniket Vashishtha , Abbavaram Gowtham Reddy, Abhinav Kumar, Saketh Bachu, Vineeth N Balasubramanian, and Amit Sharma
ICLR'25 | International Conference on Learning Representations
pdf| abstract|

Teaching Transformer Causal Reasoning through Axiomatic Reasoning
Aniket Vashishtha, Abhinav Kumar, Atharva Pandey, Abbavaram Gowtham Reddy, Kabir Ahuja, Vineeth N Balasubramanian, Amit Sharma
LLM Reasoning and Planning Workshop @ ICLR'25
pdf| abstract|

On Evaluating and Mitigating Gender Biases in Multilingual Settings
Aniket Vashishtha*, Kabir Ahuja*, and Sunayana Sitaram (* = Equal Contribution)
ACL'23 Findings | Annual Conference of the Association for Computational Linguistics
pdf| abstract|

Performance and Risk Trade-offs for Multi-word Text Prediction at Scale
Aniket Vashishtha, S Sai Prasad, Payal Bajaj, Vishrav Chaudhary, Kate Cook, Sandipan Dandapat, Sunayana Sitaram, Monojit Choudhury
EACL'23 Findings | European Chapter of the Association for Computational Linguistics
pdf| abstract|

Mining Trends of COVID-19 Vaccine Beliefs on Twitter With Lexical Embeddings: Longitudinal Observational Study
Aniket Vashishtha*, Harshita Chopra*, Ridam Pal, Ashima Ashima, Ananya Tyagi, Tavpritesh Sethi (* = Equal Contribution)
JMIR | JMIR Infodemiology
pdf| abstract| cite

VacSIM: Learning Effective Strategies for COVID-19 Vaccine Distribution using Reinforcement Learning
Raghav Awasthi, Keerat Kaur Guliani, Saif Ahmad Khan, Aniket Vashishtha , Mehrab Singh Gill, Arshita Bhatt, Aditya Nagori, Aniket Gupta, Ponnurangam Kumaraguru, Tavpritesh Sethi
Intelligence-Based Medicine | Elsevier Journal
pdf| abstract|