CV | Google Scholar | Github | linkedin | medium

"Imagination is more important than knowledge" - Albert Einstein

1. Improving reasoning ability of LLMs and reducing reward hacking in reasoning LLMs.
2. Reasoning in diffusion LLMs.
3. Weak-to-strong generalization and self-improvement.
4. LLM/VML alignment, reasoning and reducing hallucinations. 
5. Reinforcement Learning and Multi-Agent Systems, Uncertainty Estimation.
   
   

UPDATES:

• Dec 17: Gave me Prelim Exam. My talk titled : Towards Reliable Reasoning and Alignment in Large Models is available here.

• 1 paper accepted at EMNLP'25 on unceratiny aware Answer Selection for multiple LLMs.

• 1 paper accepted at Neurips'25.

• Competed Fall'24-Spring'25 internship at Capital One working on reward hacking in reasoning LLMs. Stay Tuned for paper.

• Completed Summer'24 internship at Dolby where I worked on reducing hallucinations in Video LLMs!

• EnsemW2S accepted at SafeGenAI workshop at Neurips'24. Paper is now on arxiv! 

• Robustness to Multi-Modal Environment Uncertainty in MARL accepted at Neurips'23 Multi-Agent Security workshop.

• Amazon Internship paper accepted at Interspeech'23.

RESEARCH PUBLICATIONS

REASONING LLMs | Alignment | SuperAlignment in LLMs

1. Cut the Overcredit: Precision First Process Rewards for Reasoning LLMs.
   Aakriti Agrawal, Souradip Chakraborty, Armin Saghafian, Nihal Sharma, Rizal Fathony, Nam H Nguyen, C. Bayan Bruss, Amrit Singh Bedi, Furong Huang.
   In-review at ICLR, 2026

2. Uncertainty-Aware Answer Selection for Improved Reasoning in Multi-LLM Systems.
   Aakriti Agrawal, Rohith Aralikatti, Anirudh Satheesh, Amrit Singh Bedi, Furong Huang.
   Published 30th Empirical Methods in Natural Language Processing (EMNLP), China , 2025

3. EnsemW2S: Can an Ensemble of SoTA LLMs be Leveraged to Obtain a Stronger LLM?
   Aakriti Agrawal, Mucong Ding, Zora Che, Chenghao Deng,  Anirudh Satheesh, John Langford, Furong Huang.
   Published SafeGenAI @Neurips, 2024 | In-review @ AAAI, 2026

4. Easy2Hard-Bench: Standardized Difficulty Labels for Profiling LLM Performance and Generalization.
   Mucong Ding*, Chenghao Deng*, Jocelyn Choo, Zichu Wu, Aakriti Agrawal, Avi Schwarzschild, Tianyi Zhou, Tom Goldstein, John Langford, Anima Anandkumar, Furong Huang.
   Published at Neurips Dataset Track, 2024

Vision-LLMs

1. Towards Mitigating Hallucinations in Large Vision-Language Models by Refining Textual Embeddings.
   Aakriti Agrawal, Gouthaman KV, Rohith Aralikatti, Gauri Jagatap, Jiaxin Yuan, Vijay Kamarshi, Andrea Fanelli, Furong Huang.
   In-review @   | (PS: Email for details on the project or manuscript)

Robustness

1. WAVES: Benchmarking the Robustness of Image Watermarks.
   Bang An*, Mucong Ding*, Tahseen Rabbani*, Aakriti Agrawal, Yuancheng Xu, Chenghao Deng, Sicheng Zhu, Abdirisak Mohamed, Yuxin Wen, Tom Goldstein, Furong Huang.
   Published ICML 2024 | Paper 

2. PoisonedParrot: Subtle Data Poisoning Attacks to Elicit Copyright-Infringing Content from Large Language Models.
   Michael-Andrei Panaitescu-Liess, Pankayaraj Pathmanathan, Yigitcan Kaya, Zora Che, Bang An, Sicheng Zhu, Aakriti Agrawal, Furong Huang.
   Published at NAACL'25 | SafeGenAI @Neurips'24

3. Robustness to Multi-Modal Environment Uncertainty in MARL using Curriculum Learning.
   Aakriti Agrawal,  Rohith Aralikatti, Yanchao Sun, Furong Huang.
   Published MASEC@NeurIPS 2023  | Paper | Code 

Speech Processing

Learning when to trust which teacher for weakly supervised ASR.

 Aakriti Agrawal, Milind Rao, Anit Kumar Sahu, Gopinath (Nath) Chennupati, Andreas Stolcke.

Published Interspeech 2023 | Paper 

Multi-Agent Reinforcement Learning

1. Revisiting Parameter Sharing in Multi-Agent Deep Reinforcement Learning.
   J K Terry, Nathaniel Grammel, Sanghyun Son, Benjamin J Black, Aakriti Agrawal.
   Paper

2.  RTAW: An Attention Inspired Reinforcement Learning Method for Multi-Robot Task Allocation in Warehouse Environments.
   Aakriti Agrawal,  Amrit Singh Bedi,  Dinesh Manocha.
   ICRA 2023 | Paper | Code 

3. DC-MRTA: Decentralized Multi-Robot Task Allocation and Navigation in Complex Environments.
   Aakriti Agrawal,  Senthil Arul Hariharan,  Amrit Singh Bedi,  Dinesh Manocha.
   IROS 2022 | Paper  

Robotics

1. Accurate Estimation of 3D-Repetitive-Trajectories using Kalman Filter, Machine Learning and Curve-Fitting Method for High-speed Target Interception.
   Aakriti Agrawal, Aashay Bhise, Rohitkumar Arasanipalai, Lima Agnel Tony, Shuvrangshu Jana, Debasish Ghose.
   Book Chapter in : "Artificial Intelligence for Robotics and Autonomous Systems Applications" | Paper 

2. Mid-flight propeller failure detection and control of propeller-deficient quadcopter using reinforcement learning.
   Rohitkumar Arasanipalai*, Aakriti Agrawal*, Debasish Ghose | Paper

3. A Comparative Study of Noise Cancellation Using LMS Adaptive Filter and RNN Filter.
   Aakriti Agrawal, Rohitkumar Arasanipalai, B Sainath.
   ICEPE 2018 conference | Paper