Pranjal Aggarwal

Ph.D. Student@Carnegie Mellon University

prof_pic2.jpg

I am a 1st year Ph.D. student in Language Technologies at the School of Computer Science, Carnegie Mellon University, advised by Prof. Sean Welleck. My research interests primarily focus on code generation and formal verification, inference time search for reasoning, and multimodal agents.

Currently, I’m working on developing methods for generating formally verified code using iterative refinement techniques, employing inference-time search and reinforcement learning algorithms. I’m also exploring the development of multimodal code agents capable of interacting with unified action spaces, including images and keyboard actions over long horizon tasks.

Previously, I completed my B.Tech. and M.Tech. in Computer Science at the Indian Institute of Technology Delhi (IITD), where I graduated with a GPA of 9.4. During my time at IITD, I worked on various research projects, with topics ranging from improving reasoning in deep learning models, making large scale systems efficient and demystifying advancements in machine learning and natural language generation in general. I have been fortunate to work with Prof. Mausam, Prof. Karthik Narasimhan, Prof. Chetan Arora in past.

My work has been recognized through several awards, including winning the Model Attribution Challenge at SaTML 2023, securing first place in the Tower Research Capital Data Challenge, and receiving travel grants from Google and Microsoft Research for academic conferences, with our work features in NYT[], ….

I am passionate about advancing the field of artificial intelligence and natural language processing, and I’m always excited to collaborate on new and challenging projects! Feel free to reach out to me at my email!

news

Sep 24, 2024 Our Work: “Automix: Automatically mixing Language Models’ accepted in Neurips’24. Check it here.
Aug 26, 2024 Excited to begin my Ph.D. journey in LTI, SCS at Carnegie Mellon University, advised by Prof. Sean Welleck!
May 22, 2024 Our paper “GEO: Generative Engine Optimization” has been accepted at KDD’24 in Barcelona! Find it here. We introduce a novel task of content optimization for emerging Generative Engines.
Oct 7, 2023 Our Paper: “Let’s Sample Step by Step: Adaptive-Consistency for Efficient Reasoning & Coding with LLMs’ accepted in Empirical Methods in Natural Language Processing ‘23 (Oral).
Apr 28, 2023 Our work SemSup-XC is accepted at ICML 2023. SemSup-XC achieves state-of-the-art results on extreme classification.
Dec 22, 2022 Presenting Winning Solution for MLMAC (ML Model Attribution Challenge) at Secure & Trustworthy Machine Learning’23.

selected publications

2024

  1. automix_teaser2.png
    AutoMix: Automatically Mixing Language Models
    Pranjal* Aggarwal, Aman* Madaan, Ankit Anand, and 10 more authors
    In To appear in Neural Information Processing Systems, 2024, 2024
  2. geo_teaser.png
    GEO: Generative Engine Optimization
    Pranjal Aggarwal, Vishvak Murahari, Tanmay Rajpurohit, and 3 more authors
    In Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, 2024
  3. rlhf_teaser.png
    Demystifying Reinforcement Learning with Human Feedback
    Pranjal* Aggarwal, Shreyas* Chaudhari, Khanh Nguyen, and 4 more authors
    In Under Review, 2024

2023

  1. AC_teaser.png
    Let’s Sample Step by Step: Adaptive-Consistency for Efficient Reasoning and Coding with LLMs
    Pranjal Aggarwal, Aman Madaan, Yiming Yang, and 1 more author
    In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing (Oral), Dec 2023
  2. SemSup_XC.png
    SemSup-XC: semantic supervision for zero and few-shot extreme classification
    Pranjal Aggarwal, Ameet Deshpande, and Karthik Narasimhan
    In Proceedings of the 40th International Conference on Machine Learning, Dec 2023