DiscoveryBench: Towards Data-Driven Discovery with Large Language Models
Bodhisattwa P. Majumder*, Harshit Surana*, Dhruv Agarwal*, Bhavana Dalvi Mishra*, Abhijeetsingh Meena, Aryan Prakhar, Tirth Vora, Tushar Khot, Ashish Sabharwal, Peter Clark
International Conference on Learning Representations (ICLR), 2025
pdf | website

The first data-driven discovery benchmark containing 264 tasks collected across 6 diverse domains, such as sociology and engineering, with manually derived discovery workflows from published papers to approximate the real-world challenges faced by researchers.

DiscoveryBench: Towards Data-Driven Discovery with Large Language Models
Bodhisattwa P. Majumder*, Harshit Surana*, Dhruv Agarwal*, Bhavana Dalvi Mishra*, Abhijeetsingh Meena, Aryan Prakhar, Tirth Vora, Tushar Khot, Ashish Sabharwal, Peter Clark
International Conference on Learning Representations (ICLR), 2025
pdf | website

The first data-driven discovery benchmark containing 264 tasks collected across 6 diverse domains, such as sociology and engineering, with manually derived discovery workflows from published papers to approximate the real-world challenges faced by researchers.

Latent Factor Models Meets Instructions: Goal-conditioned Latent Factor Discovery without Task Supervision
Zhouhang Xie, Tushar Khot, Bhavana Dalvi Mishra, Harshit Surana, Julian McAuley, Peter Clark, Bodhisattwa P. Majumder
North American Chapter of the Association for Computational Linguistics (NAACL), 2025
pdf | website

Instruct-LF, a goal-oriented latent factor discovery system that integrates LLM's instruction-following ability with statistical models to handle large, noisy datasets where LLM reasoning alone falls short.

DiscoveryWorld: A Virtual Environment for Developing and Evaluating Automated Scientific Discovery Agents
Peter Jansen, Marc-Alexandre Côté, Tushar Khot, Erin Bransom, Bhavana Dalvi Mishra, Bodhisattwa P. Majumder, Oyvind Tafjord, Peter Clark
arXiv, 2024
pdf | website

The first virtual environment for developing and benchmarking an agent's ability to perform end-to-end novel discovery.

CLIN: A Continually Learning Language Agent for Rapid Task Adaptation and Generalization
Bodhisattwa P. Majumder, Bhavana Dalvi Mishra, Peter Jansen, Oyvind Tafjord, Niket Tandon, Li Zhang, Chris Callison-Burch, Peter Clark
Conference on Language Modeling (COLM), 2024
pdf | website

A novel non-parametric continual learning paradigm for rapid adaptation and generalization to unseen tasks and environments for language agents. We show a dynamic, persistent, semantic memory centered around causal abstractions significantly amplifies transfer and learning without any additional parameter update.

Few-shot Dialogue Strategy Learning for Motivational Interviewing via Inductive Reasoning
Zhouhang Xie, Bodhisattwa P. Majumder, Mengjie Zhao, Yoshinori Maeda, Keiichi Yamada, Hiromi Wakaki, Julian McAuley
Findings of Association for Computational Linguistics (ACL), 2024
pdf

A framework capable of learning and applying conversation strategies in the form of natural language inductive rules from expert demonstrations.

Tailoring with Targeted Precision: Edit-Based Agents for Open-Domain Procedure Customization
Yash Kumar Lal, Li Zhang, Faeze Brahman, Bodhisattwa P. Majumder,, Peter Clark, Niket Tandon
Findings of Association for Computational Linguistics (ACL), 2024
pdf

A simple architecture with two LLM agents used sequentially, one that edits a generic how-to procedure and one that verifies its executability, performs best in customising plans and procedural texts.

Data-driven Discovery with Large Generative Models
Bodhisattwa P. Majumder*, Harshit Surana*, Dhruv Agarwal*, Sanchaita Hazra, Ashish Sabharwal, Peter Clark
International Conference on Machine Learning (ICML), 2024
Position Paper
pdf

A practical first step toward an end-to-end automation for scientific discovery. We posit that Large Generative Models (LGMs) present an incredible potential for automating hypothesis discovery, however, LGMs alone are not enough.

Skill Set Optimization: Reinforcing Language Model Behavior via Transferable Skills
Kolby Nottingham, Bodhisattwa P. Majumder, Bhavana Dalvi Mishra, Sameer Singh, Peter Clark, Roy Fox
International Conference on Machine Learning (ICML), 2024
pdf | website

Skill Set Optimization improves LLM actors through constructing and refining sets of transferable skills. Leveraging environment reward signals, generalizable skills enable significant continual improvement for frozen LLM actors.

Tell, Don't Show!: Language Guidance Eases Transfer Across Domains in Images and Videos
Tarun Kalluri, Bodhisattwa P. Majumder, Manmohan Chandraker
International Conference on Machine Learning (ICML), 2024
pdf

A novel framework that utilizes readily available or easily acquired text descriptions to guide robust transfer of discriminative knowledge from labeled source to unlabeled target data with domain gaps.

Self-Supervised Bot Play for Transcript-Free Conversational Recommendation with Rationales
Shuyang Li, Bodhisattwa P. Majumder, Julian McAuley
ACM Transactions on Recommender Systems (TORS), 2024
pdf | slides

A conversational critiquing framework to provide feedback on rationales behind a recommendation and iteratively update underlying recommendation model for faster convergence to target predictions.

To Tell The Truth: Language of Deception and Language Models
Sanchaita Hazra, Bodhisattwa P. Majumder
North American Chapter of the Association for Computational Linguistics (NAACL), 2024
Oral presentation
pdf | code & data

We show there exists algorithmic predictors that can detect novel but accurate language cues in many cases where humans failed to detect deception, opening up the possibility of human-AI collaboration in ameliorating human's ability to detect lies.

InterFair: Debiasing with Natural Language Feedback for Fair Interpretable Predictions
Bodhisattwa P. Majumder, Zexue He, Julian McAuley
Empirical Methods in Natural Language Processing (EMNLP), 2023
Oral presentation
pdf | code

Fairness in debiasing (i.e., balancing task performance and bias mitigation) is subjective and difficult to learn from data. In an interactive setup, we enable users to provide feedback and achieve a better balance, supported by controllable explanations.

Aligning Language Models to User Opinions
EunJeong Hwang, Bodhisattwa P. Majumder, Niket Tandon
Findings of Empirical Methods in Natural Language Processing (EMNLP), 2023
pdf | code

We discover that, in addition to the typical approach of prompting LLMs with demographics and ideology for personalization, utilizing the most relevant past opinions from individual users enables the model to predict user opinions more accurately.

Self-Refine: Iterative Refinement with Self-Feedback
Aman Madaan, Niket Tandon, Prakhar Gupta, Skyler Hallinan, Luyu Gao, Sarah Wiegreffe, Uri Alon, Nouha Dziri, Shrimai Prabhumoye, Yiming Yang, Sean Welleck, Bodhisattwa P. Majumder, Shashank Gupta, Amir Yazdanbakhsh, Peter Clark
Conference on Neural Information Processing Systems (NeurIPS), 2023
pdf | website

Empirical evidence on a broad array of tasks incites promising research direction: LLMs can auto-heal for better outcomes without any supervised training, or RL, or human feedback.

Large Language Models as Zero-shot Conversational Recommenders
Zhankui He*, Zhouhang Xie*, Rahul Jha, Harald Steck, Dawen Liang, Yesu Feng, Bodhisattwa P. Majumder, Nathan Kallus, Julian McAuley
The Conference on Information and Knowledge Management, (CIKM), 2023
pdf | code & datasets

The largest-scale data (till date) for conversational recommendation systems, discovering new trends for context sensitivity when LLMs are used as recommenders.

Adversarially Detecting and Remedying Inconsistencies in Natural Language Explanations
Myeongjun Jang, Bodhisattwa P. Majumder, Julian McAuley, Thomas Lukasiewicz, Oana-Maria Camburu
Association for Computational Linguistics, Main (ACL), 2023
pdf | code

An adversarial framework shows even high-quality natural language explanation do not have necessarily low-level of inconsistencies. A remedy method is proposed that shows additional knowledge-grounding improves robustness.

Towards Factual and Informative Review Generation for Explainable Recommendation
Zhouhang Xie, Sameer Singh, Julian McAuley, Bodhisattwa P. Majumder
AAAI Conference on Artificial Intelligence (AAAI), 2023
pdf | code

A personalized self-rationalizing retrieve-generate framework for factually grounded reviews to explain rating and recommendation predictions with high attribution towards past reviews and informaitve keywords.

Controlling Bias Exposure for Fair Interpretable Predictions
Zexue He, Yu Wang, Julian McAuley, Bodhisattwa P. Majumder
Findings of Empirical Methods in Natural Language Processing (EMNLP), 2022
pdf | code

Current debiasing models may over-debias. With local explanations and interventional training, we establish the fair balance between debiasing and predictability for several classification and generation tasks.

Self-Supervised Bot Play for Transcript-Free Conversational Recommendation with Rationales
Shuyang Li, Bodhisattwa P. Majumder, Julian McAuley
ACM Conference on Recommendation Systems (RecSys), 2022
Highlights of ACM RecSys' 22; invited for ACM Transactions on Recommendation Systems
pdf | code | slides

A conversational critiquing framework to provide feedback on rationales behind a recommendation and iteratively update underlying recommendation model for faster convergence to target predictions.

Knowledge-grounded Self-rationalization via Extractive and Natural Language Explanations
Bodhisattwa P. Majumder, Oana-Maria Camburu, Thomas Lukasiewicz, Julian McAuley
International Conference on Machine Learning (ICML), 2022
Spotlight presentation
pdf | code | talk

A unified framework to map extractive rationales and abstractive natural language explanations (NLE) of ML Models using commonsense. We establish new state-of-the-art in NLE generation, rationale extraction and predictive task performance.

Achieving Conversational Goals with Unsupervised Post-hoc Knowledge Injection
Bodhisattwa P. Majumder, Harsh Jhamtani, Taylor Berg-Kirkpatrick, Julian McAuley
Association for Computational Linguistics, Main (ACL), 2022
Oral presentation
pdf | code | talk

A post-hoc knowledge-injection technique that first retrieves and selects a diverse set of relevant knowledge snippets and further inject them into an initial response from an exisiting dialog model. Enriching dialog responses at decoding time with external knowledge (without re-training the existing models) promotes achieving conversational goals.

Detect and Perturb: Neutral Rewriting of Biased and Sensitive Text via Gradient-based Decoding
Zexue He, Bodhisattwa P. Majumder, Julian McAuley
Findings of Empirical Methods in Natural Language Processing (EMNLP), 2021
pdf | code

A rewriting framework that first detects sensitive components from input text and then perturbs the generation model at decoding time under a neutralizing constraint. No parallel corpus of sensitive-neutral texts is needed for training.

ReZero is All You Need: Fast Convergence at Large Depth
Thomas Bachlechner*, Bodhisattwa P. Majumder*, Henry Mao*, Gary Cottrell, Julian McAuley
Uncertainty in Artificial Intelligence (UAI), 2021
Oral presentation
pdf | code | slides

A novel deep neural network architecture that initializes an arbitrary layer as the identity map (ReZero), using a single additional learned parameter per layer to facilitate very deep signal propagation.

Unsupervised Enrichment of Persona-grounded Dialog with Background Stories
Bodhisattwa P. Majumder, Taylor Berg-Kirkpatrick, Julian McAuley, Harsh Jhamtani
Oral presentation
Association for Computational Linguistics, Main (ACL), 2021
pdf | code | slides

An unsupervised gradient-based rewriting framework to adapt potential background stories to an existing persona-grounded dialog. We constrain the generation for self-consistency with persona and promote its adherence to the story.

The GEM Benchmark: Natural Language Generation, its Evaluation and Metrics
Bodhisattwa P. Majumder as a part of GEM team
GEM workshop, Association for Computational Linguistics (ACL), 2021
pdf | website

GEM is a community-driven effort with the goal to improve how progress in natural language generation is measured. As a shared task in ACL 2021, we invite for challenge set submissions for 11 datasets and 7 languages in various NLG challenges.

Ask what's missing and what's useful: Improving Clarification Question Generation using Global Knowledge
Bodhisattwa P. Majumder, Sudha Rao, Michel Galley, Julian McAuley
North American Chapter of the Association for Computational Linguistics (NAACL), 2021
Oral presentation
pdf | code | talk

A two-stage framework that 1) estimates missing information from the global knowledge of similar contexts, and 2) conditionally generates useful questions using gradient-based decoding based on a usefulness scorer at the inference time. This work was done during an internship at Microsoft Research.

Like hiking? You probably enjoy nature: Persona-grounded Dialog with Commonsense Expansions
Bodhisattwa P. Majumder, Harsh Jhamtani, Taylor Berg-Kirkpatrick, Julian McAuley
Empirical Methods in Natural Language Processing (EMNLP), 2020
Oral presentation
pdf | code | slides

A variational learning framework to capture commonsense implications of input persona in a persona-grounded dialog agent using richer expansions obtained from existing commonsense knowledge bases.

Interview: Large-scale Modeling of Media Dialog with Discourse Patterns and Knowledge Grounding
Bodhisattwa P. Majumder*, Shuyang Li*, Jianmo Ni, Julian McAuley
Empirical Methods in Natural Language Processing (EMNLP), 2020
Oral presentation
pdf | code | data

The first large-scale analysis of discourse in media dialog ("Interview" - 105K conversations) and its impact on generative modeling of dialog turns, with a focus on interrogative patterns and use of external knowledge.

Bernard: A Stateful Neural Open-domain Socialbot
Bodhisattwa P. Majumder, Shuyang Li, Jianmo Ni, Henry Mao, Sophia Sun, Julian McAuley
Proceedings of Alexa Prize, Amazon, 2019-20
pdf

A framework for an engaging open-domain socialbot with a stateful autonomous dialog manager using non-deterministic finite automata to control multi-turn conversations. This work was done for Alexa Prize 2019.

Representation Learning for Information Extraction from Form-like Documents
Bodhisattwa P. Majumder, Navneet Potti, Sandeep Tata, James Wendt, Qi Zhao, Marc Najork
Association for Computational Linguistics (ACL), 2020
Oral presentation
pdf | blog | slides

A novel approach to learn interpretable representations for target fields using spatial and contextual knowledge for extracting structured information from form-like document images, even with unseen templates. This work was done at Google AI as a part of 2019 summer internship.

Generating Personalized Recipes from Historical User Preferences
Bodhisattwa P. Majumder*, Shuyang Li*, Jianmo Ni, Julian McAuley
Empirical Methods in Natural Language Processing (EMNLP), 2019
pdf | code | data | poster

Media coverage: Science Node, UCSD CSE News, UCSD JSOE News

A new task of personalized recipe generation to help these users: expanding a name and incomplete ingredient details into complete natural-text instructions aligned with the user's historical preferences.

Improving Neural Story Generation by Targeted Common Sense Grounding
Henry Mao, Bodhisattwa P. Majumder, Julian McAuley, Gary Cottrell
Empirical Methods in Natural Language Processing (EMNLP), 2019
pdf | code

A multi-task learning scheme to achieve quantitatively better common sense reasoning in language models by leveraging auxiliary training signals from datasets designed to provide common sense grounding.

Upcycle Your OCR: Reusing OCRs for Post-OCR Text Correction in Romanised Sanskrit
Amrith Krishna, Bodhisattwa P. Majumder, Rajesh S. Bhat, Pawan Goyal
Conference on Computational Natural Language Learning (CoNLL), 2018
pdf | code+data | supplementary

A state-of-the-art approach towards post-OCR text correction for digitising texts in Romanised Sanskrit. This work was done in a collaboration with CNeRG.

An 'Eklavya' approach to learning Context Free Grammar rules for Sanskrit using Adaptor Grammar
Amrith Krishna, Bodhisattwa P. Majumder, Anil K. Boga, Pawan Goyal
World Sanskrit Conference, 2018
pdf

A non-parametric Bayesian approach for learning (Probabilistic) Context Free Grammar productions for Sanskrit language at word-level supervised tasks such as compound type identification, identification of source and derived words from the corpora for derivational nouns and sentence-level structured prediction. This work was done at CNeRG.

Deep Recurrent Neural Networks for Product Attribute Extraction in eCommerce
Bodhisattwa P. Majumder*, Aditya Subramanian*, Abhinandan Krishnan, Shreyansh Gandhi, Ajinkya More
Preprint, arXiv, 2017
pdf | system description | video

We demonstrate the potential of neural recurrent structures in product attribute extraction by improving overall F1 scores, as compared to the previous benchmarks. This has made Walmart e-commerce achieve a significant coverage of important facets or attributes of products. This work from Walmart Labs later followed by a US patent.

Distributed Semantic Representations of Retail Products based on Large-scale Transaction Logs
Bodhisattwa P. Majumder*, Sumanth S Prabhu*, Julian McAuley
2018
report

We processed 18 million transactions consisting of unique 325,548 products from 1,551 categories to obtain vector representations which preserve product analogy. These representations were effective in identifying substitutes and complements. This work was done at Walmart Labs.

Lolcats meet Philosoraptors - What's in a 'meme'? Understanding the Dynamics of Image Macros in Social Media
Bodhisattwa P. Majumder, Amrith Krishna, Unni Krishnan, Anil K. Boga, Animesh Mukherjee
Preprint, arXiv, 2018
pdf | slides

How similar are the dynamics of meme based communities to that of text based communities? We try to explain the community dynamics by categorising each day based on temporal variations in the user engagement. Work done at CNeRG.

• A System for Information Extraction From Form-Like Documents, Google, 2020
• REDCLAN - RElative Density based CLustering and Anomaly Detection, Wal-mart, 2018
• Automated Extraction of Product Attributes from Images, Wal-mart, 2018
• System and Method for Product Attribute Extraction Using a Deep Recurrent System, Wal-mart, 2017
• Analytical Determination of Competitive Interrelationship between Item Pairs, Wal-mart, 2017