About

Niranjan Balasubramanian received his PhD from the University of Massachusetts, Amherst, where he was a part of the Center for Intelligent Information Retrieval (CIIR). Before his PhD studies, he completed his Masters degree in Computer Science at the University of Buffalo in 2003. Prior to joining the computer science department at Stony Brook University in spring 2015, he was a post-doctoral researcher in the Turing Center at the University of Washington in Seattle. His research is motivated by the challenge of building systems that can extract, understand, and reason with information present in natural language texts. His interests are in two broad areas: Natural Language Processing (NLP) and information retrieval. He has worked on various projects including Question Answering at a 4th Grade Level, Event Schema Generation from news stories, Machine Learning for Information Retrieval, Energy-efficient Mobile Search, and Automatic Wikipedia Pages.

Research topics

• Artificial Intelligence
• Computer Science
• Natural Language Processing
• Machine Learning
• Information Retrieval
• Electrical engineering
• World Wide Web
• Engineering
• Psychology
• Social psychology
• Programming language
• Data science

Selected publications

• Enabling Efficient SpMM for Sparse Attention on GEMM-Optimized Hardware with Block Aggregation

  2026-02-05

  articleOpen access

  Rapidly growing context lengths have amplified the inherent sparsity in the attention mechanism of popular Large Language Models. However, the dynamic data access patterns required by sparse attention are challenging to realize using static data paths, leading to execution inefficiency. Existing SpMM hardware acceleration techniques address these inefficiencies by dynamically configuring data paths to align with the unstructured data access patterns of sparse attention. However, these approaches are not applicable to GEMM-optimized hardware, where dynamic data paths would introduce unacceptable hardware complexity and frequency degradation.

  PublisherDOI

• Syntax Is Easy, Semantics Is Hard: Evaluating LLMs for LTL Translation

  arXiv (Cornell University) · 2026-04-08

  preprintOpen access

  Propositional Linear Temporal Logic (LTL) is a popular formalism for specifying desirable requirements and security and privacy policies for software, networks, and systems. Yet expressing such requirements and policies in LTL remains challenging because of its intricate semantics. Since many security and privacy analysis tools require LTL formulas as input, this difficulty places them out of reach for many developers and analysts. Large Language Models (LLMs) could broaden access to such tools by translating natural language fragments into LTL formulas. This paper evaluates that premise by assessing how effectively several representative LLMs translate assertive English sentences into LTL formulas. Using both human-generated and synthetic ground-truth data, we evaluate effectiveness along syntactic and semantic dimensions. The results reveal three findings: (1) in line with prior findings, LLMs perform better on syntactic aspects of LTL than on semantic ones; (2) they generally benefit from more detailed prompts; and (3) reformulating the task as a Python code-completion problem substantially improves overall performance. We also discuss challenges in conducting a fair evaluation on this task and conclude with recommendations for future work.

  PublisherDOI

• Addressing the Ecological Fallacy in Larger LMs with Human Context

  Open MIND · 2026-03-06

  preprintSenior author

  Language model training and inference ignore a fundamental linguistic fact -- there is a dependence between multiple sequences of text written by the same person. Prior work has shown that addressing this form of \textit{ecological fallacy} can greatly improve the performance of multiple smaller (~124M) GPT-based models. In this work, we ask if addressing the ecological fallacy by modeling the author's language context with a specific LM task (called HuLM) can provide similar benefits for a larger-scale model, an 8B Llama model. To this end, we explore variants that process an author's language in the context of their other temporally ordered texts. We study the effect of pre-training with this author context using the HuLM objective, as well as using it during fine-tuning with author context (\textit{HuFT:Human-aware Fine-Tuning}). Empirical comparisons show that addressing the ecological fallacy during fine-tuning alone using QLoRA improves the performance of the larger 8B model over standard fine-tuning. Additionally, QLoRA-based continued HuLM pre-training results in a human-aware model generalizable for improved performance over eight downstream tasks with linear task classifier training alone. These results indicate the utility and importance of modeling language in the context of its original generators, the authors.

  DOI

• Teaching and Evaluating LLMs to Reason About Polymer Design Related Tasks

  ArXiv.org · 2026-01-22

  articleOpen accessSenior author

  Research in AI4Science has shown promise in many science applications, including polymer design. However, current LLMs are ineffective in this problem space because: (i) most models lack polymer-specific knowledge, and (ii) existing aligned models have limited coverage of knowledge and capabilities relevant to polymer design. Addressing this, we introduce PolyBench, a large-scale training and test benchmark dataset of more than 125K polymer design-related tasks, leveraging a knowledge base of more than 13 million data points obtained from experimental and synthetic data sources to ensure broad coverage of polymers and their properties. For effective alignment using PolyBench, we introduce a knowledge-augmented reasoning distillation method that augments this dataset with structured CoT. Furthermore, tasks in PolyBench are organized from simple to complex analytical reasoning problems, enabling generalization tests and diagnostic probes across the problem space. Experiments show that small language models (SLMs) with 7B to 14B parameters, trained on PolyBench, outperform similar-sized models and remain competitive with closed-source frontier LLMs on PolyBench's test dataset, while demonstrating performance gains on external polymer benchmarks. Dataset and associated code available at https://github.com/StonyBrookNLP/PolyBench.

  PublisherOA PDF

• Addressing the Ecological Fallacy in Larger LMs with Human Context

  ArXiv.org · 2026-03-06

  articleOpen accessSenior author

  Language model training and inference ignore a fundamental linguistic fact -- there is a dependence between multiple sequences of text written by the same person. Prior work has shown that addressing this form of \textit{ecological fallacy} can greatly improve the performance of multiple smaller (~124M) GPT-based models. In this work, we ask if addressing the ecological fallacy by modeling the author's language context with a specific LM task (called HuLM) can provide similar benefits for a larger-scale model, an 8B Llama model. To this end, we explore variants that process an author's language in the context of their other temporally ordered texts. We study the effect of pre-training with this author context using the HuLM objective, as well as using it during fine-tuning with author context (\textit{HuFT:Human-aware Fine-Tuning}). Empirical comparisons show that addressing the ecological fallacy during fine-tuning alone using QLoRA improves the performance of the larger 8B model over standard fine-tuning. Additionally, QLoRA-based continued HuLM pre-training results in a human-aware model generalizable for improved performance over eight downstream tasks with linear task classifier training alone. These results indicate the utility and importance of modeling language in the context of its original generators, the authors.

  PublisherOA PDF

• Teaching and Evaluating LLMs to Reason About Polymer Design Related Tasks

  arXiv (Cornell University) · 2026-01-22

  preprintOpen accessSenior author

  Research in AI4Science has shown promise in many science applications, including polymer design. However, current LLMs are ineffective in this problem space because: (i) most models lack polymer-specific knowledge, and (ii) existing aligned models have limited coverage of knowledge and capabilities relevant to polymer design. Addressing this, we introduce PolyBench, a large-scale training and test benchmark dataset of more than 125K polymer design-related tasks, leveraging a knowledge base of more than 13 million data points obtained from experimental and synthetic data sources to ensure broad coverage of polymers and their properties. For effective alignment using PolyBench, we introduce a knowledge-augmented reasoning distillation method that augments this dataset with structured CoT. Furthermore, tasks in PolyBench are organized from simple to complex analytical reasoning problems, enabling generalization tests and diagnostic probes across the problem space. Experiments show that small language models (SLMs) with 7B to 14B parameters, trained on PolyBench, outperform similar-sized models and remain competitive with closed-source frontier LLMs on PolyBench's test dataset, while demonstrating performance gains on external polymer benchmarks. Dataset and associated code available at https://github.com/StonyBrookNLP/PolyBench.

  PublisherDOI

• Syntax Is Easy, Semantics Is Hard: Evaluating LLMs for LTL Translation

  arXiv (Cornell University) · 2026-04-08

  articleOpen access

  Propositional Linear Temporal Logic (LTL) is a popular formalism for specifying desirable requirements and security and privacy policies for software, networks, and systems. Yet expressing such requirements and policies in LTL remains challenging because of its intricate semantics. Since many security and privacy analysis tools require LTL formulas as input, this difficulty places them out of reach for many developers and analysts. Large Language Models (LLMs) could broaden access to such tools by translating natural language fragments into LTL formulas. This paper evaluates that premise by assessing how effectively several representative LLMs translate assertive English sentences into LTL formulas. Using both human-generated and synthetic ground-truth data, we evaluate effectiveness along syntactic and semantic dimensions. The results reveal three findings: (1) in line with prior findings, LLMs perform better on syntactic aspects of LTL than on semantic ones; (2) they generally benefit from more detailed prompts; and (3) reformulating the task as a Python code-completion problem substantially improves overall performance. We also discuss challenges in conducting a fair evaluation on this task and conclude with recommendations for future work.

  PublisherOA PDFDOI

• Handling Open-Vocabulary Constructs in Formalizing Specifications: Retrieval-Augmented Parsing with Expert Knowledge

  ArXiv.org · 2025-09-10

  preprintOpen accessSenior author

  We study the problem of Open-Vocabulary Constructs(OVCs) -- ones not known beforehand -- in the context of converting natural language (NL) specifications into formal languages (e.g., temporal logic or code). Models fare poorly on OVCs due to a lack of necessary knowledge a priori. In such situations, a domain expert can provide correct constructs at inference time based on their preferences or domain knowledge. Our goal is to effectively reuse this inference-time, expert-provided knowledge for future parses without retraining the model. We present dynamic knowledge-augmented parsing(DKAP), where in addition to the input sentence, the model receives (dynamically growing) expert knowledge as a key-value lexicon that associates NL phrases with correct OVC constructs. We propose ROLex, a retrieval-augmented parsing approach that uses this lexicon. A retriever and a generator are trained to find and use the key-value store to produce the correct parse. A key challenge lies in curating data for this retrieval-augmented parser. We utilize synthetic data generation and the data augmentation techniques on annotated (NL sentence, FL statement) pairs to train the augmented parser. To improve training effectiveness, we propose multiple strategies to teach models to focus on the relevant subset of retrieved knowledge. Finally, we introduce a new evaluation paradigm modeled after the DKAP problem and simulate the scenario across three formalization tasks (NL2LTL, NL2Code, and NL2CMD). Our evaluations show that DKAP is a difficult challenge, and ROLex helps improve the performance of baseline models by using dynamic expert knowledge effectively.

  PublisherOA PDFDOI

• ProST: Progressive Sub-task Training for Pareto-Optimal Multi-agent Systems Using Small Language Models

  ArXiv.org · 2025-09-02

  preprintOpen accessSenior author

  Multi-agent systems with smaller language models (SLMs) present a viable alternative to single agent systems powered by large language models (LLMs) for addressing complex problems. In this work, we study how these alternatives compare in terms of both effectiveness and efficiency. To study this trade-off, we instantiate single and multi-agent systems for the complex problems in the AppWorld environment using different sized language models. We find that difficulties with long-trajectory learning in smaller language models (SLMs) limit their performance. Even when trained for specialized roles, SLMs fail to learn all subtasks effectively. To address this issue, we introduce a simple progressive sub-task training strategy, which introduces new sub-tasks progressively in each training epoch. We find that this novel strategy, analogous to instance level curriculum learning, consistently improves the effectiveness of multi-agents at all configurations. Our Pareto analysis shows that fine-tuned multi-agent systems yield better effectiveness-efficiency trade-offs. Additional ablations and analyses shows the importance of our progressive training strategy and its ability to reduce subtask error rates.

  PublisherOA PDFDOI

• Quantifying Misattribution Unfairness in Authorship Attribution

  2025-01-01

  articleOpen accessSenior author
  PublisherOA PDFDOI

Recent grants

• III: Small: Collaborative Research: Scalable Schema-Based Event Extraction

  NSF · $391k · 2016–2019

• III: Small: Collaborative Research: Explainable Natural Language Inference

  NSF · $253k · 2018–2021

• III: Small: Collaborative Research: Modeling Pre- and Post- Conditions for Understanding Events

  NSF · $408k · 2020–2024

Frequent coauthors

• Nathanael Chambers

  43 shared

• Harsh Trivedi

  25 shared

• Tushar Khot

  24 shared

• Ashish Sabharwal

  24 shared

• Yash Kumar Lal

  23 shared

• Raymond J. Mooney

  The University of Texas at Austin

  19 shared

• Hee Young Kwon

  Korea University

  18 shared

• Tanvi Aggarwal

  Indian Institute of Technology Bombay

  18 shared