Papers by Balaji Krishnamurthy

9 papers
LM-CORE: Language Models with Contextually Relevant External Knowledge (2022.findings-naacl)

Copied to clipboard

Challenge: Large pre-trained language models can capture factual knowledge in their parameters but storing large amounts of knowledge in the model parameters is sub-optimal given the ever-growing amounts of information and resource requirements.
Approach: They propose a framework that provides explicit access to contextually relevant structured knowledge to the model and train it to use that knowledge.
Outcome: The proposed framework outperforms state-of-the-art knowledge-enhanced language models on knowledge probing tasks and can handle knowledge updates.
A Video Is Worth 4096 Tokens: Verbalize Videos To Understand Them In Zero Shot (2023.emnlp-main)

Copied to clipboard

Challenge: Existing annotated training datasets hinder development of supervised learning models for multimedia content . lack of annotating benchmarks hinders development of models with satisfactory performance . a recent study shows that large language models have zero-shot performance in multimedia understanding .
Approach: They propose to verbalize long videos to generate their descriptions in natural language . they then perform video-understanding tasks on the generated story as opposed to the original video .
Outcome: The proposed method achieves better results than baselines for video understanding.
TAN-NTM: Topic Attention Networks for Neural Topic Modeling (2021.acl-long)

Copied to clipboard

Challenge: Topic models have been widely used to learn text representations and gain insight into document corpora.
Approach: They propose a framework which processes document as a sequence of tokens through a LSTM whose contextual outputs are attended in a topic-aware manner.
Outcome: The proposed model improves on two downstream tasks: document classification and topic guided keyphrase generation.
INGENIOUS: Using Informative Data Subsets for Efficient Pre-Training of Language Models (2023.findings-emnlp)

Copied to clipboard

Challenge: Pre-trained language models have a remarkable improvement in generalization capability . however, this leads to prohibitively long training times and a detrimental environmental impact .
Approach: They propose to use submodular optimization to select highly informative subsets of training data to train multiple PTLMs using only fractions of data.
Outcome: The proposed framework achieves 99% of the performance of fully-trained models using only fraction of training data.
Learning Together to Perform Better: Teaching Small-Scale LLMs to Collaborate via Preferential Rationale Tuning (2025.acl-long)

Copied to clipboard

Challenge: Prior studies have demonstrated that LLMs generate step-by-step rationales, but limited data is available to improve their performance in commercial settings due to copyright and legal issues.
Approach: They propose a trainable framework that tunes a (small) LLM to generate outputs from a pool of diverse rationales that selectively improves the downstream task.
Outcome: The proposed framework outperforms several trainable and prompting baselines on maths problem solving, natural language inference, and commonsense reasoning.
HyHTM: Hyperbolic Geometry-based Hierarchical Topic Model (2023.findings-acl)

Copied to clipboard

Challenge: Hierarchical Topic Models (HTMs) often produce hierarchies where lower-level topics are unrelated and not specific enough to their higher-level subjects.
Approach: They propose a Hyperbolic geometry-based Hierarchical Topic Model that incorporates hierarchical information from hyperbolic geometrics to explicitly model hierarchies in topic models.
Outcome: The proposed model is significantly faster and leaves a much smaller memory footprint than the best-performing baseline.
CoSe-Co: Text Conditioned Generative CommonSense Contextualizer (2022.naacl-main)

Copied to clipboard

Challenge: Pre-trained language models (PTLMs) have been shown to perform well on natural language tasks.
Approach: They propose a commonsense contextualizer conditioned on sentences as input to make it generically usable in tasks involving natural language text.
Outcome: The proposed model improves on existing methods on CSQA, ARC, QASC and OBQA datasets.
Form2Seq : A Framework for Higher-Order Form Structure Extraction (2020.emnlp-main)

Copied to clipboard

Challenge: Document structure extraction is a widely researched area for decades due to image resolution and poor semantics.
Approach: They propose a sequence-to-sequence framework for document structure extraction using text . they use a text-based framework to classify low-level constituent elements into ten types .
Outcome: The proposed framework outperforms existing methods for document structure extraction on ICDAR 2013 dataset.
Synthesizing Human Gaze Feedback for Improved NLP Performance (2023.eacl-main)

Copied to clipboard

Challenge: Prior work on eye tracking and NLP reveals that human scanpaths can aid in understanding and performance of NLP models.
Approach: They propose a model for generating human scanpaths over text that approximates meaningful cognitive signals in human gaze patterns.
Outcome: The proposed model can approximate meaningful cognitive signals in human gaze patterns.

What is GenGO?

GenGO is an NLP powered publication search system. It currenctly indexes 30k+ papers from ACL Anthology, and implements multi-aspect summarization, semantic search, and more!

Information

About
Limitations