Papers by Ananth Balashankar

6 papers
Identifying Predictive Causal Factors from News Streams (D19-1)

Copied to clipboard

Challenge: Existing word embedding techniques are not suited to learn relationships between words in different documents and contexts.
Approach: They propose a new framework to uncover the relationship between news events and real world phenomena by measuring how word occurrence influences future occurrence.
Outcome: The proposed framework outperforms existing methods in stock price prediction errors for 12 months and 4 years.
BEFT: Bias-Efficient Fine-Tuning of Language Models in Low-Data Regimes (2026.acl-long)

Copied to clipboard

Challenge: Fine-tuning bias terms of large language models (LLMs) for downstream tasks has gained a lot of attention over the past few years.
Approach: They extensively evaluate bq, bk, v across a wide range of LLMs . they find that bv generally leads to higher downstream performance in low-data regimes compared to bQ and bK .
Outcome: The proposed method improves performance across a wide range of LLMs spanning encoder-only and decoder-free architectures up to 6.7B parameters.
Reuse Your Rewards: Reward Model Transfer for Zero-Shot Cross-Lingual Alignment (2024.emnlp-main)

Copied to clipboard

Challenge: Multilingual human preference data are difficult to obtain at scale, making it challenging to extend this framework to diverse languages.
Approach: They propose a method where a reward model is trained on preference data in one source language and applied to other target languages.
Outcome: The proposed approach is effective under comprehensive evaluation settings, including human evaluation.
Can We Improve Model Robustness through Secondary Attribute Counterfactuals? (2021.emnlp-main)

Copied to clipboard

Challenge: Recent research has explored how models rely on spurious correlations and how counterfactual data augmentation (CDA) can mitigate such issues.
Approach: They propose a context-aware methodology which takes into account the impact of secondary attributes on the model’s predictions and increases sensitivity for secondary attributes over reweighted counterfactually augmented data.
Outcome: The proposed approach improves sliced accuracy on the original dataset by 7% compared to existing methods and provides guidelines to extend this to other tasks.
Learning Faithful Representations of Causal Graphs (2021.acl-long)

Copied to clipboard

Challenge: Existing text embeddings that predict direct causal links fail to capture other indirect causal links, leading to spurious correlations in downstream tasks.
Approach: They define faithfulness property of contextual embeddings to capture geometric distance-based properties of directed acyclic causal graphs.
Outcome: The embeddings are 31.3% more faithful to human validated graphs with 800K and 200K causal links and achieve better Precision-Recall AUC in a link prediction fine-tuning task.
Improving Classifier Robustness through Active Generative Counterfactual Data Augmentation (2023.findings-emnlp)

Copied to clipboard

Challenge: Existing methods for finding meaningful counterfactuals rely on human annotation or implicit label invariance . a small amount of human-annotated counterf actual data can generate a robust dataset with learned labels.
Approach: They propose a framework that generates counterfactuals by actively sampling from regions of uncertainty and automatically labeling them with a learned auxiliary classifier.
Outcome: The proposed framework generates a large number of diverse counterfactuals and labels them with a learned classifier.

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