Papers by Rishabh Bhardwaj
Vector-Quantized Input-Contextualized Soft Prompts for Natural Language Understanding (2022.emnlp-main)
Copied to clipboard
| Challenge: | Prompt Tuning has been successful as a parameter-efficient method of conditioning large-scale pre-trained language models to perform downstream tasks. |
| Approach: | They propose to use a vector-quantized input-contextualized prompt as an extension to the soft prompt tuning framework to learn contextualization of soft prompt tokens. |
| Outcome: | The proposed prompt outperforms soft prompt tuning by an average margin of 1.19% on various language understanding tasks like SuperGLUE, QA, Relation classification, NER and NLI. |
Adapter Pruning using Tropical Characterization (2023.findings-emnlp)
Copied to clipboard
| Challenge: | Existing studies on adapter pruning have not examined the optimal number of adapter parameters needed for downstream applications. |
| Approach: | They propose an adapter pruning approach that prunes adapter parameters without changing the orientation of underlying tropical hypersurfaces. |
| Outcome: | The proposed approach prunes adapter layers without changing the orientation of underlying tropical hypersurfaces. |
KNOT: Knowledge Distillation Using Optimal Transport for Solving NLP Tasks (2022.coling-1)
Copied to clipboard
| Challenge: | Knowledge Distillation using Optimal Transport (KNOT) aims to distill the natural language semantic knowledge from multiple teacher networks to a student network. |
| Approach: | They propose to distill natural language semantic knowledge from multiple teacher networks to a student network by learning to minimize the optimal transport cost of its assigned probability distribution over the labels to the weighted sum of probabilities predicted by the (local) teacher models. |
| Outcome: | The proposed method shows improvements in the global model’s SD performance over the baseline across three NLP tasks while performing on par with Entropy-based distillation on standard accuracy and F1 metrics. |
Twitter Homophily: Network Based Prediction of User’s Occupation (P19-1)
Copied to clipboard
| Challenge: | Existing approaches to predicting Twitter users' demographic attributes exploit, select, and combine various features generated from text and network to achieve the best performance. |
| Approach: | They extend existing Twitter occupational class prediction data set and exploit social network homophily to achieve competitive performance. |
| Outcome: | The proposed method achieves better performance on a dataset with a small fraction of the training data. |
Evaluating AI for Finance: Is AI Credible at Assessing Investment Risk Appetite? (2025.emnlp-industry)
Copied to clipboard
Divij Chawla, Ashita Bhutada, Duc Anh Do, Abhinav Raghunathan, Vinod Sp, Cathy Guo, Dar Win Liew, Prannaya Gupta, Rishabh Bhardwaj, Rajat Bhardwaj, Soujanya Poria
| Challenge: | Our analysis was conducted on proprietary systems and open-weight models . FINRISKEVAL analyzed 1,720 profiles spanning a broad spectrum of possible risk categories . |
| Approach: | They evaluated proprietary AI systems and open-weight models to assess investment risk appetite using carefully curated user profiles. |
| Outcome: | The proposed models exhibit significant variance when user attributes that should not influence risk computation are changed. |
More Identifiable yet Equally Performant Transformers for Text Classification (2021.acl-long)
Copied to clipboard
| Challenge: | Current studies prove that attention weights are not unique and therefore unfit for interpretation. |
| Approach: | They propose a transformer encoder layer that decouples the relationship between key and value vector and provides identifiable weights up to the desired length of the input. |
| Outcome: | The proposed model is more identifiable than previously thought but still prone to be non-unique attentions that make them unfit for interpretation. |
Language Models are Homer Simpson! Safety Re-Alignment of Fine-tuned Language Models through Task Arithmetic (2024.acl-long)
Copied to clipboard
| Challenge: | RESTA is a tool to perform LLM realignment towards safety, which gets compromised due to downstream task fine-tuning. |
| Approach: | They propose to add a safety vector to the weights of a compromised model by arithmetic and demonstrate its generalizability on three existing safety evaluation benchmarks and a multilingual benchmark dataset. |
| Outcome: | The proposed model reduces harmfulness of the model from 18.6% to 5.1% and from 9.2% to 1.5% in parameter-efficient and full fine-tuning while maintaining most of the system's performance on the task. |
Ferret: Faster and Effective Automated Red Teaming with Reward-Based Scoring Technique (2025.findings-emnlp)
Copied to clipboard
| Challenge: | Existing red-teaming methods generate adversarial attacks to identify vulnerabilities, but they face slow performance, limited categorical diversity, and high resource demands. |
| Approach: | They propose a method that generates multiple adversarial prompt mutations per iteration and ranks them using scoring functions. |
| Outcome: | The proposed method achieves a 95% attack success rate and reduces time to a 90% ASR by 15.2%. |
Adaptation Approaches for Nearest Neighbor Language Models (2023.findings-acl)
Copied to clipboard
| Challenge: | Semi-parametric Nearest Neighbor Language Models (kNN-LMs) have produced impressive gains over purely parametric LMs, however, there has been little investigation into adapting such models for new domains. |
| Approach: | They propose to adapt kNN-LMs to expand neighborhood retrieval over an additional adaptation datastore and adapt the weights of retrieved neighbors using a learned Rescorer module. |
| Outcome: | The proposed approach outperforms purely parametric adaptation and zero-shot models and achieves perplexity improvements of 17.1% and 16% across domains. |
WalledEval: A Comprehensive Safety Evaluation Toolkit for Large Language Models (2024.emnlp-demo)
Copied to clipboard
Prannaya Gupta, Le Yau, Hao Low, I-Shiang Lee, Hugo Lim, Yu Teoh, Koh Hng, Dar Liew, Rishabh Bhardwaj, Rajat Bhardwaj, Soujanya Poria
| Challenge: | Potential harms include training data leakage, biases in responses and decision-making, and unauthorized use for purposes such as terrorism and the generation of sexually explicit content. |
| Approach: | WalledEval is a comprehensive AI safety testing toolkit designed to evaluate large language models. |
| Outcome: | The framework supports both LLM and judge benchmarking and incorporates custom mutators to test safety against various text-style mutations such as future tense and paraphrasing. |
A Robust Information-Masking Approach for Domain Counterfactual Generation (2023.findings-acl)
Copied to clipboard
| Challenge: | Domain shift is a big challenge in NLP, but many approaches fail to leverage domain-specific nuances relevant to the task at hand. |
| Approach: | They propose a method that uses frequency-based masking to transform a text from the source domain to a target domain. |
| Outcome: | The proposed method outperforms baselines on 10 out of 12 domain-counterfactual classification settings with an average of 1.7% improvement in accuracy metric. |
HYPERTTS: Parameter Efficient Adaptation in Text to Speech Using Hypernetworks (2024.lrec-main)
Copied to clipboard
| Challenge: | Neural text-to-speech (TTS) systems limited to predefined speaker styles or specific sets of speaker IDs. |
| Approach: | They propose a network that can adapt adapter parameters to new speakers . they compare two domain adaptation settings and find it to be very efficient . |
| Outcome: | The proposed Adapters improve speech synthesis performance on two domains and compare them with baselines. |
kNN-CM: A Non-parametric Inference-Phase Adaptation of Parametric Text Classifiers (2023.findings-emnlp)
Copied to clipboard
| Challenge: | Existing studies on text-discriminating properties of semi-parametric models have not been done on non-parameter models. |
| Approach: | They propose an inference-phase approach that incorporates a neighborhood search into a model to enhance the capacity of a pre-trained parametric text classifier. |
| Outcome: | The proposed model improves performance on eight SuperGLUE tasks, three adversarial natural language inference datasets, 11 question-answering (QA) datasets and two sentiment classification datasets. |