Papers by Tong Bai

13 papers
Understanding and Leveraging the Expert Specialization of Context Faithfulness in Mixture-of-Experts LLMs (2025.emnlp-main)

Copied to clipboard

Challenge: Context faithfulness is essential for reliable reasoning in context-dependent scenarios.
Approach: They propose a method that identifies and fine-tunes context-faithful experts . they propose 'context-faither fine- tuning' which selectively fine- tunes them .
Outcome: The proposed method identifies experts with specialization in context utilization and improves context grounding.
Advancing Vision-Language Models with Adapter Ensemble Strategies (2024.findings-emnlp)

Copied to clipboard

Challenge: CLIP revolutes vision-language pretraining by using contrastive learning on paired web data.
Approach: They propose to combine a "adapter ensemble" with traditional machine learning techniques to augment large-scale pretrained vision-language models.
Outcome: The proposed model outperforms baselines and derives improvement when the number of ensemble parameters increases.
HATA: Trainable and Hardware-Efficient Hash-Aware Top-k Attention for Scalable Large Model Inference (2025.findings-acl)

Copied to clipboard

Challenge: Existing top-k attention methods struggle to strike a balance between efficiency and accuracy.
Approach: They propose a top-k attention approach that integrates low-overhead techniques into the Top-k Attention process to achieve 7.2 speedup compared to vanilla full attention.
Outcome: The proposed approach achieves 7.2 speedup compared to current top-k attention methods while maintaining model accuracy.
ZEN: Pre-training Chinese Text Encoder Enhanced by N-gram Representations (2020.findings-emnlp)

Copied to clipboard

Challenge: Experimental results show that pre-trained text encoders can perform many NLP tasks with less resource.
Approach: They propose a BERT-based Chinese text encoder enhanced by n-gram representations . they show reasonable performance when ZEN is trained on a small corpus .
Outcome: The proposed encoder incorporates the comprehensive information of both the character sequence and words or phrases it contains.
LadaBERT: Lightweight Adaptation of BERT through Hybrid Model Compression (2020.coling-main)

Copied to clipboard

Challenge: Existing models that use knowledge distillation are memory-intensive and latency-prohibitive . Existing solutions that use this knowledge distilling framework are expensive .
Approach: They propose a solution that uses weight pruning, matrix factorization and knowledge distillation to learn a smaller model.
Outcome: The proposed model reduces the training overheads by an order of magnitude on public datasets while preserving state-of-the-art accuracy.
Enhancing Self-Attention with Knowledge-Assisted Attention Maps (2022.naacl-main)

Copied to clipboard

Challenge: Existing works of knowledge infusion depend on multi-task learning frameworks, which are inefficient and require large-scale retraining when new knowledge is considered.
Approach: They propose a method which integrates knowledge-generated attention maps into the self-attention mechanism and integrates it into the model.
Outcome: The proposed model outperforms existing methods on academic datasets and industry-scale ad relevance applications.
Syntax-BERT: Improving Pre-trained Transformers with Syntax Trees (2021.eacl-main)

Copied to clipboard

Challenge: Pre-trained language models like BERT achieve superior performances in various NLP tasks without explicit consideration of syntactic information.
Approach: They propose a plug-and-play framework that incorporates syntax trees into pre-trained Transformers.
Outcome: The proposed framework improves on pre-trained models on natural language understanding datasets and shows that it can be used to train pre-structured neural networks.
CogBench: Benchmarking Cognitive Alignment of Large Language Models in Educational Question Answering (2026.findings-acl)

Copied to clipboard

Challenge: Large language models (LLMs) possess strong capabilities in language understanding and generation, as well as remarkable problem-solving abilities.
Approach: They propose a benchmark to assess the cognitive alignment capabilities of large language models in educational QA.
Outcome: The proposed evaluation benchmark assesses the cognitive alignment capabilities of large language models in educational QA.
EverMemOS: A Self-Organizing Memory Operating System for Structured Long-Horizon Reasoning (2026.acl-long)

Copied to clipboard

Challenge: Existing memory systems for LLMs store isolated records and retrieve fragments . Existing systems store isolated data and fragments, limiting their ability to consolidate evolving experience and resolve conflicts.
Approach: They propose an engram-inspired memory operating system that implements an 'engram'-inspired lifecycle for computational memory.
Outcome: Experiments on LoCoMo, LongMemEval, and PersonaMeM-v2 show that EverMemeOS outperforms state-of-the-art methods on memory-augmented reasoning tasks.
Better Language Model with Hypernym Class Prediction (2022.acl-long)

Copied to clipboard

Challenge: Class-based language models (LMs) have been devised to address context sparsity in n-gram LMs for decades.
Approach: They propose to use class-based prediction to improve generalization for rare words by annealing from predicting the class to token prediction during training.
Outcome: The proposed model improves perplexity without sacrificing performance on rare words.
MoPE: Mixture of Prefix Experts for Zero-Shot Dialogue State Tracking (2024.lrec-main)

Copied to clipboard

Challenge: Existing zero-shot dialogue state tracking models suffer from domain transferring and partial prediction problems.
Approach: They propose to establish connections between similar slots in different domains to improve model transfer performance in unseen domains.
Outcome: Empirical results show that the proposed model achieves the goal accuracy of 57.13% on MultiWOZ2.1 and 55.4.
Pre-training CLIP against Data Poisoning with Optimal Transport-based Matching and Alignment (2025.emnlp-main)

Copied to clipboard

Challenge: Recent studies have shown that Contrastive Language-Image Pre-training (CLIP) models are vulnerable to data poisoning and backdoor attacks due to massive training image-caption pairs crawled from the Internet.
Approach: They propose an Optimal Transport-based framework to reconstruct image-caption pairs and propose an optimal transport-based distance measure to re-assign new captions based on the proposed optimal transport distance.
Outcome: The proposed framework reduces the attack success rates of poisoning attacks to 0% in most cases.
Unlocking the Black Box of Latent Reasoning: An Interpretability-Guided Approach to Intervention (2026.acl-long)

Copied to clipboard

Challenge: Existing methods for large language models (LLMs) lack a coherent representation of reasoning steps.
Approach: They propose a set of latent reasoning interventions that enable latent thinking and decode-time interventions that refine the latent process by imposing the identified geometric and semantic priors.
Outcome: The proposed models unlock latent capabilities and improve reasoning accuracy without any parameter updates.

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