Papers by Qiang He
UCS: Estimating Unseen Coverage for Improved In-Context Learning (2026.findings-acl)
Copied to clipboard
| Challenge: | Existing selection methods prioritize heuristic notions of relevance or diversity and provide limited insight into the coverage of a demonstration set. |
| Approach: | They propose a training-free, subset-level coverage prior that is unrevealed by a model-consistent embedding and a Smoothed Good-Turing estimator to estimate the number of unrevelled clusters within a candidate subset. |
| Outcome: | Experiments on multiple intent-classification and reasoning benchmarks show that augmenting strong baselines with UCS improves ICL accuracy by 2-6% under the same selection budget. |
QuASE: Question-Answer Driven Sentence Encoding (2020.acl-main)
Copied to clipboard
| Challenge: | Question-answering (QA) data often encodes essential information in many facets . a growing interest of QA has led to many large-scale QA datasets available to the community . |
| Approach: | They propose a question-answer driven sentence encoding framework to learn representations from QA data. |
| Outcome: | The proposed framework learns representations from QA data, using BERT or other state-of-the-art contextual language models. |
Partial Or Complete, That’s The Question (N19-1)
Copied to clipboard
| Challenge: | Existing annotation schemes aim at acquiring completely annotated structures, but partial annotations can be costly and hinder learning. |
| Approach: | They propose a method to find out that learning from partial structures can sometimes outperform learning from complete ones. |
| Outcome: | The proposed method outperforms existing methods in three different structured learning tasks. |
A2ATS: Retrieval-Based KV Cache Reduction via Windowed Rotary Position Embedding and Query-Aware Vector Quantization (2025.findings-acl)
Copied to clipboard
Junhui He, Junna Xing, Nan Wang, Rui Xu, Shangyu Wu, Peng Zhou, Qiang Liu, Chun Jason Xue, Qingan Li
| Challenge: | Long context large language models (LLMs) pose significant challenges for efficient serving due to the large memory footprint and high access overhead of KV cache. |
| Approach: | They propose a retrieval-based method to reduce the memory footprint of LLMs . they propose Windowed Rotary Position Embedding and query-aware vector quantization . |
| Outcome: | The proposed method can achieve lower performance degradation with lower overhead compared to existing methods . it can reduce the memory footprint and access overhead of long context large language models . |
Foreseeing the Benefits of Incidental Supervision (2021.emnlp-main)
Copied to clipboard
| Challenge: | Real-world applications often require improved models by leveraging a range of cheap incidental supervision signals. |
| Approach: | They propose a unified PAC-Bayesian motivated informativeness measure that characterizes the uncertainty reduction provided by incidental supervision signals. |
| Outcome: | The proposed measure quantifies the value added by incidental supervision signals to sequence tagging tasks. |
Model-based Large Language Model Customization as Service (2025.emnlp-main)
Copied to clipboard
| Challenge: | Existing large language model services require users to upload data for fine-tuning . current methods for customization are noisy and require sensitive domain data . |
| Approach: | *Llamdex is a framework that facilitates LLM customization as a service . client uploads pre-trained domain-specific *models* rather than data . |
| Outcome: | *Llamdex* framework improves domain-specific accuracy by up to 26% over state-of-the-art private data synthesis methods . |
Deeper Insights Without Updates: The Power of In-Context Learning Over Fine-Tuning (2024.findings-emnlp)
Copied to clipboard
| Challenge: | Fine-tuning and in-context learning are two prevalent methods in imbuing large language models with task-specific knowledge. |
| Approach: | They propose to use a circuit shift theory to explain why in-context learning is superior to fine-tuning for tasks with implicit patterns. |
| Outcome: | The proposed method can grasp deep patterns and significantly improve accuracy on implicit patterns, compared with fine-tuning and in-context learning. |
Scaling Behaviors of LLM Reinforcement Learning Post-Training: An Empirical Study in Mathematical Reasoning (2026.acl-long)
Copied to clipboard
Zelin Tan, Hejia Geng, Xiaohang Yu, Mulei Zhang, Guancheng Wan, Yifan Zhou, Qiang He, Xiangyuan Xue, Heng Zhou, Yutao Fan, Zhong-Zhi Li, Zaibin Zhang, Guibin Zhang, Chen Zhang, Zhenfei Yin, Philip Torr, Lei Bai
| Challenge: | elucidating scaling laws for large language models (LLMs) during pre-training remains unexplored. |
| Approach: | They characterize how model scale, data, and compute interact during pre-training . they find that large models consistently demonstrate superior compute and data efficiency . |
| Outcome: | The proposed scaling laws offer practical guidance for scaling reasoning capabilities through reinforcement learning post-training. |
CFinBench: A Comprehensive Chinese Financial Benchmark for Large Language Models (2025.naacl-long)
Copied to clipboard
Ying Nie, Binwei Yan, Tianyu Guo, Hao Liu, Haoyu Wang, Wei He, Binfan Zheng, Weihao Wang, Qiang Li, Weijian Sun, Yunhe Wang, Dacheng Tao
| Challenge: | Large language models (LLMs) have achieved remarkable performance on various NLP tasks, yet their potential in more challenging task like finance, has not been fully explored. |
| Approach: | They propose a benchmark to assess the financial knowledge of large language models (LLMs) in China. |
| Outcome: | The proposed benchmark is the most comprehensive evaluation benchmark to date for LLMs in finance. |
One-Dimensional Object Detection for Streaming Text Segmentation of Meeting Dialogue (2025.findings-acl)
Copied to clipboard
Rui He, Zhongqing Wang, Minjie Qiang, Hongling Wang, Yifan.zhang Yifan.zhang, Hua Xu, Shuai Fan, Guodong Zhou
| Challenge: | Current text segmentation models exhibit numerous limitations, such as imbalances in labels that affect the stability of model training and discrepancies between the model’s training tasks (sentence classification) and the actual text segmenting. |
| Approach: | They implement a sliding window-based segmentation method and employ two different levels of sliding window based balanced label strategies to stabilize the training process of the streaming segmentation model. |
| Outcome: | The proposed method is robust, controllable, and achieves state-of-the-art performance. |