Papers by Yuhao Jiang
AdaMix: Adaptive Mixing for Short and Long Reasoning Adapters (2026.acl-long)
Copied to clipboard
| Challenge: | Existing methods for large reasoning models have improved efficiency but still face limitations such as conflicting objectives and limited adaptability. |
| Approach: | They propose an adaptive reasoning framework that applies a uniform, computation-intensive deep reasoning strategy to all problems. |
| Outcome: | The proposed framework reduces the average response length of DeepSeek-R1-Distill-Qwen-7B by 54.9% while improving accuracy by up to 4.8% on five mathematical datasets. |
S2S-Arena: Evaluating Paralinguistic Instruction Following in Speech-to-Speech Models (2026.acl-long)
Copied to clipboard
Feng Jiang, Zhiyu Lin, Yiyang Liu, Liumeng Xue, Fan Bu, Yuhao Du, Xiangying Chen, Benyou Wang, Haizhou Li
| Challenge: | Existing benchmarks rely heavily on text-based evaluation and largely ignore paralinguistic cues such as prosody, emotion, and speaker traits. |
| Approach: | They propose a speech-native benchmark for evaluating instruction-following S2S models with explicit assessment of both semantic understanding and paralinguistic expression. |
| Outcome: | The proposed system enables more natural, robust, and human-aligned speech agents. |
RecLM: Recommendation Instruction Tuning (2025.acl-long)
Copied to clipboard
| Challenge: | Modern recommender systems aim to understand user-item relationships through past interactions, but their effectiveness is limited when handling sparse data or zero-shot scenarios. |
| Approach: | They propose a model-agnostic recommendation instruction-tuning paradigm that integrates large language models with collaborative filtering. |
| Outcome: | The proposed model-agnostic recommendation instruction-tuning paradigm improves performance across various settings and plug-and-play compatibility with state-of-the-art recommender systems. |
LoRAMoE: Alleviating World Knowledge Forgetting in Large Language Models via MoE-Style Plugin (2024.acl-long)
Copied to clipboard
Shihan Dou, Enyu Zhou, Yan Liu, Songyang Gao, Wei Shen, Limao Xiong, Yuhao Zhou, Xiao Wang, Zhiheng Xi, Xiaoran Fan, Shiliang Pu, Jiang Zhu, Rui Zheng, Tao Gui, Qi Zhang, Xuanjing Huang
| Challenge: | Experimental results show that, as the instruction data increases, LoRAMoE can significantly improve the ability to process downstream tasks, while maintaining the world knowledge stored in the LLM. |
| Approach: | They propose a framework that introduces several low-rank adapters and integrates them by using a router network to freeze the backbone model and force a portion of LoRAs to focus on leveraging world knowledge to solve downstream tasks. |
| Outcome: | The proposed framework freezes the backbone model and forces a portion of LoRAs to focus on leveraging world knowledge to solve downstream tasks, to alleviate world knowledge forgetting. |
Miner: Mining Intrinsic Mastery for Data-Efficient RL in Large Reasoning Models (2026.acl-long)
Copied to clipboard
| Challenge: | Current critic-free RL methods for large reasoning models suffer from severe inefficiency when training on positive homogeneous prompts. |
| Approach: | They propose a method that repurposes the policy’s intrinsic uncertainty as a self-supervised reward signal, with no external supervision, auxiliary models, or additional inference cost. |
| Outcome: | Evaluated across six reasoning benchmarks on Qwen3-4B and Qwend3-8B base models, the proposed method achieves state-of-the-art performance among the other four methods. |
Reinforcement Learning on Pre-Training Data (2026.acl-long)
Copied to clipboard
Siheng Li, Kejiao Li, Zenan Xu, Guanhua Huang, Kun Li, Haoyuan Wu, null Wujiajia, Zihao Zheng, Chenchen Zhang, Kun Shi, Xue Gong, Qi Yi, Ruibin Xiong, Tingqiang Xu, Yuhao Jiang, Jianfeng Yan, Yuyuan Zeng, Guanghui Xu, Jinbao Xue, Zhijiang xu, Zheng Fang, Shuai LI, Qibin Liu, Xiaoxue Li, Zhuoyu Li, Yangyu Tao, Fei Gao, Cheng Jiang, Bochao Wang, Kai Liu, Jianchen Zhu, Wai Lam, Bo Zhou, Di Wang
| Challenge: | Recent progress in large language models is driven by scaling of training compute through pre-training with nexttoken prediction (NTP) or post-training (RL) Pre-training using NTP enables models to acquire extensive knowledge and skills from general data, but it suffers from data inefficiency and catastrophic forgetting in continual learning settings. |
| Approach: | They propose to scale training compute through pre-training with next-token prediction (NTP) or post-training by scaling reinforcement learning (RL) to improve learning from general data. |
| Outcome: | Experiments on multiple benchmarks and models show that the proposed approach improves continual pre-training and provides a strong foundation for post-training on Qwen3-8B-Base. |
Low-probability Tokens Sustain Exploration in Reinforcement Learning with Verifiable Reward (2026.findings-acl)
Copied to clipboard
Guanhua Huang, Tingqiang Xu, Mingze Wang, Qi Yi, Xue Gong, Siheng Li, Ruibin Xiong, Kejiao Li, Yuhao Jiang, Bo Zhou
| Challenge: | Recent studies show that RLVR training is slow and results plateau as policy entropy collapses . low-probability regularization (Lp-Reg) reduces the number of low-quality exploratory tokens induced by RL training . |
| Approach: | They propose a method to reduce RLVR over-penalization by eliminating low-probability exploratory tokens . they propose 'Low-provability Regularization' to reduce the gradual elimination of low-quality exploratory entropy tokens. |
| Outcome: | The proposed method eliminates low-probability exploratory tokens and prevents suppression of potentially valuable low-property candidates. |
Learning Architectures from an Extended Search Space for Language Modeling (2020.acl-main)
Copied to clipboard
Yinqiao Li, Chi Hu, Yuhao Zhang, Nuo Xu, Yufan Jiang, Tong Xiao, Jingbo Zhu, Tongran Liu, Changliang Li
| Challenge: | Neural architecture search (NAS) has advanced in recent years but most NAS systems restrict search to learning architectures of a recurrent or convolutional cell. |
| Approach: | They propose a general approach to learn both intra-cell and inter-cell architectures . they implement their approach in a differentiable architecture search system . |
| Outcome: | The proposed approach outperforms the baseline on PTB and WikiText data and shows good transferability to other systems. |
SimpleDeepSearcher: Deep Information Seeking via Web-Powered Reasoning Trajectory Synthesis (2025.findings-emnlp)
Copied to clipboard
Shuang Sun, Huatong Song, Yuhao Wang, Ruiyang Ren, Jinhao Jiang, Junjie Zhang, Fei Bai, Jia Deng, Xin Zhao, Zheng Liu, Lei Fang, Zhongyuan Wang, Ji-Rong Wen
| Challenge: | Existing approaches to deep search training lack high-quality training trajectories, prohibitive computational costs and lack of high-fidelity training data. |
| Approach: | They propose a framework that synthesizes high-quality training data by simulating real user interactions in live web search environments. |
| Outcome: | The proposed framework synthesizes high-quality training data by simulating user interactions in live web search environments. |
Rhombus: Incentivizing Coordination in Parallel Thinking through Reinforcement Learning (2026.findings-acl)
Copied to clipboard
Ziyuan Nan, Qi Yi, Di Huang, Yutong Wu, Guanhua Huang, Xue Gong, Kejiao Li, Yuhao Jiang, Chenchen Zhang, Zenan Xu, Xing Hu, Bo Zhou
| Challenge: | Parallel thinking is a promising avenue for scaling test-time compute in Large Language Models . however, coordinating the exploration and aggregation stages remains challenging . |
| Approach: | They propose a parallel thinking framework that explicitly incentivizes coordination between components via end-to-end reinforcement learning. |
| Outcome: | The proposed framework improves accuracy by 6.0% over long chain-of-thought baselines while reducing wall-clock latency by 39.4% under matched token budgets. |