Papers by Binbin Zheng
Maximizing Local Entropy Where It Matters: Prefix-Aware Localized LLM Unlearning (2026.acl-long)
Copied to clipboard
| Challenge: | Existing approaches to machine unlearning treat all tokens indiscriminately and enforce uncertainty over the entire vocabulary. |
| Approach: | They propose a framework that targets the prefix in a response and minimizes uncertainty in the critical subspace. |
| Outcome: | The proposed framework achieves superior forgetting efficacy and utility preservation compared to baselines. |
MASPO: Unifying Gradient Utilization, Probability Mass, and Signal Reliability for Robust and Sample-Efficient LLM Reasoning (2026.acl-long)
Copied to clipboard
Xiaoliang Fu, Jiaye Lin, Yangyi Fang, Binbin Zheng, Chaowen Hu, Zekai Shao, Cong Qin, Lu Pan, Ke Zeng, Xunliang Cai
| Challenge: | Existing RLVR algorithms rely on rigid, uniform, and symmetric trust region mechanisms . current algorithms lack robustness, asymmetric signal reliability and inefficient gradient utilization . |
| Approach: | They propose a framework to harmonize three dimensions of RLVR algorithms, a paper argues . a binary cutoff is used to discard valuable reinforcement signals, they argue . |
| Outcome: | The proposed framework outperforms baselines in evaluating a robust RLVR solution. |
From log 𝜋 to 𝜋: Taming Divergence in Soft Clipping via Bilateral Decoupled Decay of Probability Gradient Weight (2026.acl-long)
Copied to clipboard
Xiaoliang Fu, Jiaye Lin, Yangyi Fang, Chaowen Hu, Cong Qin, Zekai Shao, Binbin Zheng, Lu Pan, Ke Zeng
| Challenge: | Standard algorithms for Large Language Models (LLMs) enforce stability via "hard clipping" but relying on log-probability gradient yields divergent weights as probabilities vanish, destabilizing LLM training. |
| Approach: | They propose a decoupled gradient policy optimization that uses a decay mechanism to decouple the probability of a boundary token. |
| Outcome: | The proposed algorithm outperforms baselines on various mathematical benchmarks. |
NeuReduce: Reducing Mixed Boolean-Arithmetic Expressions by Recurrent Neural Network (2020.findings-emnlp)
Copied to clipboard
| Challenge: | Mixed Boolean-Arithmetic (MBA) expressions are difficult to simplify because of interleaving bitwise and arithmical operations. |
| Approach: | They propose a method to learn and reduce MBA expressions using a string to string method . they propose to use a dataset to train the method to reduce MBA rules . |
| Outcome: | The proposed method outperforms all other tools in terms of accuracy, solving time, and performance overhead. |
GraphMR: Graph Neural Network for Mathematical Reasoning (2021.emnlp-main)
Copied to clipboard
| Challenge: | Existing studies have shown the effectiveness of sequence-to-sequence (Seq2Seque) on mathematics solving. |
| Approach: | They propose a graph-to-sequence neural network which can learn hierarchical information of graphs inputs to solve mathematical problems and speculate answers. |
| Outcome: | The proposed neural network outperforms other neural networks in hidden information learning and mathematics resolving. |