Bingran You, Wenjun Ke, Ria Rosenauer, Bo Liu, Wenbo Chen, Min Chen, Shuqi Xu, Yue Ma 等 72 位作者
AI agents are playing an increasingly important role in physics research. However, existing physics benchmarks only evaluate fixed questions, isolated sub-tasks, or paper reproduction with given solutions, and rarely check research quality alongside correctness. We introduce FrontierPhysics, an end-to-end benchmark with 56 tasks made by 42 physicists based on their own research, covering nine domains across theoretical, computational, and experimental physics. Each task requires agents to review the literature, develop a research plan, conduct research and experiments, and produce structured outputs and scientific manuscripts. Combining deterministic verifiers with expert-authored rubrics, FrontierPhysics evaluates whether agents' research is correct and publishable. Across 22 models in 26 configurations, no configuration passes more than 46 of its 168 rollouts. Manual review of agent trajectories shows that frontier agents' main weaknesses are insufficiently rigorous reasoning and limited expert intuition, rather than gaps in physics knowledge or execution errors.
Zhuang Zhao, Jiongchi Yu, Yao Zhang, Dongge Liu, Qiang Hu, Xiaohong Li, Yang Liu
Continuous fuzzing platforms such as OSS-Fuzz provide sustained security testing for open-source projects, but long-term build breakages reduce security coverage and waste infrastructure resources. In OSS-Fuzz, around 20% of integrated projects remain in build-failure states, leaving their fuzzers off duty. These failures expose a maintenance gap between platform maintainers, who lack project-specific context, and project maintainers, who often lack fuzzing and infrastructure expertise. Existing program-repair techniques cannot be directly applied because fuzz build breakages span evolving project and infrastructure artifacts and require repeated infrastructure-aware diagnosis. We present FuzzMedic, an automated agent for fuzz build repair. FuzzMedic performs evidence-guided diagnosis across project and fuzzing-infrastructure artifacts. It explores repairs over a repair-state graph (RSG) that records the evolving workspace and build feedback. Branching and rollback enable FuzzMedic to recover from unproductive repair paths during long-horizon debugging. Candidate patches are validated to restore executable fuzz targets rather than merely suppress build failures. We evaluate FuzzMedic on 128 reproducible OSS-Fuzz build failures extracted from 1,820 real-world failures. FuzzMedic achieves a 48.44% repair rate and outperforms three representative program-repair baselines. In real-world use, it produces 29 patches merged into upstream or OSS-Fuzz main branches, including repairs for three failures that had remained unresolved for more than one year and two cases that had never built successfully.
Tianyang Luo, Tao Feng, Zhigang Hua, Yan Xie, Shuang Yang, Ge Liu, Jiaxuan You
Reinforcement learning (RL) fine-tuning improves large language model (LLM) reasoning using reward signals on sampled rollouts. When reward labels are available for only a subset of rollouts, restricting policy updates to that subset leaves the remaining rollouts unused for policy optimization. Semi-supervised learning propagates labels from labeled to unlabeled samples, but existing methods either diffuse labels with fixed similarity weights or treat each sample independently, ignoring the relations among queries, thinking processes, and answers. We propose MemReward, a graph-based experience memory that propagates rewards to unlabeled rollouts during online policy optimization. MemReward stores queries and the rollouts (thinking processes and final answers) of an initial LLM policy as nodes of a heterogeneous graph connected by similarity and structural edges, and warms up a GNN on labeled rollouts to propagate rewards over this graph. During online RL fine-tuning, unlabeled rollouts are attached to the graph by query similarity, and the policy is updated with ground-truth rewards on labeled rollouts and GNN-predicted rewards on the rest. We evaluate MemReward on 13 benchmarks in mathematics, question answering, and code generation with verifiable rewards, and further with LLM-judge, human-preference, and continuous rewards. With ground-truth rewards on only 20% of rollouts, MemReward achieves 96.6% of the performance of fully supervised training (Oracle) on Qwen2.5-1.5B and 97.3% on Qwen2.5-3B, where it also outperforms label propagation and FixMatch-style self-training. It carries over to Llama backbones up to 8B and closely approaches Oracle on out-of-domain tasks.
Chumeng Liang, Linxuan Wang, Xinyu Peng, Huabin Liu, XiaoXu Guo, Yuxin Chen, Ge Liu, Guang Lin 等 10 位作者
Speculative decoding accelerates language model inference by verifying multiple draft tokens in a single target-model pass. Recent parallel drafters have achieved breakthrough performance in frontier production models, but their effectiveness deteriorates as the entropy of target distributions increases due to insufficient draft diversity. To overcome this bottleneck without sacrificing parallelism, we introduce UBTree, a parallel drafter that couples a Unigram proposer with a Bigram selector to construct drafting Trees. The unigram proposer is trained with the standard cross-entropy objective to generate candidate tokens independently for each position, while a lightweight bigram selector predicts transition scores between adjacent candidate pairs. Unlike the proposer, the selector is trained with a renormalized KL objective on high-temperature data. This tree-native training broadens the supervision beyond the greedy path, encouraging plausible alternative branches that improve the chance of accepting additional tokens during tree verification. Across seven standardized benchmarks with Qwen3-4B and Qwen3-8B, UBTree achieves an average speedup of $5.84$--$6.94\times$ over autoregressive decoding and outperforms DARTree in all 28 comparisons. Production-scale evaluation further demonstrates UBTree's advantage over frontier baselines such as DSpark.
Jiajun Fan, Roger Ren, Jingyuan Li, Rahul Pandey, Prashanth Gurunath Shivakumar, Ankur Gandhe, Yile Gu, Ge Liu 等 9 位作者
Reinforcement learning has made Audio Large Language Models (Audio LLMs) better reasoners, but the reward that drives it still comes from outside the model: a final-answer check, a hand-crafted process score, or an external LLM judge that stays fixed while the policy improves. We show that an Audio LLM can supply this reward itself, and can do so recursively: each improved model becomes the critic that trains its successor, so the training signal improves along with the model. We introduce AudioRSI (Recursively Self-Improving Audio Reasoners), a reinforcement learning framework in which a single model plays two roles: as an actor it generates reasoning paths, and as a critic it compares them pairwise, turning its own judgments into a process-level reward. The critic is periodically replaced by the current policy, which closes the recursive loop. Because the critic hears the same recording as the actor, it can reject reasoning that cites acoustic evidence the audio does not contain, rather than reward reasoning that merely sounds plausible. The recursion is what sustains progress: under a frozen external judge, out-of-distribution accuracy plateaus during training, while a critic refreshed from the improving policy keeps it rising. Across four out-of-distribution benchmarks, all disjoint from the AVQA training distribution, AudioRSI outperforms strong baselines including Audio Flamingo 3 and Gemini 2.5 Pro on MMAU test-mini, ranks first on MMSU, achieves a clear lead on the challenging Multi-Audio split of MMAU-Pro, and leads the 7B open-source tier on MMAR among methods without external reward models. With no reward specifying how to reason, AudioRSI also develops structured reasoning behaviors (logical reasoning, perception-aware analysis, and reduced hallucination), corroborated by an external LLM-as-Judge evaluation and a keyword-based behavior analysis.
Tao Feng, Zijie Lei, Zhigang Hua, Yan Xie, Shuang Yang, Ge Liu, Jiaxuan You
Large language models (LLMs) offer strong semantic representations for ranking, but processing candidate texts within limited contexts and repeatedly invoking an LLM for candidate-level comparisons can be costly when candidate pools contain millions of items. We propose LRanker, a framework for scaling LLM ranking to millions of candidates with test-time scaling. The framework combines two complementary components. A candidate aggregation encoder summarizes offline candidate embeddings using K-means clustering and conditions the LLM on compact global candidate information; ranking then uses inner-product scores with cached candidate embeddings. A graph-based test-time scaling mechanism allocates additional inference computation to partition-conditioned query embeddings and ensembles their ranking scores. This separates the ability to score large candidate pools from the use of additional test-time computation to improve ranking quality. We evaluate LRanker on seven tasks across three scenarios in RBench, including a recommendation task with more than 6.8M candidates. Against the strongest baseline reported for each task, the relative MRR improvements in RBench-Large are 27.5%, 3.9%, 0.9%, and 8.9% on Rec-Movie, Rec-Toy, MS MARCO, and ESCI, respectively. On RBench-Small, the reported task scores improve by 26.7% on Rec-Music and 6.1% on Routing-Balance; on the 6.8M-candidate Rec-Clothing task, the improvement over Tiger is 27.0%. Ablations support the complementary contributions of candidate aggregation and test-time ensembling. These results demonstrate a practical approach to million-scale LLM ranking, with test-time computation providing an additional means of improving ranking effectiveness.