Kun Liang

Kun Liang, Chenming Tang, Clive Bai et al.

On-policy distillation (OPD) transfers reasoning behavior by training a student on teacher feedback along student-generated trajectories, but standard full-rollout training ties every update to a costly completion and can over-allocate supervision to late positions with low marginal value for the current student. We revisit this assumption through the useful supervision horizon: student-induced rollouts can drift from teacher-preferred continuations, while aligned prefixes may already preserve the long-horizon OPD update direction. We propose ADWIN, an adaptive-window framework for OPD that treats rollout length as an online admissibility decision, training on short teacher-anchored prefixes while using delayed full-rollout probes to audit prefix--full alignment and adapt the next horizon with staleness control. Across math and code reasoning benchmarks in single-task, multi-task, and strong-to-weak settings, ADWIN improves the accuracy--compute trade-off over full-rollout OPD and prefix-based baselines, reducing end-to-end training cost by up to 4.1 times while achieving comparable or better accuracy.

Yuhu Shang, Xuexiong Luo, Lihong Wang et al.

To reduce the repetitive and complex work of instructors, exam paper generation (EPG) technique has become a salient topic in the intelligent education field, which targets at generating high-quality exam paper automatically according to instructor-specified assessment criteria. The current advances utilize the ability of heuristic algorithms to optimize several well-known objective constraints, such as difficulty degree, number of questions, etc., for producing optimal solutions. However, in real scenarios, considering other equally relevant objectives (e.g., distribution of exam scores, skill coverage) is extremely important. Besides, how to develop an automatic multi-objective solution that finds an optimal subset of questions from a huge search space of large-sized question datasets and thus composes a high-quality exam paper is urgent but non-trivial. To this end, we skillfully design a reinforcement learning guided Multi-Objective Exam Paper Generation framework, termed MOEPG, to simultaneously optimize three exam domain-specific objectives including difficulty degree, distribution of exam scores, and skill coverage. Specifically, to accurately measure the skill proficiency of the examinee group, we first employ deep knowledge tracing to model the interaction information between examinees and response logs. We then design the flexible Exam Q-Network, a function approximator, which automatically selects the appropriate question to update the exam paper composition process. Later, MOEPG divides the decision space into multiple subspaces to better guide the updated direction of the exam paper. Through extensive experiments on two real-world datasets, we demonstrate that MOEPG is feasible in addressing the multiple dilemmas of exam paper generation scenario.

Zhaoxi Zhang, Yitong Duan, Yanzhi Zhang et al. · baidu, tsinghua

Locating files and functions requiring modification in large software repositories is challenging due to their scale and structural complexity. Existing LLM-based methods typically treat this as a repository-level retrieval task and rely on multiple auxiliary tools, which often overlook code execution logic and complicate model control. We propose RepoNavigator, an LLM agent equipped with a single execution-aware tool: jumping to the definition of an invoked symbol. This unified design reflects the actual flow of code execution while simplifying tool manipulation. RepoNavigator is trained end-to-end via Reinforcement Learning (RL) directly from a base pretrained model, without relying on closed-source distillation. Experiments demonstrate that RL-trained RepoNavigator achieves state-of-the-art performance, with the 7B model outperforming 14B baselines, the 14B model surpassing 32B competitors, and the 32B model exceeding closed-source models such as GPT-5 on most metrics. These results confirm that integrating a single, structurally grounded tool with RL training provides an efficient and scalable solution for repository-level issue localization.

Kun Liang, Clive Bai, Xin Xu et al.

Recent Large Reasoning Models (LRMs) achieve strong performance by leveraging long-form Chain-of-Thought (CoT) reasoning, but uniformly applying overlong reasoning at inference time incurs substantial and often unnecessary computational cost. To address this, prior work explores various strategies to infer an appropriate reasoning budget from the input. However, such approaches are unreliable in the worst case, as estimating the minimal required reasoning effort is fundamentally difficult, and they implicitly fix the trade-off between reasoning cost and accuracy during training, limiting flexibility under varying deployment scenarios. Motivated by these limitations, we propose ORBIT, a controllable multi-budget reasoning framework with well-separated reasoning modes triggered by input. ORBIT employs multi-stage reinforcement learning to discover Pareto-optimal reasoning behaviors at each effort, followed by on-policy distillation to fuse these behaviors into a single unified model. Experiments show that ORBIT achieves (1) controllable reasoning behavior over multiple modes, (2) competitive reasoning density within each mode, and (3) integration of these frontier policies into a single unified student model while preserving clear mode separation and high per-mode performance.

Jiasheng Chen, Juan Tang, Ming Yan et al.

As is well known, differential algebraic equations (DAEs), which are able to describe dynamic changes and underlying constraints, have been widely applied in engineering fields such as fluid dynamics, multi-body dynamics, mechanical systems and control theory. In practical physical modeling within these domains, the systems often generate high-index DAEs. Classical implicit numerical methods typically result in varying order reduction of numerical accuracy when solving high-index systems.~Recently, the physics-informed neural network (PINN) has gained attention for solving DAE systems. However, it faces challenges like the inability to directly solve high-index systems, lower predictive accuracy, and weaker generalization capabilities. In this paper, we propose a PINN computational framework, combined Radau IIA numerical method with a neural network structure via the attention mechanisms, to directly solve high-index DAEs. Furthermore, we employ a domain decomposition strategy to enhance solution accuracy. We conduct numerical experiments with two classical high-index systems as illustrative examples, investigating how different orders of the Radau IIA method affect the accuracy of neural network solutions. The experimental results demonstrate that the PINN based on a 5th-order Radau IIA method achieves the highest level of system accuracy. Specifically, the absolute errors for all differential variables remains as low as $10^{-6}$, and the absolute errors for algebraic variables is maintained at $10^{-5}$, surpassing the results found in existing literature. Therefore, our method exhibits excellent computational accuracy and strong generalization capabilities, providing a feasible approach for the high-precision solution of larger-scale DAEs with higher indices or challenging high-dimensional partial differential algebraic equation systems.

Yida Cai, Kun Liang, Sanwoo Lee et al.

In recent years, large language models (LLMs) achieve remarkable success across a variety of tasks. However, their potential in the domain of Automated Essay Scoring (AES) remains largely underexplored. Moreover, compared to English data, the methods for Chinese AES is not well developed. In this paper, we propose Rank-Then-Score (RTS), a fine-tuning framework based on large language models to enhance their essay scoring capabilities. Specifically, we fine-tune the ranking model (Ranker) with feature-enriched data, and then feed the output of the ranking model, in the form of a candidate score set, with the essay content into the scoring model (Scorer) to produce the final score. Experimental results on two benchmark datasets, HSK and ASAP, demonstrate that RTS consistently outperforms the direct prompting (Vanilla) method in terms of average QWK across all LLMs and datasets, and achieves the best performance on Chinese essay scoring using the HSK dataset.

Sanwoo Lee, Kun Liang, Yunfang Wu

Recent advances in cross-prompt automated essay scoring (AES) typically train models jointly on all source prompts, often requiring additional access to unlabeled target prompt essays simultaneously. However, using all sources is suboptimal in our pilot study, and re-accessing source datasets during adaptation raises privacy concerns. We propose a source-free adaptation approach that selectively merges individually trained source models' parameters instead of datasets. In particular, we simulate joint training through linear combinations of task vectors -- the parameter updates from fine-tuning. To optimize the combination's coefficients, we propose Prior-encoded Information Maximization (PIM), an unsupervised objective which promotes the model's score discriminability regularized by priors pre-computed from the sources. We employ Bayesian optimization as an efficient optimizer of PIM. Experimental results with LLMs on in-dataset and cross-dataset adaptation show that our method (1) consistently outperforms training jointly on all sources, (2) maintains superior robustness compared to other merging methods, (3) excels under severe distribution shifts where recent leading cross-prompt methods struggle, all while retaining computational efficiency.