arXiv 论文速递

面向长程图形设计智能体，从真实用户流量中演化程序性记忆，缓解多步依赖任务中的经验复用难题，拓展智能体在专业创作场景的能力。

Professional graphic design is a long-horizon agentic task in which structured, editable artifacts emerge from many interdependent actions, yet outcomes admit no reliable programma…

用代码库自动构造可验证的编码RL环境，解决智能体训练任务稀缺与验证器可靠性问题，对可扩展的编码智能体强化学习训练有直接推动作用。

Training capable coding agents via reinforcement learning (RL) requires diverse tasks with reliable verifiers. Open-source codebases offer a rich source of such tasks, while existi…

指出现有评估只看任务完成率而忽视用户价值取向，基于真实智能体使用数据研究委托行为，对智能体安全与对齐评估有重要实证意义。

Users increasingly delegate work to autonomous AI agents, yet evaluations typically measure task completion rather than the values users prioritize. Using Value Sensitive Design, w…

Advances in large-scale neural recording have made it possible to collect data across many animals and distributed brain regions, raising the question of whether this scale can be …

证明多跳检索失败集中在可预测子群，并给出置信度评分与弃答机制，为RAG系统的可靠性与拒答能力提供理论与实用方法。

Multi-hop retrieval failures are not uniformly distributed across queries: they cluster in structurally predictable subpopulations. We prove two results formalizing this structure.…

A desirable property of a world model is the ability to learn continually across tasks, adapting to new environments without forgetting what the agent has already learnt. In partic…

Graph Convolutional Networks (GCNs) are highly sensitive to label noise, since corrupted supervision can propagate through the graph and degrade learned node representations. This …

为选择性预测这一安全门控机制提供跨ML系统的认证框架，使部署方能量化弃答保障，对高风险场景下模型可靠部署具有基础性价值。

A selective predictor acts as a safety gate: it returns an output only when the prediction appears sufficiently trustworthy. Deployments increasingly require this reliability to be…

Reinforcement learning is increasingly used to align image generators with reward signals, and Flow-GRPO recently extended this paradigm to flow-matching models by treating the den…

We need studies on conversational AI (CAI) at scale to understand human behavior and shape CAI design. However, fragmented reporting of systems and study configurations hinders rep…

Every multiparameter persistence vectorization we know of carries a one-sided Lipschitz upper bound and nothing below it: without a lower gauge there is no sense in which the featu…

Vision-language models (VLMs) rely on storing and transferring appropriate information across their sub-components. Verifying that the VLMs exhibit desired behaviours, while avoidi…