👍 99
10/27 08:00
Diffusion vision-language models (DVLMs) process their entire input at every denoising step. Since visual tokens make up 89-97% of this input on the benchmarks we study, most of the decoding computation is spent repeatedly processing the same image. Visual token pruning methods partially relieve thi
中文介绍 针对扩散视觉语言模型(DVLM)每步去噪都处理全部视觉token导致计算浪费的问题,提出证书引导的视觉路由方法,在推理时按需动态选择视觉token子集,减少重复计算,提升效率。
👍 68
10/29 08:00
We present Youtu-Parsing-Omni, an omni-modal parser that converts documents, images, audio, and video into structured records grounded in space and time, providing a perception layer for agents that read, locate, and cite evidence across modalities. Its design rests on one encoder for every modality
中文介绍 提出Youtu-Parsing-Omni全能解析器,用统一编码器和模式,将文档、图像、音频、视频解析为时空结构化记录,为智能体提供跨模态感知层,支持读取、定位和引用证据。
👍 68
10/28 08:00
Recursive self-improvement requires models not only to acquire useful training experience, but also to decide when their current experience has ceased to be useful. Existing methods often focus on what experience to generate next, while paying less attention to when fresh experience should be introd
中文介绍 提出自适应任务刷新(A-RSI)用于零数据多模态自我改进,解决模型何时需要新经验的问题。通过动态判断当前经验效用衰减,及时引入新任务,提升递归自我改进的效率和效果。
👍 38
10/29 08:00
Interactive world models generate visual content conditioned on user actions, yet maintaining spatial consistency over long-horizon exploration in real time remains challenging. Existing approaches rely on either temporal context or explicit 3D memory, both of which scale poorly because memory grows
中文介绍 提出L3G-World,用潜在3D高斯记忆实现实时无限世界生成。在隐空间维护3D场景记忆,避免显式内存增长,支持长时程交互探索并保持空间一致性,实现实时生成。
👍 36
10/27 08:00
Scaling supervised training for working agents requires diverse professional tasks that yield reliable expert demonstrations. However, missing or inconsistent task information can obstruct teacher execution and weaken the resulting supervision. We propose WorkFactory, a scalable framework for constr
中文介绍 提出WorkFactory框架,将专业意图转化为可执行工作流,为工作型智能体生成可靠专家示范。解决任务信息缺失或不一致问题,实现可扩展的监督训练。
👍 28
10/23 08:00
Locating relevant evidence in long inputs is a core capability of Multimodal Large Language Models. Prior studies attribute this capability to retrieval heads, a sparse set of attention heads that transfer information from the context to the output. However, these analyses are confined to a limited
中文介绍 系统研究多模态大模型中的检索头(retrieval heads),揭示其在不同模态和长输入下的行为模式。发现检索头在跨模态信息检索中扮演关键角色,为提升长上下文理解能力提供依据。
👍 22
10/22 08:00
Large language model agents have enabled software development through natural-language interaction, but extending this paradigm to end-to-end game development remains challenging. Building a playable game requires coordinating code, scenes, assets, animations, physics, and configurations, while allo
中文介绍 提出VibeGame框架,用AI原生引擎和自进化对抗智能体团队,通过自然语言提示实现端到端游戏开发。解决代码、场景、资产等协调问题,支持可玩游戏生成。
👍 21
10/21 08:00
A benchmark is not merely a collection of test questions; it naturally induces an empirical measure over difficulty. In this work, we develop this measure-based view of benchmark evaluation. This perspective allows us to identify and formalize two central failure modes: measure mismatch, where the b
中文介绍 将基准测试视为难度上的经验测度,形式化识别度量失配和难度偏差两种核心失败模式。该视角为基准设计提供新理论框架,有助于更准确的模型评估。
👍 20
10/26 08:00
Evaluating human-AI co-design requires testing whether coding agents can translate real users' multimodal requests into working application revisions. We introduce Co-Design Bench, a reference-free benchmark of 273 tasks curated from 37,463 deduplicated queries across 26,314 real user coding session
中文介绍 提出Co-Design Bench基准,含273个真实用户编码任务,测试编码智能体将多模态请求转化为可运行应用修改的能力。无需参考实现,评估人机协作编码质量。
👍 19
10/29 08:00
One-step and few-step text-to-image generation must accurately render the details specified in the prompt despite limited opportunities for refinement. Distribution matching distillation (DMD) uses score differences to bring the generated distribution closer to a reference distribution. Following pr
中文介绍 提出内在分布匹配(Intrinsic Distribution Matching)方法,改进一步/少步文生图模型。通过内在匹配修正分布差异,避免蒸馏中的伪影,提升细节保真度。
👍 18
10/28 08:00
Producing long-form videos from narration requires interdependent decisions that remain coherent across shots. Existing agentic editors often use multi-stage workflows driven by general models, leaving the quality-cost trade-off of compact, task-adapted policies unclear. We address this gap with WeV
中文介绍 提出WeVisCut智能体,从叙述文本直接生成连贯长视频。采用紧凑任务自适应策略,平衡质量与成本,替代多阶段通用模型流程,实现全编辑室一体化。
👍 17
10/15 08:00
Command-line interface (CLI) agents increasingly execute longer-horizon tasks, yet existing benchmarks either assume that user instructions remain fixed throughout execution or evaluate task outcomes without capturing runtime collaboration quality. In this paper, we present REALCLI, a paired benchma
中文介绍 提出REALCLI配对基准,统一评估CLI智能体与人类协作。既捕获运行时协作质量,又评估任务结果,弥补现有基准忽略动态交互的不足。
👍 16
10/16 08:00
Computer-use agents (CUAs) are commonly evaluated against fixed task specifications, but users may omit requirements, add constraints, or require approval before consequential actions. Completing the benchmark instruction does not guarantee serving the person who asked. We introduce PersonaCUABench,
中文介绍 提出PersonaCUABench,评估计算机使用智能体服务不同用户画像的能力。测试智能体处理模糊需求、约束和需批准操作等场景,超越固定指令评估,更贴近真实使用。
👍 12
10/28 08:00
Multi-Token Prediction (MTP) is now a common ingredient in language-model pre-training, as it improves downstream performance and enables faster inference through speculative decoding. We identify a flaw in standard MTP training, in which a residual component of the MTP gradient interferes with both
中文介绍 指出标准多token预测(MTP)训练中存在残差梯度干扰问题,影响性能和推测解码加速。提出修正方法,消除干扰,提升MTP训练效率和模型质量。
👍 12
10/15 08:00
Spoken Dialogue Models (SDMs) typically share a speech-token space for generation and synthesis, leading to long autoregressive sequences and costly inference. Lower-rate speech tokenizers shorten sequences but may introduce compression-quality trade-offs and costly replacement without breaking this
中文介绍 提出CoSTA方法,将压缩语音token适配到口语对话模型,用低速率tokenizer缩短自回归序列,减少推理成本。平衡压缩质量与效率,实现高效口语对话。
👍 12
10/27 08:00
Web agents require diverse, executable, and resettable website environments for reproducible training and evaluation. Human-assisted mirrors demand per-site engineering, while generative construction can diverge from specified real websites. Capture-and-replay preserves observations, with interactio
中文介绍 提出WebAnything框架,从开放网络创建可扩展的智能体环境。结合捕获-重放和生成技术,构建可执行、可重置的网站环境,支持训练和评估。
👍 11
10/28 08:00
Financial agents must form forecasts and investment judgments from decision-time evidence, with predictions evaluated against subsequent outcomes. Existing benchmarks assess financial knowledge, research workflows, or forecasting, but evaluating the full research process requires a strictly frozen i
中文介绍 提出Investor's Last Exam基准,评估AI智能体在严格冻结信息下的股权投资能力。测试决策时证据利用和前瞻性预测,避免未来数据泄漏。
👍 11
09/30 08:00
Many Indian banks serve customers through FAQ pages and scripted chatbots. Assistants based on language models could respond faster and carry out requests such as blocking a card, but only if they give correct information and follow the bank's rules. These rules prescribe a procedure, a set of requi
中文介绍 提出TauIndianBankBench,面向印度银行流程感知的工具使用对话智能体评估。测试智能体遵循银行规则和程序,完成卡片冻结等请求,验证信息正确性。
👍 10
10/28 08:00
Recent advances in multimodal large language models have increased interest in interaction paradigms for video understanding. The evolution from offline to streaming and proactive interaction has gradually shifted models from passive responders toward more proactive agents that can decide when to an
中文介绍 提出AskStream框架,让多模态大模型在流式视频理解中主动决定何时提问。在不确定性下动态询问用户,提升交互效率,从被动响应转向主动智能体。
👍 10
10/29 08:00
Agentic image editors decompose complex requests into steps, but when a step fails, they mostly retry with the same narrow set of tools and a judge that scores each output by eye directly. We argue that such agents are bounded by two spaces: the action space of what the agent can do, and the observa
中文介绍 提出EditAgent,重新思考智能体图像编辑。扩展动作空间和观察空间,让智能体在失败时能切换工具和感知策略,超越单一工具重试,提升编辑成功率。