👍 157
06/10 08:00
Many moments in the real world do not wait for a user to ask. A fire starts on a security monitor, an expression flickers across a video call, or a product a viewer wants flashes by in a livestream. Yet today's large models remain mostly turn-based by design: they answer only when addressed, and eve
中文介绍 提出JoyAI-VL-Interaction,一种实时视觉-语言交互智能体,解决现有大模型需用户主动提问的轮次式设计局限。核心技术思路是让模型能主动感知并响应实时场景(如安全监控起火、直播中商品闪过),无需用户触发。
👍 100
06/09 08:00
Data tells stories that shape society; the data journalist's job is to turn raw information into stories non-experts can trust. A high-quality news feature takes a newsroom team weeks: hunting for context, running statistics, choosing an angle, and designing visuals. Recent agents handle individual
中文介绍 提出Data Journalist Agent,将数据转化为可验证的多模态新闻故事,解决传统新闻制作周期长、依赖专家的问题。该智能体自动完成上下文搜寻、统计分析、角度选择和可视化设计,生成可信的新闻特征稿。
👍 86
06/15 08:00
Generalist robot policies must follow user instructions while reasoning about how objects, cameras, and robot actions interact in the 3D physical world. Recent vision-language-action models (VLAs) and video world-action models (WAMs) inherit strong semantic or temporal priors from large-scale founda
中文介绍 提出Geometric Action Model (GAM) 用于机器人策略学习,解决现有VLA、WAM等模型在3D物理世界中建模物体-相机-动作交互的不足。GAM显式建模几何空间关系,提升指令跟随和物理推理能力。
👍 78
06/15 08:00
DreamX-World 1.0 is a general-purpose interactive text/image-to-video world model for controllable long-horizon generation. It supports camera navigation, revisits to previously observed regions, and promptable events across photorealistic, game-style, and stylized domains. Our data engine combines
中文介绍 DreamX-World 1.0是一个通用交互式世界模型,支持可控长视频生成。可进行相机导航、重访已观察区域、事件提示等交互,覆盖写实、游戏和风格化领域。通过数据引擎结合多种合成数据实现多域泛化。
👍 76
06/12 08:00
Large Language Model (LLM) coding agents have achieved strong results on software engineering tasks, yet repository exploration remains a major bottleneck: locating relevant code consumes substantial token budget and pollutes the agent's context with irrelevant snippets. In most agents, the same mod
中文介绍 提出FastContext,为代码智能体训练高效仓库探索器,解决仓库级任务中定位相关代码消耗大量token且污染上下文的问题。通过专用模型替代通用模型进行探索,显著降低上下文开销。
👍 62
06/15 08:00
This technical report introduces VibeThinker-3B, a compact dense model with 3B parameters developed to investigate how far verifiable reasoning can be pushed within a strictly small-model regime. Building upon the Spectrum-to-Signal post-training paradigm, we systematically enhance the model through
中文介绍 VibeThinker-3B是一个3B参数小模型,探索严格小模型下可验证推理的极限。基于Spectrum-to-Signal训练范式,系统增强推理能力,在多种推理基准上取得与更大模型竞争的结果。
👍 23
06/15 08:00
Masked Diffusion Language Models (MDLMs) have emerged as a distinct paradigm for sequence generation. As MDLMs become diverse in capabilities and knowledge coverage, an important question is how to combine their knowledge. Toward this, we first investigate the unique decoding dynamics of MDLMs. We f
中文介绍 研究掩码扩散语言模型(MDLM)的集成解码策略。发现MDLM独特的解码动态,提出追踪可靠轨迹的方法,有效组合多个MDLM的知识,提升生成质量。
👍 21
06/15 08:00
Vision language models are serving as general-purpose interfaces for complex multimodal tasks. However, deployment still faces three gaps: VLMs typically incur high latency and cost when processing dense video frames and long prompts, the agent scaffold remains static after deployment, and standard
中文介绍 VisualClaw是面向物理世界的实时个性化智能体,解决VLM部署中的高延迟、静态框架和交互不足三大问题。通过轻量化视频处理、自适应脚手架和实时交互设计,实现低成本、可定制的物理世界操作。
👍 14
06/15 08:00
Multi-task learning (MTL) is essential in recommender systems to enable complementary learning among diverse user feedback. While modern industrial practices have shifted from DNNs to Transformer-centric architectures to strengthen sequence modeling and scaling capacity, they still decouple feature
中文介绍 OneRank提出统一的Transformer原生排序架构,用于多任务推荐。将特征交互和序列建模合并到单一Transformer中,替代现有DNN+Transformer分离设计,提升多任务互补学习效果。
👍 14
06/15 08:00
Visual world models (VWMs) synthesize interactive, action-conditioned rollouts from a single context image. However, it remains an open question how robust these models are to adversarial perturbations. Standard adversarial attacks fail to assess this vulnerability because attackers lack ground-trut
中文介绍 BadWorld首次研究视觉世界模型(VWM)的对抗鲁棒性。针对VWM缺乏真实反馈的标准对抗攻击无法生效,提出针对交互式视频生成的新型对抗攻击方法,揭示模型脆弱性。
👍 14
06/15 08:00
We introduce Qwen-RobotWorld, a language-conditioned video world model for embodied intelligence. With natural language as a unified action interface, it predicts physically grounded future visual trajectories from current observations across robotic manipulation, autonomous driving, indoor navigati
中文介绍 Qwen-RobotWorld是语言条件视频世界模型,统一具身世界建模。以自然语言作为统一动作接口,从当前观测预测物理上合理的未来视觉轨迹,适用于机器人操控、自动驾驶、室内导航等任务。
👍 13
06/15 08:00
As LLM agents are deployed in long-horizon sessions, context accumulation drives up inference costs. Existing approaches utilize text pruning or dynamic memory eviction to minimize token footprints; however, their unconstrained sequence mutations alter layouts, introducing prefix mismatches and cach
中文介绍 TokenPilot为LLM智能体提供缓存高效的上下文管理,解决长会话中上下文积累导致推理成本高的问题。通过保留上下文布局、避免前缀不匹配,在减少token占用的同时保持缓存友好性。
👍 13
06/10 08:00
Modern Lean theorem provers achieve strong performance only with substantial training and inference compute, driven in part by scarce verified proof data and the long reasoning traces of formal proof search, making both supervised fine-tuning (SFT) and sampling expensive. We introduce Pythagoras-Pro
中文介绍 Pythagoras-Prover通过增强型Lean形式化方法提升形式化证明效率。解决训练数据稀缺和推理轨迹过长的问题,在更少计算资源下达到与大型方法相当的性能。
👍 11
06/13 08:00
Advanced agents are increasingly demonstrating the potential to operate as autonomous engineers, creating a growing demand for evaluation benchmarks that capture the complexity of real-world development. Such environments typically involve both complex code and large-scale data (i.e., file system).
中文介绍 CODA-BENCH是面向代码智能体的数据密集型任务基准,涵盖复杂代码和大规模文件操作的真实开发场景。用于评估智能体在数据驱动的工程任务中的实际表现。
👍 10
04/08 08:00
Web agents act through long interaction sequences, yet existing benchmarks evaluate only terminal success, discarding all process information and offering little guidance on improvement. In this work, we conduct a process-level analysis of web agents. We introduce WebStep, a benchmark of 1,800 task
中文介绍 WebStep提供过程级Web智能体评估,通过语义状态跟踪分析交互过程中的每一步。包含1800个任务,不仅评估最终成功,还定位失败环节,为改进提供细粒度指导。
👍 9
06/13 08:00
Efficient and scalable agentic intelligence requires models that can deliver both low-latency responses and strong reasoning capabilities while remaining practical to train, serve, and deploy. In this report, we present Ling-2.6 and Ring-2.6, a family of models designed to address this challenge at
中文介绍 Ling-2.6和Ring-2.6是万亿参数规模的智能体模型家族,实现低延迟强推理与高效训练部署。采用MoE架构,在多项基准上达到领先性能,同时保持实用部署能力。
👍 9
06/15 08:00
As LLMs advance, post-training reinforcement learning (RL) increasingly relies on multi-dimensional rewards to cultivate comprehensive capabilities. This shift demands new algorithms capable of optimizing diverse and potentially competing objectives simultaneously. To address this, existing methods
中文介绍 GD²PO提出群体动态奖励解耦策略优化,解决多奖励冲突问题。针对LLM后训练中多种奖励目标(如有用性、安全性)的竞争,通过动态分组和解耦优化各项能力。
👍 9
06/15 08:00
Multi-turn LLM serving accumulates dialogue history whose Key-Value (KV) cache grows with every turn and every user, quickly exceeding the model weights themselves and making memory -- not compute -- the binding constraint on throughput. Non-uniform KV compression, which allocates heterogeneous budg
中文介绍 Tangram实现非均匀KV缓存压缩,解决多轮对话中KV缓存膨胀导致的显存瓶颈。通过异构预算分配,在保持生成质量的同时显著减少内存占用,提升吞吐量。
👍 8
06/12 08:00
Phone agents are increasingly expected to complete real mobile workflows rather than merely predict the next screen action. However, much of the current mobile-agent literature still evaluates agents primarily as GUI controllers that observe a screen, emit taps and swipes, and are scored by target a
中文介绍 PhoneHarness是手机使用智能体评估框架,混合GUI、CLI和工具动作。不同于仅预测屏幕点击,要求智能体完成真实移动工作流,支持多模态交互评估。
👍 8
06/12 08:00
We introduce Nemotron 3 Ultra, a 550 billion total and 55 billion active parameter Mixture-of-Experts Hybrid Mamba-Attention language model. We pre-trained Nemotron 3 Ultra on 20 trillion text tokens, then extended the context length to 1M tokens, and post-trained using Supervised Fine Tuning (SFT),
中文介绍 Nemotron 3 Ultra是550B参数(55B活跃)的MoE混合Mamba-Transformer模型。在20T文本上预训练,扩展至1M上下文,经SFT、RLHF等后训练,在推理和智能体任务上表现优异。