跳到正文
研踪PaperTrail

论文库

检索计算机领域论文,按研究方向、作者或关键词筛选。

部分采集异常
正在读取账号与研究记录…

全站更新概况

+1680新增论文
+1333其中 AI 相关
+187已有论文修订
+59新增 AI 导读

arXiv 最近完成扫描:2026-11-18 04:28 · 全站最近入库:2026-11-17 23:35各信源进度 →

统计口径与数据新鲜度

统计 2026-11-17 09:07 — 2026-11-18 09:07(北京时间) · 点击数字查看对应简报;AI 相关包含在新增论文内。

最近完成检查:2026-11-18 04:28最近发现新增版本:2026-11-17 23:35库内论文最新更新时间:2026-11-17 02:59

每分钟刷新已保存的数据。空检查不代表采集异常;新收录可能包含历史论文。AI 导读按新增版本计数。

25,786 已收录论文21,491 AI 相关7,725 有修订112 HF 社区发现

浏览论文库

时间范围跟随排序:版本更新、本站收录或首次发表。最新收录可能包含历史论文。部分信源异常,查看状态 →

找到 166 篇论文 · 按版本更新时间排序上轮扫描起点 2026-11-18 04:27(北京时间)
版本更新 · v2arXiv:2606.02355

SIRI: Self-Internalizing Reinforcement Learning with Intrinsic Skills for LLM Agent Training

Zhongyu He, Yuanfan Li, Fei Huang, Tianyu Chen, Siyuan Chen, Xingyang Li, Meng Hsuan Yu, Xiangrong Liu 等 12 位作者

原文摘要

Long-horizon LLM agents can benefit from reusable skills, yet existing skill-based methods often rely on external skill generators during training or persistent skill retrieval at inference, increasing engineering complexity, context length, and deployment latency. We propose Self-Internalizing Reinforcement learning with Intrinsic skills (SIRI), a three-phase framework that enables agents to discover, validate, and internalize skills without external skill generators or inference-time skill banks. SIRI first warms up the policy with GiGPO to acquire basic interaction ability and collect successful skill-free trajectories. It then performs self-skill mining, where the current policy summarizes compact skills from its own successful plain rollouts and validates them through paired skill-augmented and skill-free rollouts. Finally, SIRI distills only beneficial skill-guided action tokens into the plain policy using trajectory-level utility and action-level advantage. At inference, the agent runs with the original prompt only. On ALFWorld and WebShop with Qwen2.5-7B-Instruct, SIRI improves GiGPO from 0.908 to 0.930 on ALFWorld and from 0.728 to 0.813 on WebShop, outperforming prompt-based, RL-based, and memory-augmented baselines. Further analysis shows that our self-mining strategy can achieve performance comparable to distillation with closed-source large model.

cs.AI · cs.LG

新论文 · v1arXiv:2611.21020

SWE-OPD: On-Policy Distillation for Agentic Coding

Yuhang He, Yanyu Chen, Chuanzhe Guo, Jie Zhu, Zichen Tian, Hongyu Ge, Zixuan Ren, Huang He 等 14 位作者

原文摘要

In this work, we study on-policy distillation (OPD) for transferring agentic coding capabilities from a large teacher model to a smaller student model. Prior work has focused primarily on reasoning and short-horizon agentic tasks. We present SWE-OPD, a minimal OPD study for software engineering, where the student interacts with executable repositories and the teacher supervises the prefixes visited by the student. Across three teacher--student pairs, two training sources, and three SWE benchmarks, we find that on-policy supervision transfers coding capability: supervision from Qwen3.6-27B increases the SWE-bench Pro resolve rate of Qwen3.5-9B from 35.9% to 39.4%. Refreshing trajectories is not strictly required either, as an offline variant with cached teacher scores still improves over the student baselines while cutting end-to-end GPU-hours by 59.4%--72.6% and halving peak GPU allocation, at benchmark-dependent accuracy. Furthermore, we extend the offline variant to corss-tokenizer scenarios: distilling GLM-4.5-Air from DeepSeek-V4-Flash-0731 raises its SWE-bench Pro and Multilingual resolve rate from 35.8% to 39.7% and from 41.8% to 46.2%, around 4% and 4.6% improvement, respectively. We also find that a decreasing loss, however, does not necessarily indicate improved agent performance: under top-$K$ reverse-KL loss, one student exhibits severe repetition despite a decreasing loss. We identify a support-mass gradient term as a potential mechanism underlying this failure.

cs.AI · cs.LG

新论文 · v1arXiv:2611.20971

FAVR: Flow-Aware Visual Residual Alignment for Vision-Language-Action Policies

Qixuan Wu, Xiaole Yang, Yuguang Yang, Canyu Chen, Linlin Yang, Haixia Pan, Guodong Guo, Baochang Zhang

原文摘要

Pretrained video representations provide rich temporal information for robotic manipulation, yet effectively integrating them into existing vision-language-action (VLA) policies remains challenging. We present Flow-Aware Visual Residual Alignment (FAVR), a lightweight framework for flow-aware visual conditioning of frozen VLA policies. FAVR maintains a compact temporal visual state through linear Test-Time Training (TTT) and retrieves action-conditioned visual features throughout flow-based action generation. By exploiting the affine structure of the frozen action head, we explicitly characterize how injected visual features influence predicted action velocities and formulate flow-aware channel selection as a box-constrained optimization problem. We develop a FISTA-based solver for joint gate optimization and derive an efficient closed-form approximation through diagonalization and rescaling. With only 1.74-2.19M trainable parameters and 5,000 adaptation steps, FAVR improves pretrained VLA policies across benchmarks. On RoboTwin 2.0, FAVR improves LingBot-VLA from 87.12% to 90.04% under clean conditions and from 86.88% to 89.38% under randomized conditions. A controlled ablation on five RoboTwin tasks further shows that flow-aware gating increases average success from 72.0% with a gate shared across flow steps to 82.0%, demonstrating the benefit of adapting visual conditioning throughout action generation.

cs.CV

新论文 · v1arXiv:2611.20758

UniGraph: A Unified Graph-based Index for Dynamic Hybrid Dense-Sparse Vector Retrieval

Keyu Chen, Xiaoyao Zhong, Jiabao Jin, Peng Cheng, Wangze Ni, Jingkuan Song, Heng Tao Shen

原文摘要

Hybrid dense-sparse vector retrieval combines semantic signals from dense representations with lexical signals from sparse representations, typically ranking items by a query-dependent weighted sum of dense and sparse vector distances. Existing systems commonly search separate dense and sparse vector indexes and rerank the union of their candidates, incurring two retrieval passes and potentially discarding items that rank highly only after the semantic and lexical signals are combined. Constructing a graph-based index directly under a hybrid distance can mitigate this premature candidate loss. However, because varying the hybrid weight alters the underlying neighborhood relationships, a single static graph built at a fixed weight cannot accommodate dynamic query-specific weights. Therefore, we propose a Unified Hybrid Graph (UniGraph), which merges neighbor sets computed at multiple hybrid weights into a single bounded-degree graph index, enabling hybrid vector retrieval with different query-specific weights without index reconstruction. We further develop an adaptive weight-guided search framework that chooses between sparse- and dense-guided entry points according to the query weight, together with a dense-first hybrid distance pruning strategy that avoids unnecessary exact sparse distance computations. Extensive experiments show that UniGraph provides a favorable Recall-QPS trade-off across different query weights. In the mixed-weight setting, UniGraph delivers 27.6%~52.3% higher QPS than the fastest baseline on each dataset at Recall@100 greater than or equal to 0.95. UniGraph has been deployed at scale in Ant Group's production environment, where hybrid retrieval with query-dependent weights is in strong demand.

cs.IR · cs.DB

版本更新 · v3arXiv:2611.12063

From Token Bills to Benchmark Weeks: Making Auto-Research for Data Systems Verifiable and Affordable

Kaiwen Zheng, Srijan Srivastava, Zheng Yan, Xinpeng Zhao, Shaohua Li, Chieh-Jan Mike Liang, Tianyu Chen, Haoxian Chen 等 9 位作者

原文摘要

AI auto-research is poised to forge bespoke data systems, from key-value stores and indexes to query engines, each engineered for a specific workload and hardware platform. Yet this promise rests on fragile ground. An agent optimizes what it can measure, not necessarily what users need. Natural-language requirements leave ample room for specification gaming, producing systems that blaze through benchmarks while silently losing updates or violating consistency. Worse, each research iteration consumes tokens, while validating a candidate under real concurrency, memory pressure, and failure conditions can stretch the entire research process into weeks. In this paper, we envision a more cost-efficient and verifiable approach to auto-research. Precise and composable specifications anchor the search, lightweight performance critics prune costly dead ends, and falsification and deductive verification make correctness an integral part of the process rather than an afterthought. By treating correctness, performance, research time, and token cost as equally important objectives, we aim to transform AI-generated data systems from plausible artifacts into trustworthy, production-grade systems whose behavior can be understood, audited, and ultimately relied upon.

cs.DB

新论文 · v1arXiv:2611.20050

Clinician Assessment of Laryngeal Disease with Auditory AI before Laryngoscopy: A Paired Reader Study

Siyuan Hou, Shuting Yang, Zengrui Jin, Liyu Cheng, Siyin Wang, Yunhan Shi, Xueyan Li, Rong Hu 等 16 位作者

原文摘要

Background: Before laryngoscopy, clinicians assess the likely cause of hoarseness and whether to suspect a precancerous or malignant lesion. Whether voice-based artificial intelligence (AI) supports pre-laryngoscopic categorization is unknown. Methods: We developed an auditory large language model that generates a patient-level assessment card (provisional category, voice-quality ratings, and rationale) and a prespecified malignancy alert from voice recordings, patient information, and acoustic measures obtained before laryngoscopy at one center. We evaluated both in a held-out test set and a later cohort. In a sequential reader study, 32 otolaryngology clinicians categorized held-out cases without AI and, 1 to 2 weeks later, reviewed their initial categories with the card (without the alert). Results: Across 800 paired assessments, reader accuracy was 49.8% before and 56.8% after review (difference, 7.0 percentage points; 95% CI, 1.7 to 13.5; P < 0.001); 99 categories changed from incorrect to correct and 43 from correct to incorrect (ratio, 2.30; 95% CI, 1.22 to 5.31). Accuracy rose 14.7 points when the model was correct and fell 8.7 when incorrect. Provisional-category accuracy was 72% in held-out testing and 67% in temporal validation. The malignancy alert flagged 85.7% and 86.8% of precancerous or malignant lesions. Sensitivity for precancerous or malignant lesions increased from 53.9% to 75.0% after AI-assisted review (difference, 21.1 percentage points; 95% CI, 4.3 to 42.0). Conclusions: In this single-center study, clinicians' pre-laryngoscopic categorization accuracy was higher after reviewing the AI assessment card, although the sequential design cannot fully separate AI assistance from reassessment; gains depended on model correctness without clearly selective adoption. Randomized and external evaluations, including of the malignancy alert, are needed.

eess.AS · cs.AI

新论文 · v1arXiv:2611.20045

Security of Deep Research Agents: Attack Surface, Defenses, Benchmarks, and Agenda Seeding against Research Planning

Yang Luo, Guangpu Chen, Yichen Liu, Zhaoyu Chen, Yifeng Gao, Zhiyang Ding, Qingni Shen, Zhonghai Wu

原文摘要

Deep research and agentic search agents autonomously retrieve content from the open web and generate reports with citations, and their outputs are often taken directly as research conclusions. This paper systematizes security research in this area along four stages, namely retrieval, planning, evidence integration, and report generation. We present a unified threat model, a taxonomy of attacks and defenses, and a review of evaluation benchmarks, and we align the area dimension by dimension with retrieval-augmented generation (RAG) poisoning, prompt injection, and generative engine optimization (GEO). Coding 34 primary studies shows that only 26.5% of them involve the planning stage, and none performs an end-to-end poisoning evaluation on commercial products. To address this gap, we propose the agenda seeding attack, in which the attacker writes no instructions and only introduces a fictitious evaluation metric. Across 3,520 sandboxed research runs under the same budget, GEO and prompt injection are nearly ineffective, whereas agenda seeding makes the attacker's target the top pick in 16% to 28% of reports. Injection detection and query anchoring fail against it, verification prompting hurts utility, and our proposed provenance tagging preserves utility but can be bypassed by multi-site bait. Finally, we propose eight research directions.

cs.CR · cs.AI

版本更新 · v2arXiv:2611.17160

XBot-Omni: An Omni-Modal Model for Embodied Understanding and Interaction

Kerui Gu, Xing Fu, Shuoyang Sun, Cheng Peng, Jingru Duan, Huayu Gu, Zhenyu Chen, Jiaxin Huang 等 13 位作者

原文摘要

Embodied interaction requires an agent to integrate multimodal perception with verbal and nonverbal behaviors to effectively communicate its intentions, including speech, body motion, and facial expressions. We present XBot-Omni, a Thinker--Talker--Actor architecture that extends the Omni paradigm to robotic modalities, including body motion and facial expressions, with ``X'' representing the extensibility to robotic-specific modalities. The Thinker processes text, images, audio, and video to formulate contextually appropriate responses, which are then realized by the Talker and Actor through speech and coordinated bodily expressions, respectively. In this report, we innovatively introduce the concept of ``Actor'' as a dedicated component for embodied agents: a coordinated, streaming module that leverages the native representations of the Thinker and Talker within an end-to-end pipeline, introducing only modest latency overhead. This unified architecture enables multimodal generation conditioned on omni-modal inputs directly to prevent information loss in staged designs. We comprehensively evaluate XBot-Omni across seven benchmarks, demonstrating strong multimodal understanding capabilities on both general-domain and robot-centric videos, as well as high-quality, natural, and well-coordinated motion generation. Furthermore, our results validate its capability to generate embodied multimodal responses directly from omni inputs, meeting the diverse requirements of embodied interaction. XBot-Omni enables robots not only to perceive the omni world but also to express themselves through speech and bodily behaviors, extending Omni models from conversational interaction to embodied communication in physical environments.

cs.RO

新论文 · v1arXiv:2611.19974

Take a Step Back: A Dual-Granularity Raw Feature Empowered Intermediate Fusion Graph Convolutional Network for Multimodal Recommendation

Jinfeng Xu, Zheyu Chen, Shuo Yang, Jinze Li, Wenhao Yuan, Ziyue Peng, Puzhen Wu, Shujie Li 等 11 位作者

原文摘要

Although multimodal recommendations have shown the ability to learn user preferences and item properties by leveraging explicit multimodal information, they remain constrained by two critical limitations: low-quality representation learning and modality fusion. For the former, existing solutions primarily focus on multi-granularity representation learning; however, we argue that training such representations is difficult due to the inherent data sparsity problem in recommendation scenarios. For the latter, existing solutions often rely on contrastive learning to align representations of different modalities to facilitate modality fusion, yet alignment-only objectives may over-compress valuable modality-specific information. To this end, we suggest \underline{``Take a Step Back"} and propose a \textbf{D}ual-granularity raw feature empowered \textbf{I}ntermediate fusion \textbf{G}raph \textbf{C}onvolutional \textbf{N}etwork (\textbf{DIGCN}) for multimodal recommendation. DIGCN leverages the multimodal understanding capabilities of Large Language Models (LLMs) to construct dual-granularity raw features directly. Furthermore, DIGCN introduces an intermediate fusion graph convolutional network equipped with dynamic cross-modality weight correction, designed to capture cross-modality signals and achieve effective modality fusion. Comprehensive experiments on four real-world datasets validate that DIGCN consistently outperforms a variety of strong baselines. Additional analyses confirm its superiority under varying levels of data sparsity and cold-start scenarios. Notably, DIGCN demonstrates high extensibility, capable of further benefiting from additional contrastive learning-based alignment tasks.

cs.IR

新论文 · v1arXiv:2611.19937

AllocFlow: End-to-End Portfolio Trading Action Generation from Raw Market Data

Annan Wang, Yuchen Xie, Yu Cen, Leon Liangyu Chen, Qianyu Julie Zhu, Xin Li, Tianyi Wei, Weisi Lin

原文摘要

Real-world portfolio management requires converting noisy market observations into coordinated, simultaneous allocations across multiple assets under shared capital constraints. Advances in forecasting models and alpha discovery help identify opportunities. However, traditional pipelines often decouple signal forecasting from portfolio construction, potentially misaligning learning with allocation and execution objectives. We present AllocFlow, an end-to-end flow-matching framework that directly generates joint portfolio allocations from raw multi-asset market histories. To capture dependencies among portfolio positions, AllocFlow employs a cross-asset flow Transformer conditioned on temporal market representations to generate continuous, signed position weights across all instruments jointly. To address the sample inefficiency and instability of learning continuous allocation policies from scratch, we establish a two-stage training scheme: strategy-supervised imitation shapes a structured distribution of candidate portfolios, followed by group-relative reinforcement learning (RL) that refines joint decisions using portfolio-level financial rewards and a frozen policy reference. On hourly cryptocurrency perpetual futures, RL refinement raises the annualized net return of the imitation policy from $27.79\%$ to $51.38\%$ and its Calmar ratio from $0.955$ to $4.043$, surpassing the teacher strategy on every metric, and AllocFlow reaches a Calmar ratio of $4.90$ on six further months of unseen 2026 data without retraining.

cs.LG · q-fin.TR

新论文 · v1arXiv:2611.19912

AgenticJevNav: Agentic Object Search and Navigation via Event-Driven Verification

Shuning Zhang, Shuli Zhang, Yunheng Wang, Wei Lu, Jiaxi Zhang, Yihang Kang, Tao Wang, Xiangyu Chen 等 11 位作者

原文摘要

Agentic navigation must maintain goals and adapt actions from incomplete evidence despite execution failures and delayed inference. We present AgenticJevNav, which links goal-directed planning, evidence memory, event-triggered verification, and bounded execution. The Agent plans from RGB, odometry, memory, and feedback, retaining goal ownership while authorizing a bounded action set. Jev reviews structured evidence after blocked/failed execution or two consecutive low-progress segments. Version, freshness, and once-only checks govern decision application. A single executor and independent RGB reflex constrain motion. Goal revisions revoke obsolete requests and actions before replanning. Across six simulated task families, AgenticJevNav achieves 73.33% SR over 60 trials, 21.67 percentage points above the next-best Agent-only aggregate control under differing historical protocols. A real-world GENISOM L1 quadruped deployment achieves 70.0% success over 30 trials, including bounded search reporting. Together, the simulation and physical-robot results demonstrate the effectiveness of the proposed architecture, while remaining failures highlight opportunities to improve grounding, approach planning, and runtime.

cs.RO · cs.AI

新论文 · v1arXiv:2611.19904

DiffWAM: Learning World Action Model from Consecutive Visual Difference

Xinheng Zhang, Haoyu Xiong, Chao Jing, Xiaohan Zhang, Linshan Li, Hongbang Fan, Xingyu Chen

原文摘要

World Action Models (WAMs) have emerged as an important paradigm in robot learning, jointly predicting future scene states and robot actions from current observations and instructions. However, future states entangle task-related changes with large amounts of task-irrelevant content, such as static backgrounds, which weakens the model's ability to capture changes relevant to task execution. To address this, we propose Difference World Action Model (DiffWAM), a Mixture-of-Experts (MoE) framework that enhances world representations by jointly learning future-state prediction and consecutive visual difference prediction. We define consecutive visual differences as differences between the latent representations of consecutive observations, providing an explicit learning signal for temporal variation. The framework comprises a World Expert for future frame prediction, an Action Expert for action generation, and a Diff Expert for predicting these latent differences along the trajectory. A structured unidirectional attention mask conditions the Diff, World, and Action Experts solely on the current-frame representation while blocking attention between experts. This design allows both the World and Diff Experts to be removed during inference, introducing no additional inference overhead. DiffWAM achieves state-of-the-art performance on the LIBERO and RoboTwin 2.0 benchmarks and substantially improves over FastWAM on real-world manipulation tasks. Specifically, it attains average success rates of 98.8%, 92.65%, and 88.3%, outperforming the state-of-the-art baseline (i.e., FastWAM) by 1.75, 1.55, and 15.0 percentage points, respectively.

cs.RO

新论文 · v1arXiv:2611.19847

Reflection Is Not Always Worth It: Budgeted Reflection Allocation for Recommendation Agents

Jinfeng Xu, Zheyu Chen, Ziyue Peng, Jianheng Tang, Zheng Lin, Wenhao Yuan, Jian Chen, Shujie Li 等 9 位作者

原文摘要

This work aims to make recommendation agents more efficient by reducing unnecessary LLM reflection while preserving full-catalog ranking and inference reranking. Since BRAVE changes only how training-time reflection is allocated, it does not introduce new user data requirements or additional LLM exposure beyond the base agent protocol. The main potential risk is that a poorly calibrated allocator may assign less reflection to users whose preferences are sparse, niche, or under-represented, which could weaken personalization for those users or indirectly affect item exposure. This risk can be mitigated through deployment-time auditing and conservative allocation design. In practice, reflection budgets should be evaluated not only by aggregate Recall or NDCG, but also across user-history lengths, item popularity groups, and application-relevant cohorts when such analysis is permitted. The allocator should rely on task-side ranking evidence rather than sensitive attributes, and fallback policies can reserve a small portion of the budget for random or fairness-aware coverage when subgroup robustness is important. Privacy risks can be controlled with the same protections required by the base recommendation agent: minimizing user text sent to the LLM controller, anonymizing user identifiers, retaining only necessary interaction evidence, and applying standard access-control and logging policies. Overall, BRAVE is designed as a cost-aware allocation layer, and its ethical deployment depends on validating that the saved computation does not come at the expense of vulnerable users or responsible exposure.

cs.IR

新论文 · v1arXiv:2611.19803

Pros4D: Modeling Evolving 4D Worlds in Latent Space for Dynamic Spatial Reasoning

Rongxin Gao, Dongxuan Liu, Chu Li, Chengyang Ji, Jialing Huang, Runyu Chen, Kairun Wen, Chenxin Li 等 15 位作者

原文摘要

Dynamic spatial reasoning from monocular video enables vision-language models (VLMs) to understand how 3D environments evolve over time, supporting physical world understanding and embodied intelligence. Existing efforts primarily address observed-dynamics understanding, whereas agents in dynamic environments also require future-dynamics prediction beyond the observation boundary. Moreover, external geometric processing and textual descriptions can increase inference complexity or lose information when representing continuous dynamics. These limitations motivate a shared internal state that preserves observed dynamics and remains predictive beyond the input video. Accordingly, we present Pros4D, a framework for modeling evolving 4D worlds in latent space that unifies retrospective understanding of observed-dynamics with prospective reasoning for future-dynamics prediction. Specifically, Pros4D-Engine derives geometry-grounded supervision and constructs paired examples for both tasks, while Pros4D-Bench evaluates them across eight spatial task categories spanning dynamic cameras, objects, and scenes. At the model level, Pros4D compresses observed dynamics into latent observed states and autoregressively rolls them forward into latent future states. 4D Latent Alignment (4DLA) grounds these states through complementary visual-semantic and geometric supervision, while joint language supervision connects them to task reasoning. Experiments across benchmarks show consistent improvements over strong baselines in both capabilities.

cs.CV

新论文 · v1arXiv:2611.19771

JEPA4R: Embodied JEPA World Models with 4D Point Planning

Jiarui Dong, Zhiheng Liang, Yanjun Li, Zeyu Zhang, Heejeong Nam, Haoyu Chen, Hao Tang

原文摘要

Joint-embedding predictive architecture (JEPA) world models predict future embeddings instead of pixels and plan in latent space. Two challenges remain: 2D images leave the metric geometry that manipulation acts on implicit and entangled with appearance, and what these models imagine stays hidden, as they are mostly judged in simulation by short-horizon latent planning. We present JEPA4R, a JEPA world model that observes single-depth-camera point clouds, predicts and plans in their latent space, and acts through a policy head on its frozen latents; we call this 4D point planning. A decoder on the frozen world model shows what it imagines as 4D point clouds, with a mean point error within 1.1 cm over 200 steps. Under the policy-learning protocols of OGBench and CortexBench, JEPA4R leads pixel JEPA models on OGBench Cube-Single, Cube-Double, Cube-Triple and Scene, and ranks first on MetaWorld and Adroit Relocate with 98.4% and 88.0%, above the runner-up, R3M, at 96.0% and 65.3%. Trained only in a digital twin, JEPA4R and its policy heads run zero-shot on the real arm with both a gripper and a dexterous hand.

cs.RO · cs.CV

新论文 · v1arXiv:2611.19577

ElevatorII: High-Performance Fully-Static Cross-ISA Binary Translation

Hongyu Chen, James McGowan, Michael Franz

原文摘要

The established method for running software compiled for one instruction-set architecture (ISA) on a different ISA is by emulation in conjunction with just-in-time (JIT) compilation. This requires run-time support on the target that needs to be trusted and whose compilation decisions are hard to predict or reconstruct. We present ElevatorII, a fully static, ahead-of-time cross-ISA binary-to-binary compiler from x86 to AArch64 that challenges the established consensus. Our approach eliminates the need for run-time support on the target, removes the hard to reproduce effects of dynamic compilation, and provides the capability to exhaustively test the actual binary that is eventually executed. Our approach combines superset disassembly with observation-guided path selection. Superset disassembly overcomes the code vs. data differentiation problem by creating a separate control flow path for every possible interpretation of every byte in the input binary. It guarantees that there is pre-translated code for every possible jump destination in the original text segment, even for computed jumps that are not statically decidable. Static and observation-guided path selection then creates additional optimized "confirmed code paths" that overlay the regular grid. On the 11 SPEC CPU 2026 Integer benchmarks that do not use C++ exceptions, ElevatorII runs at 2.5x native AArch64 time, on par with Box64, the fastest dynamic translator, and 5x faster than QEMU. It also translates binaries obfuscated with virtualization and branch functions, which no static disassembler can follow. Our approach trades execution performance for space: our output code is 51x the size of code compiled from source. We alleviate this by separating cold code to a .text.unlikely ELF section: only 2-3% hot code winds up in the main ELF section, while the rest will likely never be loaded into memory.

cs.CR · cs.PL

版本更新 · v2arXiv:2610.23203

SimpleVLN: Towards Simple and Frontier Vision-Language Navigation

Lingyi Hong, Weiren Cai, Jiayu Chen, Xinyu Zhou, Bailin He, Yunfei Guo, Qiao Sun, Boyu Ma 等 10 位作者

原文摘要

Vision-and-Language Navigation (VLN) requires embodied agents to ground language in visual observations and make long-horizon decisions. Existing methods often address these demands through task-specific architectures. As multimodal large language models (MLLMs) become increasingly capable, an important question remains: Have we truly unlocked the potential of native MLLMs for VLN before introducing task-specific architecture? We present SimpleVLN, an architecture-preserving framework developed through studies of navigation formulation, data construction, and optimization. At the formulation level, SimpleVLN combines single-turn decision inputs with fixed-budget global-local memory to efficiently use navigation history without unbounded context growth, while action chunks enable efficient output of multiple primitive actions per model invocation. For data construction, we combine expert demonstrations with policy-dependent DAgger experience and develop a scalable synthesis pipeline that generates large-scale tasks for reinforcement learning. At the optimization level, a progressive IL-DAgger-RL pipeline advances from action imitation to task-level outcome optimization. Comprehensive experiments validate the effectiveness of these three aspects. We instantiate SimpleVLN across multiple MLLM backbones and model sizes, demonstrating its generality. SimpleVLN achieves 76.7% and 80.1% success rates on the R2R-CE and RxR-CE splits, respectively, substantially outperforming existing methods. These results show that the navigation capabilities of native MLLMs can be effectively unlocked without task-specific architectural designs.

cs.RO

新论文 · v1arXiv:2611.19261

dit.xpu : Taming DiT Sparse Inference on Mobile SoCs

Jiayu Chen, Jinye Hu, Maoliang Li, Anqian Cen, Hailong Zou, Zihao Zheng, Xinming Wei, Guojie Luo 等 11 位作者

原文摘要

Diffusion Transformers (DiTs) enable on-device image generation and editing but incur high attention costs at high resolutions. Existing sparse attention reduces computation, yet dynamic sparsity introduces preprocessing and irregular data accesses that poorly fit mobile NPUs. We present DiT.XPU, a sparse DiT inference framework for heterogeneous mobile SoCs. DiT.XPU combines tile-aligned sparse attention with anchor-based error compensation by exploiting slowly varying dense-sparse output residuals across denoising steps. Heterogeneous XPU collaboration overlaps NPU dense attention with GPU sparse calibration, while mask-guided dataflow reduces data transfers and layout conversion by overlapping data movement with computation. Across five DiT models on two Snapdragon smartphones, DiT.XPU achieves up to 1.56x end-to-end denoising speedup over the Hexagon-NPU backend and 11.5x attention-kernel speedup over the fastest alternative, while preserving generation quality close to dense attention. Code is available at https://github.com/IF-LAB-PKU/dit.xpu.

cs.CV

版本更新 · v2arXiv:2611.17332

ArchSpace:记录、评审与引用 LLM 架构研究

ArchSpace: Recording, Reviewing, and Referencing LLM Architecture Research

Jie Wang, Tong Zhu, Jiahao Kuang, Zhe Xu, Mowen Ruan, Yunkai Chai, Yuliang Liu, Dehong Zheng 等 14 位作者

AI 速读 基于摘要

ArchSpace 是一个公开平台,用于记录 LLM 架构提案及其实验与推理、由社区评审,并供后续工作引用,同时借助缩放定律在小规模实验上预测更大规模的模型损失。

cs.CL · cs.LG

查看原文摘要

Model architecture shapes the scaling behavior, training efficiency, and inference cost of LLMs, making it a central focus of both industry and academia. In industry, teams such as those behind DeepSeek, Qwen, and Kimi routinely introduce architectural innovations, yet public reports rarely document which ideas succeeded, which were abandoned, and why. In academia, limited compute and inconsistent baselines often restrict evaluations to small-scale, single-point comparisons, making it difficult to establish whether improvements hold at larger scales or translate into broader impact. ArchSpace addresses both gaps as a public platform for Recording, Reviewing, and Referencing Architecture Research: each proposal is recorded with its experiments and reasoning, reviewed by the community, and referenced by later work. To evaluate ideas before costly large-scale training, ArchSpace leverages scaling laws to predict model loss at larger scales from smaller experiments, a practice widely adopted in industry. We provide initial references that make the platform usable: Qwen3.5 and OLMo3 baselines, fitted scaling laws, and a reusable training recipe. Ideas validated through this review can in turn become new baselines, and ArchSpace thereby aims to broaden the range of ideas that shape the next generation of models.

版本更新 · v4arXiv:2610.13357

ECGGuard: Deep Watermarking of Electrocardiogram Signals With Error-Correcting

Kuan-Yu Chen, Kuan-Chen Wang, Yu Tsao, Jian-Jiun Ding

原文摘要

Wearable electrocardiogram (ECG) recordings pass through multiple networks and systems before automated interpretation, leaving patient metadata vulnerable to loss. Digital watermarking can embed this metadata in the signal, but conventional methods fail under time shifts, and high waveform similarity does not guarantee the same downstream diagnosis. We present \rev{ECGGuard}, a deep watermarking framework for single-lead ECG. A fully convolutional generator embeds the payload without signal quantization, a detector decodes the message frame by frame and checks for watermark presence, and during training a frozen arrhythmia classifier penalizes changes that would alter its predictions. Because high bit accuracy does not ensure whole-message recovery, the payload is protected by extended Hamming codes, \rev{a validation-selected channel ordering}, and soft-decision decoding. On the MIT-BIH Arrhythmia Database, \rev{averaged over three training runs, ECGGuard} keeps bit accuracy above \rev{98\%} under time shifts of 10 to 50 samples, where transform-domain baselines drop to chance. The coded 16-bit payload is recovered exactly in \rev{99.95\%} of clean segments\rev{, in over 96\%} under time shifts, 10\% sample deletion, or 10\% packet loss\rev{, and in 99.1\% for unseen subjects}. \rev{For the guiding classifier, the diagnostic loss limits the drop in ventricular-beat sensitivity to 1.6 percentage points on the test records (35.0 without it) and to 12.2 points for unseen subjects (67.0 without it); the drop grows to 10.3 points for independently trained classifiers, and the protection does not transfer to an external database.} \rev{ECGGuard} has 0.56 million parameters and needs about \rev{19}~ms per segment on a desktop CPU.

eess.SP · cs.CR · cs.CV

新论文 · v1arXiv:2611.17797

Vision-as-a-Tool: Visual Harness for Multimodal Agentic Coding

Haolin Yang, Wenshan Wu, Jin Jiang, Yu Cheng, Yan Xia, Ting Song, Yangyu Huang, Zhiliang Peng 等 16 位作者

原文摘要

Multimodal agentic coding requires robust visual reasoning to analyze observations and evaluate generated artifacts. Current frameworks largely rely on a Thinking-with-Images paradigm, where agents depend on their intrinsic vision to execute visually grounded tasks. However, this architectural coupling introduces a fundamental bottleneck. Agent performance remains strictly bounded by internal perceptual limitations. Furthermore, text-only agents lack native perception. Without specialized tools, they cannot perform visual reasoning. To overcome these constraints, we introduce the Vision-as-a-Tool (VaaT) paradigm. Prior visual tool-use merely manipulates images for an agent's intrinsic vision to read. In contrast, VaaT reframes general-purpose multimodal large language models (MLLMs) as callable external utilities. As a result, text-only agents can completely offload visual reasoning during agentic execution. Meanwhile, vision-capable agents can dynamically augment their native perception on demand. We instantiate this paradigm in VHarness, a unified framework driven by five atomic actions. We comprehensively evaluate VHarness across Image-to-HTML, Image-to-SVG, and Image-to-3D generation tasks. Experimental results demonstrate that VHarness consistently improves visual fidelity and geometric accuracy across both text-only and multimodal coding agents. These findings highlight that modularizing visual reasoning into callable utilities provides a flexible and scalable foundation for future agentic coding systems.

cs.AI · cs.CV

新论文 · v1arXiv:2611.16677

Manifold Attracting Flow Matching: Physics-Constrained Sampling via Continuous Geometric Dynamics

Guanyu Chen, Jinchen Wang, Shengze Xu, Xingyu Ren, Tieyong Zeng, Antonios Armaou, Dong Ni

原文摘要

Generating scientific data under previously unseen equality constraints requires satisfying the prescribed constraints without destroying the physical structure learned from data. Existing inference-time approaches often enforce physical constraints by repeatedly correcting predicted endpoints, but such discrete corrections can distort learned generative structure and introduce unstable feedback. In this work, we propose Manifold Attracting Flow Matching (MAFM), which continuously evolves a persistent anchor through an ODE whose dynamics drive it analytically toward the constraint manifold. Specifically, its geometric correction preserves model-guided tangent motion while prescribing the normal motion to contract constraint violations toward the manifold. This construction separates model tracking from constraint attraction and provides theoretical guarantees for residual contraction and local feedback stabilization. Experiments across geometric and partial differential equation benchmarks demonstrate improved generation quality over variant constraint-guided baselines while maintaining low constraint residuals. MAFM provides a flexible framework for coordinating learned physical structure and explicit constraints through controllable sampling dynamics, without retraining the underlying generative model.

cs.LG

新论文 · v1arXiv:2611.16433

daVinci-Prover: Agent-Native Mid-Training for Formal Theorem Proving

Wenhong Zhu, Xuefeng Li, Jiashuo Zhang, Yu Chen, Tiantian Mi, Qipeng Guo, Pengfei Liu

原文摘要

Advances in large language models (LLMs) are rapidly transforming formal theorem proving. Yet further progress remains constrained by the lack of a transparent and effective data recipe for systematically improving theorem-proving models. In this work, we argue that effective agentic provers require strong mathematical skill priors to ground theorem proving and agentic trajectories to enable long-horizon, interactive proof construction. To this end, we synthesize an agent-native 33.7B-token mid-training corpus covering decomposition, search, self-verification, and revision, together with verifier-grounded agentic trajectories. We further introduce the daVinci-Prover harness, which enables models to construct proofs and interact with a Lean verifier in an isolated workspace, and use it to collect approximately 36K high-quality agent trajectories for supervised fine-tuning. Our mid-training recipe yields an average improvement of 13.7 points over the matched Qwen3-8B SFT-only baseline across MiniF2F, FATE-M, and ProofNet. When applied to the Qwen3.5 series, our approach also yields strong performance across theorem-proving benchmarks and achieves near-saturating MiniF2F performance with memory compaction. The results highlight agent-native mid-training as an effective foundation for developing stronger interactive theorem-proving agents.

cs.CL

版本更新 · v2arXiv:2611.14009

Coding Agent Memory Post-Training II: Unlocking the Memory Potential of Pre-trained File Operations for Long-Horizon Tasks via On-Policy Distillation

Lirui Luo, Kelong Mao, Xinwei Yang, Rongqing Li, Luyu Chen, Kieran Wong, Yudong Guo, Xinrui Wang 等 13 位作者

原文摘要

Language-model agents increasingly tackle long-horizon tasks whose interaction histories exceed the model's active context. They must therefore decide what to write down and what to read back, and when to do either. Recent work has begun to use on-policy distillation (OPD) to make memory control part of the policy, relying on memory tools and compaction protocols that lie outside either model's pre-training. However, this weakens the distillation on both sides: the teacher cannot give useful guidance on an interface it has never used, and the student samples too few meaningful memory actions for that guidance to take effect. To study this problem, we introduce Coding Agent Memory Gym II (CAMG-II), a suite of long-horizon agentic-RL environments spanning Shop, Coding, AutoResearch, General, and Visual. Alongside each environment's native task interface, CAMG-II provides ordinary Bash access and an episode-persistent workspace, enabling the policy to create, revise, search, and reuse files as memory across context replacements. We also introduce Coding Agent Memory Post-Training II (CAMP-II), which trains a single policy jointly across all five environments with fully asynchronous OPD to learn this file-based memory behavior directly from a frozen teacher with a prior for file operations. On CAMG-II, CAMP-II obtains 2.97 times the gain of the strongest OPD baseline. The memory behavior CAMP-II learns generalizes to two external benchmarks, Terminal-Bench 2.1 and SWE-Bench Pro V2, with improved success rates.

cs.LG · cs.AI · cs.CL

新论文 · v1arXiv:2611.16374

ManiLadder: Benchmarking Robot Manipulation Through a Categorized and Multi-Difficulty Level Task Ladder

Chongkai Gao, Jieao Shi, Zhaoyu Mu, Yudi Lin, Xuanwei Liu, Wei Ken Ng, Baiye Cheng, Anqi Chen 等 17 位作者

原文摘要

Recent advances in robotic manipulation have enabled increasingly complex and dexterous behaviors, yet existing benchmarks still focus on simple pick-and-place tasks and lack a systematic framework for measuring progress in robot learning along the axis of task difficulty. We present ManiLadder, a large-scale and organized simulation benchmark for robotic manipulation. ManiLadder contains 112 carefully designed tasks stratified into 4 difficulty levels, with 8 categories in each level, spanning rigid-body and deformable-object manipulation, single-arm and dual-arm settings, and gripper and dexterous-hand end-effectors. To calibrate difficulty, we fix the data amount the same for every task and use a unified flow- matching imitation learning policy under the same training and evaluation protocols, then use its evaluation success rate as an operational empirical difficulty signal. All tasks are iteratively refined until they fall into target difficulty ranges, which are then verified across different random seeds, policies, and agentic learning systems. We further analyze the task properties, compare different policy performances and training settings, analyze task transfer and data scaling behaviors, and conduct real-world experiments showing that the designed tasks preserve their difficulty ordering beyond simulation. ManiLad- der provides a calibrated ladder for measuring, comparing, and advancing robotic manipulation capabilities. More results are available at https://maniladder-benchmark.github.io/.

cs.RO

版本更新 · v2arXiv:2611.11083

Who Benefits from Wikipedia's Growth Mentorship? The Role of Newcomer Information Needs

Yubo Zhou, Pengyu Chen, Ceren Budak

原文摘要

Evaluations of Wikipedia's programs for new editors usually ask whether a program works on average; we ask which newcomers benefit when their question to a mentor gets a reply. English Wikipedia's Growth mentorship system assigns every new account a volunteer mentor, and about a quarter of the questions newcomers send never get a reply. Using the first questions of 34,073 newcomers (2021 to 2026), we compare newcomers who received a reply with newcomers who did not, within groups that are similar on 164 characteristics recorded before the question (propensity score stratification). (1) Does a reply help newcomers keep editing? Yes, in the short term: in the 14 days after the reply, newcomers who received one were 1.8 percentage points (pp) more likely to edit an article (22% vs. an estimated 20% without a reply) and 1.8pp more likely to edit on two or more days. Later in the first year, the only difference that remains significant after correction for multiple testing is in the share who edit an article at least once between six months and one year after the reply (+1.5pp). (2) Does the effect depend on what the newcomer asked? Yes. We sort questions by Morrison's (1993) information-seeking types, which earlier studies measured mostly through self-reports, using a language model checked against human labels. A reply has significant effects for questions about rules and norms (+3.4pp), questions that show the newcomer's earlier editing (+3.4pp) and specific how-to questions (+2.4pp). For newcomers who do not know what to work on, a reply has no detectable effect, and the effect is 4.2pp smaller than for newcomers with other questions. (3) Which replies are followed by more editing? Among newcomers who received a reply, those whose reply came sooner or stayed closer to the topic of their question edited more.

cs.SI · cs.HC

版本更新 · v4arXiv:2611.01666

Stand-In Effects and Pooling Amplification: Lessons from Biological World Models for Testing Action Effects

Bowen Xu, Boyu Chen, Jiaying Li

原文摘要

Whether an action-conditioned world model responds to actions is usually checked by replacing or removing the actions and measuring the loss in accuracy. Taking biological experiments as the guiding case, we show that such diagnostics report action effects where there are none: under autocorrelated designs, future actions stand in for history the model does not see. When the design law is known, a conditional randomization test written as an e-value is valid in finite samples whatever the model. A second failure arises whenever many sites share one action sequence, as in world models: the pooled score turns small model errors along the action contrast into an e-value near zero, so a valid test loses its power although the effect is real. Tempering the score changes the bet, not the validity. The log e-value is concave in the tempering factor and zero at zero, so too small a factor costs little and too large a one can cost everything; every single factor we tried, estimated or tuned, failed in some setting, and mixtures over a fixed grid never did. Our own preregistered analysis of optogenetic ERK experiments failed by pooling; a tempered statistic, registered before the data were obtained, then rejected the null on a never-opened field (E = 1.1e10). In three registered tests on models trained by others, the pooled test rejected in 0 of 96 units of a diffusion model, and reached the power bar in 10 of 26 Atari games under IRIS's token likelihood and 1 of 6 control tasks under TD-MPC2; mixtures fixed in advance reached it in 96 units, 24 games and 6 tasks. On recorded robot video with iVideoGPT the mixture rejected in 155 of 159 trajectories; the pooled test did not collapse there (143). Under exact nulls the statistics made 14 rejections in 26,664 tests.

cs.LG · q-bio.QM · stat.ME

新论文 · v1arXiv:2611.14908

Hy ASR 3.0 Technical Report

Aoxiong Yin, Auden Li, Chao Huang, Chuanyang Hu, Eitan Wang, Ethan Zou, Fangming Feng, Guangyu Chen 等 63 位作者

原文摘要

We present Hy ASR 3.0, a large language model (LLM)-based automatic speech recognition (ASR) system that delivers strong recognition accuracy, supports flexible contextual conditioning, and maintains robust recognition performance across diverse speech conditions. Hy ASR 3.0 supports Mandarin Chinese, English and 13 Chinese dialect groups, including Cantonese. Among the evaluated systems, it achieves the lowest aggregate error rates on public Mandarin Chinese, English, and Cantonese benchmarks and the lowest error rates on the corresponding internal test sets, alongside first-tier dialect recognition performance. Beyond audio-only transcription, Hy ASR 3.0 supports multiple forms of external context, including hotword lists, question--answer dialogues, and summaries, enabling the use of task-relevant information during recognition. The model also leverages intra-utterance context to disambiguate homophonic and phonetically similar keywords using information from other parts of the same utterance, without requiring a separate contextual input. Evaluations demonstrate strong contextual recognition in both external-context and intra-utterance settings, as well as robust recognition of long-form speech, children's and older adults' speech, low-volume speech, and speech in noisy environments. This report describes the model architecture and training methodology and presents evaluations across recognition accuracy, context utilization, and robustness.

cs.CL · eess.AS

版本更新 · v2arXiv:2608.04222

TIDE: A Physically Diverse 3D Turbulence Benchmark Dataset for Advancing Scientific Machine Learning

Yilong Dai, Yiming Sun, Yiheng Chen, Shengyu Chen, Peyman Givi, Xiaowei Jia, Runlong Yu

原文摘要

Turbulence is a central testbed for machine learning on physical dynamics because its governing laws are known exactly. However, most existing studies remain in 2D, while 3D turbulence has fundamentally different physics and is far more costly to simulate. Existing 3D resources also typically provide only one realization per configuration, making it difficult to distinguish learning the dynamics from fitting the statistics of a single flow. In this paper, we introduce TIDE (Turbulent Incompressible DNS Ensembles), a 256^3 DNS corpus and benchmark for 3D incompressible turbulence, with 15 configurations on eight controlled axes, independent ensembles, pressure fields, and equation-level verification. The benchmark includes five tasks, standardized learned baselines, controlled generalization splits, and physical-fidelity metrics alongside pointwise error. Across the main forecasting configurations, current learned models trained under a fixed reference recipe barely outperform persistence and still make about twice the error of a spectral solver given the true equations. Moreover, lower pointwise error can coincide with severely distorted small-scale dynamics, showing that accuracy alone does not ensure physical fidelity. Rollout stability depends on the configuration, the training seed, and how densely training samples each trajectory. Joint training across neighboring regimes does not close the transfer gaps we test, and the forced-to-decay contrast bounds a conditioning limit: an estimate of the current forcing state removes 20-41% of the error a spectral solver makes without forcing input, while history or forcing inputs help only some learned architectures. Closing these accuracy, fidelity, and conditioning gaps is the central open problem made measurable by TIDE. https://github.com/Dyloong1/TIDE-dataset-benchmark

physics.flu-dyn · cs.LG

新论文 · v1arXiv:2611.14469

C²-Routing:面向高效视觉语言推理的互补裁剪与压缩

C$^2$-Routing: Complementary Cropping and Compression for Efficient Vision-Language Inference

Guorui Lu, Qinyu Chen

AI 速读 基于摘要

提出 C²-Routing,根据图像与问题自适应选择裁剪或压缩来减少视觉 token,并用安全门控制精度损失。

cs.CV · cs.LG

查看原文摘要

Visual Question Answering (VQA) imposes substantial computational overhead on vision-language models (VLMs), motivating model-agnostic input reduction methods that can be readily applied across diverse VLMs, including in black-box settings. Simply enforcing a visual-token cap reduces computation, but it can discard information needed to answer the question. A single reduction strategy also struggles to accommodate different scenarios. Our key observation is that cropping and compression are complementary: cropping highlights small targets and local details, whereas whole-image compression better preserves context for large targets or distributed evidence. Based on this observation, we propose C$^2$-Routing, which adaptively selects cropping or compression according to the input image and question. Since both strategies may still discard key information, we further design a safety gate that uses the predicted joint correctness distribution under original and reduced inputs to guide fallback and protect accuracy. To accommodate consumer GPU resource constraints, all training and evaluation experiments run on RTX 3090 GPUs. To measure gains beyond token capping, we use full-image baselines with the same cap of 1,024 visual tokens. We conduct extensive evaluation on three VLMs and 18 datasets spanning local details, global scenes, scene text, and documents. With the visual budget already capped, C$^2$-Routing further reduces end-to-end FLOPs by 36.53% and visual tokens by 42.36% on average, with a mean relative task-score decrease of only 0.77%.