微信内可能无法直接打开本站。请点右上角 ··· → 在浏览器打开 ，或复制链接后用系统浏览器访问。
综合
官方
企业
汇聚
When Self-Consistency Backfires: Majority Vote Hurts the Majority of Hard Science Problems for Small LLMs
arXiv:2608.…
11403v1 Announce Type: new Abstract: Self-consistency (SC) via majorit…
On the full GPQA Diamond benchmark (198 graduate-level science questio…
RSS 官方收录 · 可信分层展示
From Numbers to Judgment: Specialist LLM Agents and Reinforcement Learning for European Listed Real Estate
arXiv:2608.…
11381v1 Announce Type: new Abstract: We study whether the localized nu…
Larix maps a 16-lens European listed-real-estate analysis framework to…
RSS 官方收录 · 可信分层展示
Can Frontier LLMs Match Natively Multimodal Embeddings? A Comparison on Hard-Negative Text-to-Image Retrieval
arXiv:2608.…
11343v1 Announce Type: new Abstract: Multimodal retrieval and classifi…
The March 2026 release of Gemini Embedding 2, Google's first natively …
RSS 官方收录 · 可信分层展示
Inverse Theory of Mind Modeling for Content Recommendation: From Web Browsing to Dynamic Intelligent Interfaces
arXiv:2608.…
11354v1 Announce Type: new Abstract: Modern recommender systems treat …
As interfaces evolve from static layouts toward generative UIs and imm…
RSS 官方收录 · 可信分层展示
Apodex Discovery: Reality Benchmarks and Environments for Evaluating and Building Discoverative Artificial Intelligence
arXiv:2608.11341v1 Announce Type: new Abstract: Apollo did not reach the Moon merely because its engineers could solve difficult equations.…
It succeeded by turning a distant ambition into a mission architecture…
AI now faces a similar transition: frontier models can solve difficult…
RSS 官方收录 · 可信分层展示
Deployment Decision Reliability: A Generalizability-Theory Framework for Sizing Long-Horizon Agent Evaluations
arXiv:2608.11323v1 Announce Type: new Abstract: Enterprise practitioners read agent leaderboards as if they ranked agent capability.…
We show, across three open agent-trace benchmarks (TheAgentCompany, $\…
Leaderboards rank specialization, not capability.
RSS 官方收录 · 可信分层展示
Glance, Scrutinize, and Think: Advancing Video Anomaly Detection from Training-Free to Agentic Reasoning
arXiv:2608.11260v1 Announce Type: new Abstract: Video Anomaly Detection (VAD) aims to identify anomalous events and localize their temporal intervals.…
Existing approaches exhibit a "when-what" dissociation: traditional DN…
We attribute this to the absence of a unified reasoning paradigm.
RSS 官方收录 · 可信分层展示
Adaptive Hybrid Particle Swarm Optimization with Gradient Descent
arXiv:2608.…
11258v1 Announce Type: new Abstract: Gradient injection helps Particle…
We propose Adaptive Hybrid PSO (AHPSO), which uses a sigmoid function …
RSS 官方收录 · 可信分层展示
Symbolic Machine Learning for Vapor-Liquid Equilibrium Prediction in Cx-N2 Binary Mixtures
arXiv:2608.…
11255v1 Announce Type: new Abstract: Accurate prediction of vapor--liq…
While deep learning models can provide accurate predictions, they ofte…
RSS 官方收录 · 可信分层展示
Local verification cannot detect non-transportability: a cohomological theory of context preservation in agentic reasoning
arXiv:2608.…
11252v1 Announce Type: new Abstract: Agentic AI systems routinely tran…
We prove this class of safeguard is structurally incomplete.
RSS 官方收录 · 可信分层展示
AgonAlpha: Autonomous Alpha Discovery via Prompt Economy and Scalable Agentic Search
arXiv:2608.…
11250v1 Announce Type: new Abstract: Language models can propose many …
We present AgonAlpha, an architecture that searches over frozen resear…
RSS 官方收录 · 可信分层展示
EvoGraph-Mem: Failure-Aware Editable Graph Memory for Long-Term Language Agents
arXiv:2608.11248v1 Announce Type: new Abstract: Long-term memory is essential for language agents operating across extended interactions and evolving tasks.…
Existing memory-augmented agents mainly focus on storing and retrievin…
In particular, previously distilled insights can become outdated, over…
RSS 官方收录 · 可信分层展示
Conformity Mitigations in Large Language Models Lie on a Single Resistance-Receptivity Frontier
arXiv:2608.…
11247v1 Announce Type: new Abstract: Recent advances in language model…
Each agent sees what the others assert before it answers, so peer opin…
RSS 官方收录 · 可信分层展示
Towards the Harness of Embodied Agents
arXiv:2608.…
11246v1 Announce Type: new Abstract: The success of coding agents has …
We ask whether the same paradigm extends to embodied agents in the phy…
RSS 官方收录 · 可信分层展示
Towards Sustainable Learning in Online Education: A Reinforcement Learning Approach
arXiv:2608.…
11245v1 Announce Type: new Abstract: Online education offers unprecede…
To address these challenges, we introduce AI Tutor, a reinforcement le…
RSS 官方收录 · 可信分层展示
BEST-KAG: Enhancing Question Answering of Building Engineering Standards with Multimodal Knowledge Graph Modeling and Large Language Model
arXiv:2608.11244v1 Announce Type: new Abstract: Construction standards are critical for building safety and sustainability.…
Existing standard application workflows rely on keyword-based document…
To address these limitations, this study develops a multimodal knowled…
RSS 官方收录 · 可信分层展示
The Off-Support Barrier: Why Semantic Safety Constraints Are Not Learning-Problem Invariants, and What Follows for Prior Design, Containment, and Verification
arXiv:2608.…
11243v1 Announce Type: new Abstract: We argue that a single structural…
, the agent does not escape its sandbox) is an off-support object.
RSS 官方收录 · 可信分层展示
RecSys Factory: Bounding LLM Agent Autonomy to Decision Points in the Industrial Recommender Lifecycle
arXiv:2608.…
11241v1 Announce Type: new Abstract: Deploying LLM agents into industr…
Any two can be maximized against the third.
RSS 官方收录 · 可信分层展示
VQ-bench: A Composable Vector Quantization Framework
arXiv:2608.11240v1 Announce Type: new Abstract: Vector quantization is an old problem but has recently become central to AI infrastructure.…
It is therefore experiencing a surge of renewed engineering and resear…
This paper provides a unified framework for developing and benchmarkin…
RSS 官方收录 · 可信分层展示
Towards Query-Agnostic RAG Evaluation via Query Coverage and Claim Verifiability
arXiv:2608.…
11238v1 Announce Type: new Abstract: Retrieval-augmented generation im…
We propose Q-CARE, a query-agnostic and fully reference-free framework…
RSS 官方收录 · 可信分层展示
CORA-Diff: Confidence-Oriented Residual Acceptance for Efficient Diffusion Language Model Inference
arXiv:2608.11235v1 Announce Type: new Abstract: Diffusion language models (DLMs) update many tokens in parallel, yet practical decoders often use a fixed denoising horizon.…
Many predictions stabilize early, but blockwise decoding continues unt…
Existing accelerators often rely on learned filters, modified scores, …
RSS 官方收录 · 可信分层展示
Geometry-aware Incremental Neural Operator for Long-Horizon PDE prediction
arXiv:2608.11237v1 Announce Type: new Abstract: Neural operators have shown strong potential for learning solution operators of partial differential equations (PDEs).…
However, long-horizon autoregressive prediction remains challenging: l…
Existing methods mainly improve state representations and operator bac…
RSS 官方收录 · 可信分层展示
InfraBench: Evaluating Infrastructure Agents Across Layers, Lifecycle, and Risk
arXiv:2608.11234v1 Announce Type: new Abstract: Managing modern computing infrastructure has become a steadily harder problem due to the ever-increasing complexity.…
Recent advances in AI agents create a timely opportunity to automate i…
We present InfraBench, a benchmark suite for evaluating AI agents on r…
RSS 官方收录 · 可信分层展示
LinearKV: One Cached State Suffices for Position-Independent Caching in Hybrid LLMs
arXiv:2608.11231v1 Announce Type: new Abstract: LLM serving is increasingly accelerated by position-independent caching (PIC).…
Existing PIC methods, however, are built for full-attention models, wh…
Hybrid LLMs break these primitives---they replace most attention layer…
RSS 官方收录 · 可信分层展示
The Edge-based Contiguous p-median Problem with Connections to Logistics Districting
arXiv:2608.…
11230v1 Announce Type: new Abstract: This paper introduces the edge-ba…
Two binary programming models are introduced, both of which incorporat…
RSS 官方收录 · 可信分层展示
Forecasting Side Effects of Activation Steering
arXiv:2608.…
11227v1 Announce Type: new Abstract: Activation steering modifies a la…
While effective, steering often produces unintended side effects on ot…
RSS 官方收录 · 可信分层展示
Synchronizing Beliefs with Second-Order Theory-of-Mind in Human-Autonomy Teams (Extended Version)
arXiv:2608.…
11229v1 Announce Type: new Abstract: Comparative feedback, asking peop…
Preference-based reward learning typically casts the human teacher as …
RSS 官方收录 · 可信分层展示
打破机器人「数据焦虑」，穹彻智能想建一条「训练数据」生产线
如果说 2025 年是具身智能爆发的一年，那么今年，行业竞争已经开始进入新的阶段。从国内外越来越多机器人公司发布新产品，到资本持续加码，具身智能正成为 AI 领域最受关注的方向之一。…
但随着行业快速发展，一个新的共识正在形成：真正限制机器人进化速度的，不再只是模型，而是数据。
与大语言模型可以利用互联网海量文本不同，机器人需要学习的是现实世界中的操作能力。
RSS 官方收录 · 可信分层展示
DeepSeek 的「斩杀线」，正在重塑 AI 行业的世界观
作者｜Techno 之王 编辑｜靖宇 8 月 4 日，「DeepSeek 斩杀线」竟然冲上微博热搜。V4 Flash 正式版凭借极致的性价比，在全球 AI 大模型的「成本-性能」坐标系上划出了一条残酷的行业生死线——比它贵的没它强，比它弱的没它便宜。…
开发者社区群情振奋，甚至外溢成了更大人群津津乐道的话题。
当看起来 AI 世界的定价权已经易手，所有人为极致性价比欢呼的时候，两天后，8 月 6 日，DeepSeek 在开放平台后台发布了一条简短的…
RSS 官方收录 · 可信分层展示
AI Agent 手机下半场，「国标」L3 只是起点
一张证书，暂时还不会影响到你「换机」节奏。作者｜张勇毅 编辑｜靖宇 时间回到 7 月下旬，如果你关注智能手机领域的话，应该会记得大量手机厂商开始官宣自己的手机，通过了国家人工智能 L3 级认证。…
这轮集体官宣的起点，是 2026 年 7 月 17 日公布的首批人工智能终端智能化分级测试结果。
按照首批测试名单，移动终端中共有 11 款产品达到 L3，包括 9 部手机和 2 台平板。
RSS 官方收录 · 可信分层展示
下滑 · j/k · m/u · i 信流 · a 稍后 · o 原文 · e 详情 · t 今日 · f 搜索