2026-09-22 AI日更 | GPT-6 Astra 一天上线视频功能，智能体同日进入芯片设计与精神科质控 链接到标题

今日 AI 主线从能力发布转向交付与治理并进：OpenAI 同日成立数学顾问组、发布全球标准路线图并扩展 Academy，GPT-6 Astra 已支持 Higgsfield 一天上线视频功能；智能体进入芯片设计与精神科质控等专业场景。Grok 代理团队提交 2500 个 PR，引发工程可靠性与审查负担争议。

📖 本期 Watch List 深度导读 链接到标题

今日有三条阅读主线。一是 OpenAI 治理与落地密集更新：数学与 AI 顾问组、全球标准路线图、Academy 新学习路径，以及 Higgsfield 借助 GPT-6 Astra 一天上线视频功能，安全、教育与商用生态同步推进。二是 arXiv 上长上下文与注意力效率研究集中出现：半径有界稀疏 prefill、MoE 路由与注意力耦合、从注意力拓扑检测幻觉等，工程团队值得深读。三是智能体进入专业深水区：芯片设计探索更高层抽象、精神科接诊引入临床质控，Naveen Rao 播客则从能源墙角度讨论算力成本。

🌐 X 平台 AI 热点快讯 链接到标题

话题 1:Z.ai Open-Sources ZCode After Secret Code Uploads 链接到标题

• 分类:AI · News
• 概况:热度时间:21 hours ago,相关帖子数:3200
• 是什么事:Z.ai 在 ZCode 代码被秘密上传后，宣布将 ZCode 开源。
• 为什么重要:这一事件凸显了 AI 模型代码的泄露与开源边界问题，可能影响代码生成模型的透明度、安全性和后续社区采用。
• 讨论概况:X 上的讨论集中在 Z.ai 开源是否为被动应对、秘密上传对模型安全的影响，以及 ZCode 的实际能力与许可证限制。

话题 2:SpaceXAI Engineer Ships 2,500 PRs Using Grok Bot Agent Teams 链接到标题

• 分类:AI · News
• 概况:热度时间:7 hours ago,相关帖子数:1200
• 是什么事:一名 SpaceX AI 工程师使用 Grok 机器人代理团队提交了 2,500 个代码拉取请求（PR）。
• 为什么重要:该事件展示了 AI 智能体在规模化软件工程中的实际应用能力，可能推动软件开发流程向人机协同自动化转变。
• 讨论概况:X 上的讨论聚焦于这些 PR 的真实质量与可维护性，争议点包括自动化提交是否加重代码审查负担、效率数据是否可信，以及该模式能否在关键生产环境中推广。

话题 3:Ling-3.0-flash-VL AI Transforms Product Photos into Marketing Plans and Prototypes 链接到标题

• 分类:AI · News
• 概况:热度时间:13 hours ago,相关帖子数:1200
• 是什么事:Ling-3.0-flash-VL 发布，能够将产品照片自动转化为营销计划和产品原型。
• 为什么重要:该技术展示了多模态视觉语言模型在商业创意生成与设计流程自动化上的新应用，可能降低营销与原型制作门槛。
• 讨论概况:X 上的讨论集中在生成质量与实际可用性，以及它是否会取代营销和设计人员，还是作为辅助工具。

今日 X 上的 AI 舆情小结 链接到标题

今日舆论主线围绕 AI 智能体和多模态模型加速进入实际生产环节展开，从代码生成、规模化软件工程到营销设计自动化均有体现。共识在于这些工具能降低门槛、提升效率，并推动人机协同自动化；但分歧也很明显：Z.ai 开源被视为泄露后的被动应对，Grok 提交的大量 PR 被质疑质量与审查负担，Ling-3.0-flash-VL 则引发“取代还是辅助”的争论。潜在风险集中在模型代码泄露与开源边界不清、自动化产出在关键生产环境中的可靠性不足，以及营销、设计等岗位可能受到的冲击。

💡 大佬观点(Influencer Insights) 链接到标题

今日大佬观点暂缺,推荐阅读 Watch List 深度内容。

📚 附录:今日 Watch List 更新源列表 链接到标题

时间窗口:最近 3 天;覆盖 22 个源;共 37 条更新

All-In Podcast (A_full) 链接到标题

• Naveen Rao: 4D Computing, AI’s Energy Wall & Beating Biology
  • 发布时间:2026-09-22 05:03 北京时间
  • 摘要:【待翻译】- The cost of a token, power contracts & the gap to close.
    • IREN is a vertically integrated AI Cloud platform, delivering data centers, compute and software for AI training and inference.
    • Oracle connects the data, applications, and infrastructure that turn AI into business outcomes—with the flexibility, choice, and control to optimize as AI evolves.
    • EY helps tech innovators scale from startup to exit to megacap.
    • We’re focused on giving every person the tools to reach their full potential and making sure the benefits of technology are distributed to all.
  • EN 要点:
    • (0:00) Welcome Naveen Rao
    • (4:45) Is energy really the problem
    • The cost of a token, power contracts & the gap to close
    • (10:24) Cutting out the middleman: abstractions, dynamical systems & a new kind of machine

Stratechery by Ben Thompson (A_full) 链接到标题

• Frontier Overhangs
  • 发布时间:2026-09-21 18:00 北京时间
  • 摘要:【待翻译】- Listen to this post:.
    • There has been, over the last week, what I think is a healthy debate about the philosophy and psychology that undergirds the views of meaningful segments of the AI community, particularly those obsessed with doomsday scenarios.
    • It is, in the end, difficult to reason with a philosophy that grants equivalent moral weight to not just all beings — human or not — who exist today, but who may ever exist in the future; this tilts the scales in such an absurd fashion towards safetyism that innovation is impossible and freedom is intolerable.
    • Worse, it taps into the psychology of religion, where dissent is not brooked and questioning the premise is heresy.
    • This analysis is in a similar vein but from the opposite direction: “Pacing the Frontier” is framed as — and I believe motivated by — a concern about safety, but it also happens to address several distinct problems faced by the frontier labs.
  • EN 要点:
    • Listen to this post :
    • Log in to listen
    • There has been, over the last week, what I think is a healthy debate about the philosophy and psychology that undergirds the views of meaningful segments of the…
    • It is, in the end, difficult to reason with a philosophy that grants equivalent moral weight to not just all beings — human or not — who exist today, but who ma…

OpenAI Blog (A_full) 链接到标题

• Advisory Group on Mathematics and Artificial Intelligence

  • 发布时间:2026-09-21 20:00 北京时间
  • 摘要:【待翻译】- OpenAI is working with an independent Advisory Group on Mathematics and Artificial Intelligence to guide the review and communication of emerging AI results.
    • This piece from OpenAI Blog explains how Advisory Group on Mathematics and Artificial Intelligence shapes the broader AI and infrastructure landscape.
    • It also surfaces practical implications for founders, operators, and investors following Advisory Group on Mathematics and Artificial Intelligence.
  • EN 要点:
    • OpenAI is working with an independent Advisory Group on Mathematics and Artificial Intelligence to guide the review and communication of emerging AI results.

• Higgsfield AI ships new video features in a day with GPT-6 Astra

  • 发布时间:2026-09-21 20:00 北京时间
  • 摘要:【待翻译】- With GPT-6 Astra, Higgsfield AI makes video ad creation easier for small businesses and brings new creative tools to market faster.
    • This piece from OpenAI Blog explains how Higgsfield AI ships new video features in a day with GPT-6 Astra shapes the broader AI and infrastructure landscape.
    • It also surfaces practical implications for founders, operators, and investors following Higgsfield AI ships new video features in a day with GPT-6 Astra.
  • EN 要点:
    • With GPT-6 Astra, Higgsfield AI makes video ad creation easier for small businesses and brings new creative tools to market faster.

• Building standards for the next phase of AI

  • 发布时间:2026-09-21 18:00 北京时间
  • 摘要:【待翻译】- OpenAI outlines a path to shared global AI standards, calling for coordinated evaluation, reporting, and governance to improve safety.
    • This piece from OpenAI Blog explains how Building standards for the next phase of AI shapes the broader AI and infrastructure landscape.
    • It also surfaces practical implications for founders, operators, and investors following Building standards for the next phase of AI.
  • EN 要点:
    • OpenAI outlines a path to shared global AI standards, calling for coordinated evaluation, reporting, and governance to improve safety.

• Expanding OpenAI Academy with new learning paths

  • 发布时间:2026-09-21 15:00 北京时间
  • 摘要:【待翻译】- Explore new OpenAI Academy learning paths for employees, developers, leaders, educators, and students to build and demonstrate practical AI skills.
    • This piece from OpenAI Blog explains how Expanding OpenAI Academy with new learning paths shapes the broader AI and infrastructure landscape.
    • It also surfaces practical implications for founders, operators, and investors following Expanding OpenAI Academy with new learning paths.
  • EN 要点:
    • Explore new OpenAI Academy learning paths for employees, developers, leaders, educators, and students to build and demonstrate practical AI skills.

• How V7 gives AI agents institutional memory

  • 发布时间:2026-09-21 08:00 北京时间
  • 摘要:【待翻译】- Using GPT-5.6, V7 turns scattered company files into context agents can use to complete complex, source-linked work.
    • This piece from OpenAI Blog explains how How V7 gives AI agents institutional memory shapes the broader AI and infrastructure landscape.
    • It also surfaces practical implications for founders, operators, and investors following How V7 gives AI agents institutional memory.
  • EN 要点:
    • Using GPT-5.6, V7 turns scattered company files into context agents can use to complete complex, source-linked work.

ArXiv cs.AI (B_intro+search) 链接到标题

• RBS-Attention: Radius-Bounded Sparse Prefill for Long-Context Large Language Models

  • 发布时间:2026-09-21 12:00 北京时间
  • 摘要:【待翻译】- arXiv:2609.20971v1 Announce Type: new.
    • Abstract: Long-context large language model inference is increasingly limited by prefill, where dense self-attention processes the entire prompt before generation begins.
    • Sparse block selection can reduce this cost, but a block centroid may hide a highly relevant token among many irrelevant ones.
    • We call this failure mode mean dilution and propose RBS-Attention, a training-free sparse-prefill method with two complementary selection branches.
  • EN 要点:
    • arXiv:2609.20971v1 Announce Type: new
    • Abstract: Long-context large language model inference is increasingly limited by prefill, where dense self-attention processes the entire prompt before generati…
    • Sparse block selection can reduce this cost, but a block centroid may hide a highly relevant token among many irrelevant ones
    • We call this failure mode mean dilution and propose RBS-Attention, a training-free sparse-prefill method with two complementary selection branches

• Attention-Aware Routing: Coupling Routing and Attention in MoEs

  • 发布时间:2026-09-21 12:00 北京时间
  • 摘要:【待翻译】- arXiv:2609.20974v1 Announce Type: new.
    • Abstract: In Mixture-of-Experts language models, the router typically selects and weights experts based on the token’s hidden state, utilizing limited contextual information.
    • We propose Attention-Aware Routing (AAR), which augments the router with temporal and spectral features extracted from a sliding window of attention weights that represent a summary of the model’s contextual state, disentangled from the hidden state.
    • Keeping the base transformer entirely frozen, we train only the routing parameters, isolating routing as the sole variable.
  • EN 要点:
    • arXiv:2609.20974v1 Announce Type: new
    • Abstract: In Mixture-of-Experts language models, the router typically selects and weights experts based on the token’s hidden state, utilizing limited contextua…
    • We propose Attention-Aware Routing (AAR), which augments the router with temporal and spectral features extracted from a sliding window of attention weights tha…
    • Keeping the base transformer entirely frozen, we train only the routing parameters, isolating routing as the sole variable

• CaLR: Causal Latent Revision for Robust Diffusion Reasoning

  • 发布时间:2026-09-21 12:00 北京时间
  • 摘要:【待翻译】- arXiv:2609.20981v1 Announce Type: new.
    • Abstract: Autoregressive (AR) models suffer from local greediness, while diffusion language models (DLMs) often lack the strict causal structure required for reasoning.
    • To combine the advantages and overcome the drawbacks of the dual, we propose Causal Latent Revision (CaLR), a framework that reformulates reasoning as constrained latent optimization.
    • By adopting a causal topology matrix (CTM) from an expert model and implicit differentiation, CaLR performs gradient-guided ``thought revision" to enforce logical consistency, enabling dynamic self-correction of intermediate steps during parallel generation.
  • EN 要点:
    • arXiv:2609.20981v1 Announce Type: new
    • Abstract: Autoregressive (AR) models suffer from local greediness, while diffusion language models (DLMs) often lack the strict causal structure required for re…
    • To combine the advantages and overcome the drawbacks of the dual, we propose Causal Latent Revision (CaLR), a framework that reformulates reasoning as constrain…
    • By adopting a causal topology matrix (CTM) from an expert model and implicit differentiation, CaLR performs gradient-guided ``thought revision" to enforce logic…

• LoRA Enhanced Contrastive Learning with SAS Vision Transformers

  • 发布时间:2026-09-21 12:00 北京时间
  • 摘要:【待翻译】- arXiv:2609.21061v1 Announce Type: new.
    • Abstract: Automatic target recognition (ATR) with synthetic aperture sonar (SAS) supports advanced naval capabilities, but deep learning is constrained by scarce target imagery, background clutter, and human-in-the-loop assessment.
    • We adapt DINOv3 Vision Transformer (ViT) models to underwater SAS ATR using a three-stage parameter-efficient framework.
    • Stage 1 uses Low-Rank Adaptation (LoRA) while freezing the ViT backbone, bridging the gap between natural-image pretraining and underwater acoustic propagation.
  • EN 要点:
    • arXiv:2609.21061v1 Announce Type: new
    • Abstract: Automatic target recognition (ATR) with synthetic aperture sonar (SAS) supports advanced naval capabilities, but deep learning is constrained by scarc…
    • We adapt DINOv3 Vision Transformer (ViT) models to underwater SAS ATR using a three-stage parameter-efficient framework
    • Stage 1 uses Low-Rank Adaptation (LoRA) while freezing the ViT backbone, bridging the gap between natural-image pretraining and underwater acoustic propagation

• Detecting Hallucination in LLMs: Tracing the Topological Signatures of Impaired Context Sharing

  • 发布时间:2026-09-21 12:00 北京时间
  • 摘要:【待翻译】- arXiv:2609.21096v1 Announce Type: new.
    • Abstract: In this work, we examine the topology of information flow patterns within attention graphs to effectively distinguish hallucinated from non-hallucinated responses.
    • We analyze the Forman-Ricci curvature to identify structural patterns indicating information bottlenecks in attention graphs.
    • We then introduce a method that captures both semi-local and global information-flow characteristics of attention heads associated with hallucinated responses.
  • EN 要点:
    • arXiv:2609.21096v1 Announce Type: new
    • Abstract: In this work, we examine the topology of information flow patterns within attention graphs to effectively distinguish hallucinated from non-hallucinat…
    • We analyze the Forman-Ricci curvature to identify structural patterns indicating information bottlenecks in attention graphs
    • We then introduce a method that captures both semi-local and global information-flow characteristics of attention heads associated with hallucinated responses

• Decoupling Internal Representational Changes and Causal Importance in Fine-Tuned Large Language Models

  • 发布时间:2026-09-21 12:00 北京时间
  • 摘要:【待翻译】- arXiv:2609.21113v1 Announce Type: new.
    • Abstract: Fine-tuning has emerged as a widely adopted approach for adapting LLMs to a variety of downstream tasks.
    • However, how it reshapes their internal mechanisms remains poorly understood.
    • To address this, we investigate how fine-tuning alters internal representations in LLMs, including attention patterns and layer-wise activations, and examine whether these changes are linked to task-relevant components identified by EAP (e.g., attention heads and logit-level activations) that drive task performance.
  • EN 要点:
    • arXiv:2609.21113v1 Announce Type: new
    • Abstract: Fine-tuning has emerged as a widely adopted approach for adapting LLMs to a variety of downstream tasks
    • However, how it reshapes their internal mechanisms remains poorly understood
    • To address this, we investigate how fine-tuning alters internal representations in LLMs, including attention patterns and layer-wise activations, and examine wh…

• TinyCeNN-LM: Quality-Gated Conversion of Pretrained Attention with CeNN-Inspired Cellular-Recurrent Layers

  • 发布时间:2026-09-21 12:00 北京时间
  • 摘要:【待翻译】- arXiv:2609.21139v1 Announce Type: new.
    • Abstract: Replacing attention in a pretrained language model is a compatibility problem: a plausible substitute may alter representations expected by later layers.
    • TinyCeNN-LM introduces a \emph{quality-gated post-training conversion} framework using CeNN-inspired cellular-recurrent layers with bounded local processing, compact recurrent memory, routing, fusion, and accept-or-rollback validation.
    • Three implementations are studied: Integrated Memory, MemoryFusion, and PDelta3-GDN2-CLVR+Local32.
  • EN 要点:
    • arXiv:2609.21139v1 Announce Type: new
    • Abstract: Replacing attention in a pretrained language model is a compatibility problem: a plausible substitute may alter representations expected by later laye…
    • TinyCeNN-LM introduces a \emph{quality-gated post-training conversion} framework using CeNN-inspired cellular-recurrent layers with bounded local processing, co…
    • Three implementations are studied: Integrated Memory, MemoryFusion, and PDelta3-GDN2-CLVR+Local32

• Clinician-Grounded Quality Assurance for AI-Assisted Psychiatric Intake

  • 发布时间:2026-09-21 12:00 北京时间
  • 摘要:【待翻译】- arXiv:2609.21149v1 Announce Type: new.
    • Abstract: Before patients can use AI-assisted psychiatric intake systems, health systems need practical ways to routinely evaluate these tools against their clinical standards for quality assurance.
    • Because clinicians may use different intake styles, evaluation for this task must (1) support comparison across interviewing approaches, (2) minimize clinician burden, and (3) measure clinically relevant performance for health systems deploying these technologies.
    • We present a clinician-grounded evaluation platform built around a memory-augmented patient simulator for open-ended AI interviewing, InterviewPlayground.
  • EN 要点:
    • arXiv:2609.21149v1 Announce Type: new
    • Abstract: Before patients can use AI-assisted psychiatric intake systems, health systems need practical ways to routinely evaluate these tools against their cli…
    • Because clinicians may use different intake styles, evaluation for this task must (1) support comparison across interviewing approaches, (2) minimize clinician…
    • We present a clinician-grounded evaluation platform built around a memory-augmented patient simulator for open-ended AI interviewing, InterviewPlayground

• Can Agents Design Better Chips with a Higher Level Abstraction?

  • 发布时间:2026-09-21 12:00 北京时间
  • 摘要:【待翻译】- arXiv:2609.21157v1 Announce Type: new.
    • Abstract: Large Language Model (LLM) agents are increasingly being explored for chip design, but most existing approaches operate directly at RTL.
    • We ask whether agents can design better chips by leveraging higher-level abstractions.
    • We compare Direct RTL Design, Agent-based HLS Design, Post-Compiler HLS Refinement, and Post-HLS RTL Refinement, and combine Agent-based HLS Design with Post-HLS RTL Refinement as Agent-based HLS with RTL Refinement (AHRR).
  • EN 要点:
    • arXiv:2609.21157v1 Announce Type: new
    • Abstract: Large Language Model (LLM) agents are increasingly being explored for chip design, but most existing approaches operate directly at RTL
    • We ask whether agents can design better chips by leveraging higher-level abstractions
    • We compare Direct RTL Design, Agent-based HLS Design, Post-Compiler HLS Refinement, and Post-HLS RTL Refinement, and combine Agent-based HLS Design with Post-HL…

• SpecOpt: Contact-Diff Reasoning for Agentic Molecule Optimization Toward Binding Specificity

  • 发布时间:2026-09-21 12:00 北京时间
  • 摘要:【待翻译】- arXiv:2609.21165v1 Announce Type: new.
    • Abstract: Off-target protein binding is a major source of adverse effects for small-molecule drugs, yet most structure-based molecular design methods focus on generating selective compounds de novo rather than improving the selectivity of existing, well- characterized drugs.
    • We introduce specificity optimization (SpecOpt), a molecular design task that seeks constrained structural modifications to an existing compound that increase its binding preference for an intended target over known off-targets while preserving its structural identity and drug-like properties.
    • To enable systematic evaluation, we construct a ChEMBL-derived benchmark from compound-target interaction data, identifying intended targets through curated drug-mechanism annotations and off- targets through measured activities.
  • EN 要点:
    • arXiv:2609.21165v1 Announce Type: new
    • Abstract: Off-target protein binding is a major source of adverse effects for small-molecule drugs, yet most structure-based molecular design methods focus on g…
    • We introduce specificity optimization (SpecOpt), a molecular design task that seeks constrained structural modifications to an existing compound that increase i…
    • To enable systematic evaluation, we construct a ChEMBL-derived benchmark from compound-target interaction data, identifying intended targets through curated dru…

ArXiv cs.CL (B_intro+search) 链接到标题

• Do small language models know what they don’t know?

  • 发布时间:2026-09-21 12:00 北京时间
  • 摘要:【待翻译】- arXiv:2609.20824v1 Announce Type: new.
    • Abstract: We explore whether entropy-based confidence signals can be leveraged to improve the accuracy of Small Language Models (SLMs) with fewer than 3 billion parameters, running entirely on consumer hardware.
    • We evaluate seven distinct approaches, including token-level entropy early stopping, semantic entropy estimation, and uncertainty-aware routing to larger expert models, across 7 model pairs and 5 standard NLU benchmarks.
    • Our key finding is that token-level entropy is effectively blind in SLMs: in 91% of dataset-model combinations, mean token entropy is near zero regardless of answer correctness, rendering token-based confidence signals unusable at this scale.
  • EN 要点:
    • arXiv:2609.20824v1 Announce Type: new
    • Abstract: We explore whether entropy-based confidence signals can be leveraged to improve the accuracy of Small Language Models (SLMs) with fewer than 3 billion…
    • We evaluate seven distinct approaches, including token-level entropy early stopping, semantic entropy estimation, and uncertainty-aware routing to larger expert…
    • Our key finding is that token-level entropy is effectively blind in SLMs: in 91% of dataset-model combinations, mean token entropy is near zero regardless of an…

• HERMES: Contrast-Aware Knowledge Graph Reasoning from Clinical Notes for Patient Outcome Prediction

  • 发布时间:2026-09-21 12:00 北京时间
  • 摘要:【待翻译】- arXiv:2609.20825v1 Announce Type: new.
    • Abstract: Clinical predictive models often rely on structured Electronic Health Record data, such as time-series and procedure codes.
    • While recent approaches have begun leveraging unstructured clinical notes, they typically encode them as flat sequences, which may lose explicit relational and temporal structure present in clinical narratives.
    • In response, we propose HERMES, a graph-based framework that operates exclusively on clinical text while preserving clinical relationships.
  • EN 要点:
    • arXiv:2609.20825v1 Announce Type: new
    • Abstract: Clinical predictive models often rely on structured Electronic Health Record data, such as time-series and procedure codes
    • While recent approaches have begun leveraging unstructured clinical notes, they typically encode them as flat sequences, which may lose explicit relational and…
    • In response, we propose HERMES, a graph-based framework that operates exclusively on clinical text while preserving clinical relationships

• TALON: A Temporally Aware Longitudinal Framework for Radiology Report Generation

  • 发布时间:2026-09-21 12:00 北京时间
  • 摘要:【待翻译】- arXiv:2609.20826v1 Announce Type: new.
    • Abstract: Current radiology report generation (RRG) models usually produce descriptive reports based on a single examination or only the most recent prior examination, limiting their ability to perform accurate and meaningful longitudinal comparisons and detect subtle interval changes.
    • Although recent approaches have begun to incorporate multiple prior examinations, they usually aggregate a fixed-length history without explicitly modeling the role-dependent relevance of each prior examination before fusion.
    • To address this, we propose TALON, a Temporally Aware LONgitudinal RRG framework that adaptively integrates variable-length patient histories.
  • EN 要点:
    • arXiv:2609.20826v1 Announce Type: new
    • Abstract: Current radiology report generation (RRG) models usually produce descriptive reports based on a single examination or only the most recent prior exami…
    • Although recent approaches have begun to incorporate multiple prior examinations, they usually aggregate a fixed-length history without explicitly modeling the…
    • To address this, we propose TALON, a Temporally Aware LONgitudinal RRG framework that adaptively integrates variable-length patient histories

• From Discharge Notes to Patient Understanding: Persona-Grounded, Open-Ended Simulation of LLMs as Discharge Educators

  • 发布时间:2026-09-21 12:00 北京时间
  • 摘要:【待翻译】- arXiv:2609.20827v1 Announce Type: new.
    • Abstract: Hospital discharge education is an interactive teaching task: a clinician adapts a discharge plan to a patient’s literacy, recall, and personality.
    • Existing LLM evaluations target static or artifact-generation tasks and do not measure patient understanding under open-ended dialogue.
    • We introduce DischargeBench, a persona-grounded simulation in which a candidate LLM educator conducts a multi-turn session with a Virtual Patient, while an Education Monitor Agent regulates patient realism without modifying the educator, protecting the evaluation signal.
  • EN 要点:
    • arXiv:2609.20827v1 Announce Type: new
    • Abstract: Hospital discharge education is an interactive teaching task: a clinician adapts a discharge plan to a patient’s literacy, recall, and personality
    • Existing LLM evaluations target static or artifact-generation tasks and do not measure patient understanding under open-ended dialogue
    • We introduce DischargeBench, a persona-grounded simulation in which a candidate LLM educator conducts a multi-turn session with a Virtual Patient, while an Educ…

• Beyond WER: Entity and Disfluency Recall in Accented Conversational ASR

  • 发布时间:2026-09-21 12:00 北京时间
  • 摘要:【待翻译】- arXiv:2609.20828v1 Announce Type: new.
    • Abstract: ASR systems optimised for Word Error Rate (WER) often miss named entities and filled pauses in accented conversational English, both critical for language-learning feedback.
    • We present a three-stage pipeline for speakers from India, Indonesia, and Latin America: (1) heuristic SQL filters curating entity-rich training data at 2.8x the entity density of random sampling, (2) regional LoRA adapters fine-tuned on Qwen2.5-Omni-3B producing both verbatim and corrected transcripts in a single forward pass, and (3) a six-category error taxonomy validated by an LLM-based judge (83.8% agreement, 210 human-labelled samples).
    • The pipeline achieves 80-85% entity recall (up from 53-55%), 76-86% filler recall (up from <5%), and 6-10% WER across 6k test utterances, outperforming Whisper and a commercial ASR on entity recall while matching a zero-shot 30B model with 10x fewer parameters.
  • EN 要点:
    • arXiv:2609.20828v1 Announce Type: new
    • Abstract: ASR systems optimised for Word Error Rate (WER) often miss named entities and filled pauses in accented conversational English, both critical for lang…
    • We present a three-stage pipeline for speakers from India, Indonesia, and Latin America: (1) heuristic SQL filters curating entity-rich training data at 2.8x th…
    • The pipeline achieves 80-85% entity recall (up from 53-55%), 76-86% filler recall (up from <5%), and 6-10% WER across 6k test utterances, outperforming Whisper…

• SAGE: Schema-Guided LLMs for Grant Review

  • 发布时间:2026-09-21 12:00 北京时间
  • 摘要:【待翻译】- arXiv:2609.20829v1 Announce Type: new.
    • Abstract: Grant reviewers must apply detailed criteria to application forms, budgets, and supporting documents while producing assessments that colleagues can inspect.
    • We present SAGE, Schema-Guided Aspect-Based Grant Evaluation, a system that translates a grant rubric into structured checks and links its judgements to evidence from the application package.
    • We evaluate SAGE in two stages on 35 nonprofit grant applications.
  • EN 要点:
    • arXiv:2609.20829v1 Announce Type: new
    • Abstract: Grant reviewers must apply detailed criteria to application forms, budgets, and supporting documents while producing assessments that colleagues can i…
    • We present SAGE, Schema-Guided Aspect-Based Grant Evaluation, a system that translates a grant rubric into structured checks and links its judgements to evidenc…
    • We evaluate SAGE in two stages on 35 nonprofit grant applications

• Reviser: Revision-Capable Text Generation via Autoregressive Cursor Actions

  • 发布时间:2026-09-21 12:00 北京时间
  • 摘要:【待翻译】- arXiv:2609.20830v1 Announce Type: new.
    • Abstract: Revision-capable generation is appealing because it can insert or revise earlier content, but many non-autoregressive and edit-based approaches obtain this flexibility through repeated sequence-level computation.
    • We propose Reviser, a decoder-only Transformer that generates a response as a sequence of cursor-relative actions on a mutable canvas.
    • At each step, Reviser predicts exactly one action token: INSERT(token), MOVE($\Delta$), or STOP, and is autoregressive over edit-history actions rather than final text order.
  • EN 要点:
    • arXiv:2609.20830v1 Announce Type: new
    • Abstract: Revision-capable generation is appealing because it can insert or revise earlier content, but many non-autoregressive and edit-based approaches obtain…
    • We propose Reviser, a decoder-only Transformer that generates a response as a sequence of cursor-relative actions on a mutable canvas
    • At each step, Reviser predicts exactly one action token: INSERT(token), MOVE($\Delta$), or STOP, and is autoregressive over edit-history actions rather than fin…

• Recursive Language Models Generalize Out of Domain

  • 发布时间:2026-09-21 12:00 北京时间
  • 摘要:【待翻译】- arXiv:2609.20831v1 Announce Type: new.
    • Abstract: We study when limiting what a language model can see improves learning.
    • We compare standard CoT, the more general learner that reads the full trace, with recursive language models, which restricts itself by solving each subtask in an isolated context.
    • In-distribution, this generality comes for free: CoT can efficiently simulate the recursive rule, so the IID generalization guarantee changes only by a constant factor, and recursion does not offer much.
  • EN 要点:
    • arXiv:2609.20831v1 Announce Type: new
    • Abstract: We study when limiting what a language model can see improves learning
    • We compare standard CoT, the more general learner that reads the full trace, with recursive language models, which restricts itself by solving each subtask in a…
    • In-distribution, this generality comes for free: CoT can efficiently simulate the recursive rule, so the IID generalization guarantee changes only by a constant…

• TatBLiMP: A Benchmark of Linguistic Minimal Pairs for Tatar

  • 发布时间:2026-09-21 12:00 北京时间
  • 摘要:【待翻译】- arXiv:2609.20832v1 Announce Type: new.
    • Abstract: We introduce TatBLiMP, the first benchmark of linguistic minimal pairs for Tatar (tt, ISO 639-3 tat), a Qypchaq Turkic language written in Cyrillic.
    • To our knowledge it is the first grammaticality evaluation for Tatar language models of any kind, since even the 101-language MultiBLiMP does not include Tatar.
    • TatBLiMP covers 16 morphosyntactic phenomena in 1248 sentence pairs.
  • EN 要点:
    • arXiv:2609.20832v1 Announce Type: new
    • Abstract: We introduce TatBLiMP, the first benchmark of linguistic minimal pairs for Tatar (tt, ISO 639-3 tat), a Qypchaq Turkic language written in Cyrillic
    • To our knowledge it is the first grammaticality evaluation for Tatar language models of any kind, since even the 101-language MultiBLiMP does not include Tatar
    • TatBLiMP covers 16 morphosyntactic phenomena in 1248 sentence pairs

• Transsion’s Speaker-Attributed Multilingual ASR System for the MLC-SLM 2026 Challenge

  • 发布时间:2026-09-21 12:00 北京时间
  • 摘要:【待翻译】- arXiv:2609.20833v1 Announce Type: new.
    • Abstract: This paper presents the Transsion Speech Team submission to Task 1 of the MLC-SLM 2026 Challenge, which focuses on speaker-attributed transcription for multilingual conversational speech.
    • We propose a cascaded framework consisting of three components: a speaker diarization module, a long-form multilingual ASR module, and a speaker-transcription fusion module.
    • The diarization module is built upon DiariZen and produces speaker-homogeneous segments through local speaker activity estimation and global speaker clustering.
  • EN 要点:
    • arXiv:2609.20833v1 Announce Type: new
    • Abstract: This paper presents the Transsion Speech Team submission to Task 1 of the MLC-SLM 2026 Challenge, which focuses on speaker-attributed transcription fo…
    • We propose a cascaded framework consisting of three components: a speaker diarization module, a long-form multilingual ASR module, and a speaker-transcription f…
    • The diarization module is built upon DiariZen and produces speaker-homogeneous segments through local speaker activity estimation and global speaker clustering

ArXiv cs.LG (B_intro+search) 链接到标题

• Sparse Priors for Efficient Distribution Learning

  • 发布时间:2026-09-21 12:00 北京时间
  • 摘要:【待翻译】- arXiv:2609.20883v1 Announce Type: new.
    • Abstract: Despite the widespread use and success of generative AI techniques today, theoretical guarantees on learning a distribution supported in $d$ dimensions from $n$ samples degrade as $O(n^{-1/\Theta(d)})$, though shown to be minimax optimal.
    • We hypothesize that present bounds are too pessimistic because smoothness assumptions are not enough to capture the structure of distributions that often appear in real applications.
    • Consequently, we introduce the class of sparse priors and define the “Sparse Dimension” as a measure of sparsity of a prior over the space of all distributions.
  • EN 要点:
    • arXiv:2609.20883v1 Announce Type: new
    • Abstract: Despite the widespread use and success of generative AI techniques today, theoretical guarantees on learning a distribution supported in $d$ dimension…
    • We hypothesize that present bounds are too pessimistic because smoothness assumptions are not enough to capture the structure of distributions that often appear…
    • Consequently, we introduce the class of sparse priors and define the “Sparse Dimension” as a measure of sparsity of a prior over the space of all distributions

• BI-Agent and BI-Bench: Towards Automating End-to-End Business Intelligence

  • 发布时间:2026-09-21 12:00 北京时间
  • 摘要:【待翻译】- arXiv:2609.20886v1 Announce Type: new.
    • Abstract: Business intelligence (BI) is a cornerstone of enterprise decision-making and is widely used by enterprise users in software such as Power BI and Tableau.
    • In traditional BI workflows, users need to prepare data by (1) identifying relevant tables, (2) performing data transformations, and (3) building join relationships, before they can (4) answer their business questions.
    • These steps can be complex and time-consuming, making BI challenging.
  • EN 要点:
    • arXiv:2609.20886v1 Announce Type: new
    • Abstract: Business intelligence (BI) is a cornerstone of enterprise decision-making and is widely used by enterprise users in software such as Power BI and Tabl…
    • In traditional BI workflows, users need to prepare data by (1) identifying relevant tables, (2) performing data transformations, and (3) building join relations…
    • These steps can be complex and time-consuming, making BI challenging

• Elastic Threshold Attention: Learned Contextual Sparsity for Long-Context Decoding

  • 发布时间:2026-09-21 12:00 北京时间
  • 摘要:【待翻译】- arXiv:2609.20888v1 Announce Type: new.
    • Abstract: Massive KV caches can cause severe memory-bandwidth bottlenecks during long-context decoding.
    • Sparse attention methods mitigate this via selective loading, but that comes at a cost: rigid heuristics drop necessary context, leading to quality degradation.
    • We introduce \textbf{Elastic Threshold Attention (ETA)}, an end-to-end trainable architecture that achieves hardware-accelerated decoding speed without sacrificing dense model quality.
  • EN 要点:
    • arXiv:2609.20888v1 Announce Type: new
    • Abstract: Massive KV caches can cause severe memory-bandwidth bottlenecks during long-context decoding
    • Sparse attention methods mitigate this via selective loading, but that comes at a cost: rigid heuristics drop necessary context, leading to quality degradation
    • We introduce \textbf{Elastic Threshold Attention (ETA)}, an end-to-end trainable architecture that achieves hardware-accelerated decoding speed without sacrific…

• Bio-MF: Low-Latency and High-Fidelity EEG-to-fNIRS Cross-Modal Generation for Hybrid Motor-Imagery Brain–Computer Interfaces

  • 发布时间:2026-09-21 12:00 北京时间
  • 摘要:【待翻译】- arXiv:2609.20904v1 Announce Type: new.
    • Abstract: Hybrid motor-imagery brain-computer interfaces (MI-BCIs) combining EEG and fNIRS can outperform EEG-only systems by exploiting complementary electrophysiological and hemodynamic information.
    • To obtain such hybrid information when paired EEG-fNIRS acquisition is unavailable or inconvenient, recent studies have focused on EEG-to-fNIRS cross-modal generation.
    • However, existing methods still suffer from slow generation and often require pretraining, limiting their use in real-time MI-BCI scenarios.
  • EN 要点:
    • arXiv:2609.20904v1 Announce Type: new
    • Abstract: Hybrid motor-imagery brain-computer interfaces (MI-BCIs) combining EEG and fNIRS can outperform EEG-only systems by exploiting complementary electroph…
    • To obtain such hybrid information when paired EEG-fNIRS acquisition is unavailable or inconvenient, recent studies have focused on EEG-to-fNIRS cross-modal gene…
    • However, existing methods still suffer from slow generation and often require pretraining, limiting their use in real-time MI-BCI scenarios

• Continuous Delayed-Memory Stochastic Gradient Descent and Continuous-Time Reinforcement Learning from History of Astrophysical Time Series Studies

  • 发布时间:2026-09-21 12:00 北京时间
  • 摘要:【待翻译】- arXiv:2609.20906v1 Announce Type: new.
    • Abstract: Quasars are luminous objects in the universe that exhibit stochastic brightness variations encoding information about the supermassive black holes powering them, and modeling these variations from ground-based survey data time series, known as light curves, is a statistical challenge.
    • This paper reviews how stochastic differential equations (SDEs) have been adapted with neural network parameterizations to overcome this challenge in history.
    • We create the Continuous-Delayed-Memory Stochastic Gradient Descent which depend on the past state of the discrete iteration process.
  • EN 要点:
    • arXiv:2609.20906v1 Announce Type: new
    • Abstract: Quasars are luminous objects in the universe that exhibit stochastic brightness variations encoding information about the supermassive black holes pow…
    • This paper reviews how stochastic differential equations (SDEs) have been adapted with neural network parameterizations to overcome this challenge in history
    • We create the Continuous-Delayed-Memory Stochastic Gradient Descent which depend on the past state of the discrete iteration process

• Do Quantum Models Scale Like LLMs?

  • 发布时间:2026-09-21 12:00 北京时间
  • 摘要:【待翻译】- arXiv:2609.20912v1 Announce Type: new.
    • Abstract: In this work, we study the neural scaling laws of RydbergGPT, an autoregressive transformer model trained on qubit projective measurement data gathered from interacting Rydberg atom arrays.
    • The quantum system is known to exhibit a finite-size remnant of a critical point as the laser detuning parameter is varied.
    • We find that near the critical point the transformer loss as a function of training dataset size is well described by a power-law with a loss floor correction.
  • EN 要点:
    • arXiv:2609.20912v1 Announce Type: new
    • Abstract: In this work, we study the neural scaling laws of RydbergGPT, an autoregressive transformer model trained on qubit projective measurement data gathere…
    • The quantum system is known to exhibit a finite-size remnant of a critical point as the laser detuning parameter is varied
    • We find that near the critical point the transformer loss as a function of training dataset size is well described by a power-law with a loss floor correction

• When AI Reviews Train AI Reviewers: Scientific-Judgment Collapse and Mitigation

  • 发布时间:2026-09-21 12:00 北京时间
  • 摘要:【待翻译】- arXiv:2609.20942v1 Announce Type: new.
    • Abstract: Large language models (LLMs) increasingly participate in scientific evaluation, both as automated reviewers and as assistants to human reviewers.
    • As model-generated reviews enter public data and future training corpora, AI peer review can become recursive: later reviewers learn from judgments produced by earlier models.
    • We study one step of this feedback loop in a controlled setting.
  • EN 要点:
    • arXiv:2609.20942v1 Announce Type: new
    • Abstract: Large language models (LLMs) increasingly participate in scientific evaluation, both as automated reviewers and as assistants to human reviewers
    • As model-generated reviews enter public data and future training corpora, AI peer review can become recursive: later reviewers learn from judgments produced by…
    • We study one step of this feedback loop in a controlled setting

• Efficient Bayes-Adaptive Reinforcement Learning with Temporal Logic Specifications

  • 发布时间:2026-09-21 12:00 北京时间
  • 摘要:【待翻译】- arXiv:2609.20954v1 Announce Type: new.
    • Abstract: We present a novel end-to-end model-based Reinforcement Learning (RL) algorithm for efficient policy synthesis under given Linear Temporal Logic (LTL) specifications (e.g., safety or reachability) in unknown environments.
    • To do so, a Limit-Deterministic B{"u}chi Automaton (LDBA) representation of the LTL task is synchronised with a Bayes-Adaptive Markov Decision Process (BAMDP) representation of the environment, which allows us to leverage an enhanced exploration-exploitation trade-off that is achieved via Bayesian RL, as opposed to traditional non-Bayesian approaches.
    • We further propose a novel Bayes-Adaptive Monte-Carlo Planning (BAMCP) algorithm to allow for approximate Bayes-optimal strategy synthesis in the synchronised BAMDP construct.
  • EN 要点:
    • arXiv:2609.20954v1 Announce Type: new
    • Abstract: We present a novel end-to-end model-based Reinforcement Learning (RL) algorithm for efficient policy synthesis under given Linear Temporal Logic (LTL)…
    • To do so, a Limit-Deterministic B{"u}chi Automaton (LDBA) representation of the LTL task is synchronised with a Bayes-Adaptive Markov Decision Process (BAMDP)…
    • We further propose a novel Bayes-Adaptive Monte-Carlo Planning (BAMCP) algorithm to allow for approximate Bayes-optimal strategy synthesis in the synchronised B…

• From Switching to Dynamic Regret: A Simple Reduction via Unbiased Random Sequences

  • 发布时间:2026-09-21 12:00 北京时间
  • 摘要:【待翻译】- arXiv:2609.20968v1 Announce Type: new.
    • Abstract: In non-stationary online learning, dynamic regret has attracted increasing attention as a measure of how well an online learner performs against a time-varying comparator sequence.
    • Despite considerable advances, attaining optimal bounds for strongly convex and exp-concave losses often involves intricate analysis.
    • In this paper, we present a \textit{simple} framework that reduces dynamic regret minimization to switching regret minimization.
  • EN 要点:
    • arXiv:2609.20968v1 Announce Type: new
    • Abstract: In non-stationary online learning, dynamic regret has attracted increasing attention as a measure of how well an online learner performs against a tim…
    • Despite considerable advances, attaining optimal bounds for strongly convex and exp-concave losses often involves intricate analysis
    • In this paper, we present a \textit{simple} framework that reduces dynamic regret minimization to switching regret minimization

• Generative inversion for early ranking of competing geologic interpretations

  • 发布时间:2026-09-21 12:00 北京时间
  • 摘要:【待翻译】- arXiv:2609.20978v1 Announce Type: new.
    • Abstract: High-consequence subsurface decisions are often made under severe data scarcity.
    • Experts may arrive at competing interpretations of the same subsurface system, yet early in a project there is rarely a practical way to determine which one is most realistic.
    • This uncertainty can persist until several wells are drilled, often costing millions of dollars.
  • EN 要点:
    • arXiv:2609.20978v1 Announce Type: new
    • Abstract: High-consequence subsurface decisions are often made under severe data scarcity
    • Experts may arrive at competing interpretations of the same subsurface system, yet early in a project there is rarely a practical way to determine which one is…
    • This uncertainty can persist until several wells are drilled, often costing millions of dollars