System translated (Gemini)

🤖 AI 速览

Today’s main theme is the continued descent of AI from general assistants into specialized services and real infrastructure. OpenAI launched solutions for legal scenarios, while Anthropic and Notion strengthened project-level and team-level workflows; meanwhile, data centers, consumer-side …
📋 文章元数据
发布时间
2026-09-18
类型
ai-daily
字数
7138
阅读时长
34 min

2026-09-18 AI Daily | OpenAI Enters the Legal Sector: AI Adoption Now Hinges on Specialized Data, Evidence Chains, and Compute Costs Link to heading

The main theme today is AI’s ongoing shift from general-purpose assistants to specialized services and tangible infrastructure. OpenAI is launching solutions for the legal sector, while Anthropic and Notion are enhancing project and team-level workflows. At the same time, developments in data centers, consumer access points, and several papers on trustworthy systems suggest that the next stage of competition will focus not just on model capabilities, but also on evidence traceability, access control, long-term memory, and sustainable costs.

📖 In-depth Guide to This Issue’s Watch List Link to heading

Three threads are most worthy of attention for today’s Watch List. First, “AI Infrastructure and Mainstream User Adoption”: Meta’s narrative on data centers and employment, along with Joanna Stern’s discussion on the iPhone and AI for the masses, provide supply-side and consumer-side perspectives on the next phase of AI’s expansion. Second, several arXiv papers point to “Trustworthy AI Systems,” covering topics like independently challengeable AI-assisted claims, evidence tracing with GraphEcho, document version control, and sensor digital twins. These are a must-read for teams focused on enterprise-level Agent risk management. Third, “Data and Memory Capabilities”: NeMo Data Designer, multimodal synthetic data, CapMem for egocentric video memory, and papers on agents in healthcare and trading scenarios demonstrate that models are shifting from pure generation to long-term memory, verifiable decision-making, and specialized execution.

🌐 AI Hotspots on X Link to heading

Topic 1: Anthropic Launches Redesigned Projects for Claude Code Link to heading

  • Category: AI · News
  • Overview: Trending time: 6 hours ago, Related posts: 3,900
  • What it is: Anthropic has launched a redesigned Projects feature for Claude Code, aimed at better organizing code context, tasks, and collaboration workflows.
  • Why it matters: This indicates a shift for AI programming assistants from single-turn conversational tools to project-level workspaces, enhancing the model’s utility in long-term code comprehension, context management, and the software development lifecycle.
  • Discussion summary: Discussions on X are focused on whether the new Projects feature will significantly improve the developer experience in Claude Code, its competitiveness against tools like Cursor and GitHub Copilot, and the reliability and cost of project-level context management in real-world engineering.

Topic 2: Anthropic Unifies Claude Chat and Cowork into Single Seamless Experience Link to heading

  • Category: AI · News
  • Overview: Trending time: 1 day ago, Related posts: 11,000
  • What it is: Anthropic has integrated Claude’s chat features with collaborative work scenarios to create a more seamless, unified experience.
  • Why it matters: This shows that AI assistants are evolving from simple conversational tools into collaborative productivity platforms deeply integrated into workflows, which could influence the direction of enterprise AI applications and agent-based products.
  • Discussion summary: Discussions on X center on whether the new experience can actually boost office productivity, Claude’s differentiating advantages over ChatGPT and Gemini, and the potential data privacy, access control, and enterprise adoption barriers that come with deeper workflow integration.

Topic 3: Z.ai’s AI Agent Triples Throughput on Domestic Hardware in Two Weeks Link to heading

  • Category: AI · News
  • Overview: Trending time: 15 hours ago, Related posts: 2,900
  • What it is: Z.ai’s AI Agent has reportedly tripled its throughput on domestic hardware within two weeks through optimization.
  • Why it matters: This suggests that model and system optimization can significantly unlock the performance of domestic AI hardware, helping to reduce reliance on foreign hardware and boost the competitiveness of the local compute ecosystem.
  • Discussion summary: Discussions on X are focused on the testing methodology behind the threefold increase, the specific contributions of hardware and software optimizations, whether the results can be replicated across different tasks and scales, and the practical implications of this achievement for the performance and commercial viability of domestic chips.

Topic 4: Ex-OpenAI Engineer Launches Jev for Fast AI Decisions Link to heading

  • Category: AI · News
  • Overview: Trending time: 2 days ago, Related posts: 72,000
  • What it is: A former OpenAI engineer has launched a new project called Jev, which focuses on enabling AI systems to make decisions faster.
  • Why it matters: If Jev can significantly reduce AI decision latency, it could help accelerate the adoption of agents, multi-step automation, and real-time applications, enhancing AI’s utility in finance, customer service, programming, and enterprise processes.
  • Discussion summary: Discussions on X are centered on whether Jev truly offers a speed advantage far superior to existing solutions, the transparency of its technical details, whether it can strike a balance between accuracy and safety, and if the trend of ex-OpenAI talent founding startups will further intensify competition in the AI agent space.

Topic 5: Notion Launches Skills API for Team AI Prompts Link to heading

  • Category: AI · News
  • Overview: Trending Time: 6 hours ago, Related Posts: 377
  • What it is: Notion has launched the Skills API, allowing teams to integrate reusable AI prompts and workflow capabilities into Notion AI.
  • Why it matters: This indicates that AI products are moving from single-turn conversations to reusable, manageable team-level workflows, helping enterprises standardize prompt engineering, knowledge bases, and business processes.
  • Discussion Summary: Discussions on X focus on whether the Skills API will make enterprise AI workflows easier to implement. Some are optimistic that Notion can enhance team collaboration and automation capabilities, while others are concerned about prompt governance, permission security, result reliability, and dependency on a specific platform.

Topic 6: Dreamforce Spotlights AI Integrations and Safety Debates with Top Executives Link to heading

  • Category: AI · News
  • Overview: Trending Time: 2 days ago, Related Posts: 23,000
  • What it is: The Salesforce Dreamforce conference focused on the implementation of enterprise AI integration and featured several executives discussing AI safety and governance issues.
  • Why it matters: This reflects that AI is shifting from showcasing model capabilities to large-scale application in enterprise software, business processes, and security compliance. The integration methods of related platforms may influence the speed and standards of enterprise AI adoption.
  • Discussion Summary: Discussions on X mainly center on whether enterprise AI already has a practical ROI, the implementation boundaries for AI agents and office automation, and how large tech companies should shoulder security, privacy, and governance responsibilities while promoting AI commercialization.

AI Public Opinion Summary on X Today Link to heading

The main theme of today’s public opinion is that AI is shifting from single-turn conversations and model capability demonstrations to deep integration with project-level context, team collaboration, reusable workflows, and enterprise business processes, with agents and AI programming tools becoming key areas of competition. The general consensus is that workflow integration, synergistic optimization of software and hardware, and lower decision latency are expected to boost the actual productivity of AI and accelerate its enterprise adoption. However, the market remains divided on whether these products can deliver sustainable, real ROI, and whether the user experience and performance advantages of different vendors are sufficient to create differentiation. Regarding Z.ai’s “triple throughput,” Jev’s speed advantage, and the actual effectiveness of various new features, discussions generally call for more transparent testing standards and reproducible results, while also questioning whether speed improvements come at the cost of accuracy and reliability. Potential risks are concentrated in enterprise data privacy, permissions and prompt governance, loss of control over agents and liability attribution, and dependency on a single platform. If security, compliance, and long-term cost issues are not resolved, the transition of AI from demonstration to large-scale application could still be hindered.

💡 Influencer Insights Link to heading

No influencer insights today. In-depth content from the Watch List is recommended.

📚 Appendix: Today’s Watch List Update Source List Link to heading

Time window: Last 3 days; 22 sources covered; 36 updates in total.

Lex Fridman Podcast (A_full) Link to heading

  • #502 – Psychiatry, Insane Asylums, Mental Illness, ECT, Lobotomies, Freud & Jung
    • Release Time: 2026-09-17 08:40 Beijing Time
    • Abstract: [Pending Translation] - Andrew Scull is a historian of psychiatry.
      • See below for timestamps, transcript, and to give feedback, submit questions, contact Lex, etc.
      • Wispr Flow: AI-powered voice dictation app.
      • Fin: AI agent for customer service.
      • LMNT: Zero-sugar electrolyte drink mix.
    • EN Key Points:
      • Andrew Scull is a historian of psychiatry
      • Thank you for listening ❤ Check out our sponsors:
      • See below for timestamps, transcript, and to give feedback, submit questions, contact Lex, etc
      • CONTACT LEX:

All-In Podcast (A_full) Link to heading

  • Meta’s Dina Powell McCormick: The Case for Data Centers, Backlash, AI Job Boom & Meta’s Future

    • Published: 2026-09-18 00:54 Beijing Time
    • Summary: - 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.
      • Keel Infrastructure owns the power, land, and connectivity that HPC and AI run on - backed by secured energy assets and established grid interconnections across North America.
    • EN Highlights:
      • (0:00) Dina Powell McCormick joins The Besties
      • (1:52) Impact of Meta’s data center in Richland Parish, LA: $50K teacher checks, revitalizing a small town community
      • (11:52) How Meta is alleviating data center concerns: pollution, electricity, water, noise, aesthetics; Misinformation campaigns
      • (19:57) America’s Workforce Academy, the job growth in re-industrializing America

  • Brad Gerstner: No AI Bubble, Semis Eat the Nasdaq & AI’s Take Off Problem

    • Published: 2026-09-17 11:04 Beijing Time
    • Summary: - 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.

  • Keel Infrastructure owns the power, land, and connectivity that HPC and AI run on - backed by secured energy assets and established grid interconnections across North America.

    • EN Highlights:
      • (0:00) Welcome Brad Gerstner
      • (1:01) Trump Accounts, Every Child a Capitalist & The CAC Scan
      • (5:07) Can AI revenue pay for the CapEx
      • (8:53) The Build Out Issue: Gigawatts, TAM, Token Growth, and Margin Expansion

Stratechery by Ben Thompson (A_full) Link to heading

  • An Interview with Joanna Stern About the iPhone Duo and AI for Normal People
    • Published: 2026-09-17 18:00 Beijing Time
    • Summary: [Translation Pending] - $15 / month or $150 / year.
      • Substantial analysis of the news of the day delivered via three weekly emails or podcasts.
      • Stratechery Interviews.
      • Interviews with leading public CEOs, private company founders, and discussions with fellow analysts.
      • A twice-weekly podcast from John Gruber and myself: 15 minutes an episode, not a minute less, not a minute more.
    • EN Highlights:
      • An Interview with Joanna Stern About the iPhone Duo and AI for Normal People

OpenAI Blog (A_full) Link to heading

  • Introducing Astra for Law
    • Published: 2026-09-17 08:00 Beijing Time
    • Summary: [Translation Pending] - OpenAI for Law brings frontier intelligence for law, custom firm workflows, connected legal data sources, and legal-grade controls for confidential client work.
      • This piece from OpenAI Blog explains how Introducing Astra for Law shapes the broader AI and infrastructure landscape.
      • It also surfaces practical implications for founders, operators, and investors following Introducing Astra for Law.
    • EN Highlights:
      • OpenAI for Law brings frontier intelligence for law, custom firm workflows, connected legal data sources, and legal-grade controls for confidential client work.

Lex Fridman (B_intro+search) Link to heading

  • Psychiatry, Insane Asylums, Mental Illness, ECT, Lobotomies, Freud & Jung | Lex Fridman Podcast #502
  • Publication Time: 2026-09-17 08:34 Beijing Time
    • Summary: Andrew Scull is a historian of psychiatry.
      • See below for timestamps, transcript, and to give feedback, submit questions, contact Lex, etc.
      • Feedback - give feedback to Lex:.
      • AMA - submit questions, videos or call-in:.
    • EN Key Points:
      • Andrew Scull is a historian of psychiatry
      • Thank you for listening ❤ Check out our sponsors:
      • See below for timestamps, transcript, and to give feedback, submit questions, contact Lex, etc
      • Transcript:

ArXiv cs.AI (B_intro+search) Link to heading

  • Making AI-Assisted Claims Independently Challengeable: Publication Authority and a Protocol for Falsifiable Publication Records

    • Publication Time: 2026-09-17 12:00 Beijing Time
    • Summary: arXiv:2609.17631v1 Announce Type: new.
      • Abstract: AI-assisted claims can appear authoritative when evidence, analysis, human authorization, presentation, and correction history refer to different states.
      • Provenance, attestation, and transparency expose history but alone do not specify the publication transition examined here.
      • We develop Publication Authority as an exact-state, non-transferable, single-use publication capability and instantiate it in PAC-2026 (Publication-Accountability Calculus), a machine-readable AIJIM Protocol candidate.
    • EN Key Points:
      • arXiv:2609.17631v1 Announce Type: new
      • Abstract: AI-assisted claims can appear authoritative when evidence, analysis, human authorization, presentation, and correction history refer to different stat…
      • Provenance, attestation, and transparency expose history but alone do not specify the publication transition examined here
      • We develop Publication Authority as an exact-state, non-transferable, single-use publication capability and instantiate it in PAC-2026 (Publication-Accountabili…

  • EvolveTrade: Experience-Driven Policy Refinement for Self-Evolving LLM Trading Agents

    • Publication Time: 2026-09-17 12:00 Beijing Time

  • Abstract: - arXiv:2609.17632v1 Announce Type: new.

    • Abstract: Large language model (LLM) trading agents can combine market data, news, and executable analysis, but their behavior is often controlled by static hand-written tool-use policies that are fixed before deployment.
    • This limits their ability to adapt how they gather evidence, invoke tools, verify signals, and manage risk under changing market regimes.
    • We introduce EvolveTrade, a self-evolving framework that treats the system prompt of a tool-using trading agent as a text-parameterized policy.

  • EN Key Points:

    • arXiv:2609.17632v1 Announce Type: new
    • Abstract: Large language model (LLM) trading agents can combine market data, news, and executable analysis, but their behavior is often controlled by static han…
    • This limits their ability to adapt how they gather evidence, invoke tools, verify signals, and manage risk under changing market regimes
    • We introduce EvolveTrade, a self-evolving framework that treats the system prompt of a tool-using trading agent as a text-parameterized policy

  • One Color Preprocessing Improves DSATUR

    • Published: 2026-09-17 12:00 Beijing Time
    • Abstract: - arXiv:2609.17633v1 Announce Type: new.
      • Abstract: The Graph Coloring Problem (GCP) is NP-hard and DSATUR stands as one of the fastest heuristics for it despite producing colorings that typically use more colors than state-of-the-art coloring algorithms.
      • We propose SSLD (Semidefinite Spectral Learning with DSATUR), which improves DSATUR by preprocessing a first good color class before letting DSATUR complete coloring the rest of the given graph.
      • We obtain this color class from a Semidefinite Programming (SDP), similar to an SDP used to compute the Lov'asz theta number.
    • EN Key Points:
      • arXiv:2609.17633v1 Announce Type: new

  • Abstract: The Graph Coloring Problem (GCP) is NP-hard and DSATUR stands as one of the fastest heuristics for it despite producing colorings that typically use m…

    • We propose SSLD (Semidefinite Spectral Learning with DSATUR), which improves DSATUR by preprocessing a first good color class before letting DSATUR complete col…
    • We obtain this color class from a Semidefinite Programming (SDP), similar to an SDP used to compute the Lov'asz theta number

  • Physics-Constrained Digital Twins for Sensor Integrity in Urban Pedestrian Flow: Detecting Stealthy False Data Injection with Conformal Guarantees

    • Published: 2026-09-17 12:00 Beijing Time
    • Abstract: - arXiv:2609.17635v1 Announce Type: new.
      • Abstract: City pedestrian counting systems now feed economic indicators, planning decisions and safety operations, yet the twins built on top of them treat the incoming stream as ground truth.
      • We study what happens when it is not.
      • We formalise stealthy false data injection for city-scale pedestrian sensing, where the map from latent flow to observation is far more rank deficient than in the power and water networks for which stealth has been characterised.
    • EN Key Points:
      • arXiv:2609.17635v1 Announce Type: new
      • Abstract: City pedestrian counting systems now feed economic indicators, planning decisions and safety operations, yet the twins built on top of them treat the…
      • We study what happens when it is not
      • We formalise stealthy false data injection for city-scale pedestrian sensing, where the map from latent flow to observation is far more rank deficient than in t…

  • What You Can’t See Is Still What You Learn: A Preregistered Sixty-Society Confirmation That Evidence Masking Drives Compositional Generalization

    • Published: 2026-09-17 12:00 Beijing Time
    • Abstract: - arXiv:2609.17637v1 Announce Type: new.
      • Abstract: Restricting what a module can read may improve what a system learns to compute.

  • We test this in a preregistered confirmation with sixty four-cell systems sharing a frozen language-model backbone and communicating through learned continuous packets.

  • Five conditions vary evidence masking, ownership markers, and replacement of foreign evidence with neutral filler, across six initialization clusters, each with two data orders, on one fresh task world.

  • EN Key Points:

    • arXiv:2609.17637v1 Announce Type: new
    • Abstract: Restricting what a module can read may improve what a system learns to compute
    • We test this in a preregistered confirmation with sixty four-cell systems sharing a frozen language-model backbone and communicating through learned continuous…
    • Five conditions vary evidence masking, ownership markers, and replacement of foreign evidence with neutral filler, across six initialization clusters, each with…

  • CapMem: A Benchmark for Caption-Based Episodic Memory in Egocentric Video

    • Published: 2026-09-17 12:00 Beijing Time
    • Abstract: [To be translated] - arXiv:2609.17688v1 Announce Type: new.
      • Abstract: Wearable assistants require episodic memory over egocentric video, yet current vision-language models face bounded frame budgets, growing visual-token costs, and long-context retrieval failures.
      • Under these practical constraints, we study whether textual captions can serve as reusable episodic memory.
      • We define the Episodic Memory Video Caption QA task and introduce CapMem, a human-annotated benchmark with 75 videos totaling 33.7 hours, and 1,000 multiple-choice questions across 16 scenarios.
    • EN Key Points:
      • arXiv:2609.17688v1 Announce Type: new
      • Abstract: Wearable assistants require episodic memory over egocentric video, yet current vision-language models face bounded frame budgets, growing visual-token…
      • Under these practical constraints, we study whether textual captions can serve as reusable episodic memory

  • We define the Episodic Memory Video Caption QA task and introduce CapMem, a human-annotated benchmark with 75 videos totaling 33.7 hours, and 1,000 multiple-cho…

  • GraphEcho: Structural Redundancy and Evidence Provenance in LLM Graph Agents

    • Publication Time: 2026-09-17 12:00 Beijing Time
    • Abstract: [Translation Pending] - arXiv:2609.17695v1 Announce Type: new.
      • Abstract: A large language model (LLM) agent can follow more graph paths without acquiring more independent evidence.
      • GraphEcho tests whether agents mistake these repeated encounters for additional corroboration.
      • The benchmark varies path counts and evidential origins while holding evidence content fixed, and evaluates both judgments and active exploration.
    • EN Key Points:
      • arXiv:2609.17695v1 Announce Type: new
      • Abstract: A large language model (LLM) agent can follow more graph paths without acquiring more independent evidence
      • GraphEcho tests whether agents mistake these repeated encounters for additional corroboration
      • The benchmark varies path counts and evidential origins while holding evidence content fixed, and evaluates both judgments and active exploration

  • GVD: Governed Versioning and Deduplication for Document Repositories

    • Publication Time: 2026-09-17 12:00 Beijing Time
    • Abstract: [Translation Pending] - arXiv:2609.17696v1 Announce Type: new.
      • Abstract: Document repositories evolve continuously.
      • Guidelines and policies are revised, superseded, and re-uploaded, so the same content recurs in different wording and newer versions refine or contradict earlier ones.
      • These inconsistencies belong to the growing collection rather than to any single document, yet existing work treats versioning, duplicate detection, and contradiction detection as isolated pairwise tasks and stops once a pair is labeled.
    • EN Key Points:
      • arXiv:2609.17696v1 Announce Type: new
      • Abstract: Document repositories evolve continuously

  • Guidelines and policies are revised, superseded, and re-uploaded, so the same content recurs in different wording and newer versions refine or contradict earlie…

  • These inconsistencies belong to the growing collection rather than to any single document, yet existing work treats versioning, duplicate detection, and contrad…

  • NeMo Data Designer: An Extensible Framework for Multimodal Synthetic Data Generation

    • Release Time: 2026-09-17 12:00 Beijing Time
    • Summary: [Translation Pending] - arXiv:2609.17699v1 Announce Type: new.
      • Abstract: We present NeMo Data Designer (NDD), an open-source, general-purpose framework for multi-modal synthetic data generation (SDG).
      • Designed to be intuitive to use, NDD provides a declarative configuration format in which human and/or agent users define each dataset column, with column types spanning text, code, structured outputs, images, embeddings, and statistical samplers that are explicitly configured to steer dataset diversity.
      • Additional column types and functionality can be introduced using the framework’s flexible plugin system.
    • EN Key Points:
      • arXiv:2609.17699v1 Announce Type: new
      • Abstract: We present NeMo Data Designer (NDD), an open-source, general-purpose framework for multi-modal synthetic data generation (SDG)
      • Designed to be intuitive to use, NDD provides a declarative configuration format in which human and/or agent users define each dataset column, with column types…
      • Additional column types and functionality can be introduced using the framework’s flexible plugin system

  • A Systematic Evaluation of the COTQ Provincial Land Cover Product: Structural Consistency, Spectral Separability, and Relative Positioning Against ESA, ESRI, and Google Products

    • Release Time: 2026-09-17 12:00 Beijing Time
    • Summary: [Translation Pending] - arXiv:2609.17731v1 Announce Type: new.

  • Abstract: High-resolution land use and land cover (LULC) products derived from Sentinel-2 imagery are widely used for environmental monitoring and land management, yet their performance can vary across regions with complex ecological gradients and heterogeneous surface conditions.

  • In Quebec, these limitations motivated the development of a provincial 10-m land-cover product, the COTQ, designed to support annual monitoring of land occupation and soil artificialisation.

  • This study presents a systematic evaluation of the COTQ product relative to three global 10-m LULC datasets: ESA WorldCover, ESRI LandCover, and Google DynamicWorld.

    • EN Highlights:
      • arXiv:2609.17731v1 Announce Type: new
      • Abstract: High-resolution land use and land cover (LULC) products derived from Sentinel-2 imagery are widely used for environmental monitoring and land manageme…
      • In Quebec, these limitations motivated the development of a provincial 10-m land-cover product, the COTQ, designed to support annual monitoring of land occupati…
      • This study presents a systematic evaluation of the COTQ product relative to three global 10-m LULC datasets: ESA WorldCover, ESRI LandCover, and Google DynamicW…

ArXiv cs.CL (B_intro+search) Link to heading

  • Enhancing Extubation Failure Prediction with LLM-Derived Features from Respiratory Therapy Clinical Notes

    • Published: 2026-09-17 12:00 Beijing Time
    • Abstract: [Awaiting translation] - arXiv:2609.17532v1 Announce Type: new.
      • Abstract: Invasive mechanical ventilation is a lifesaving therapy, but timely, safe discontinuation is essential to preventing extubation failure (EF) and related risks to health.
      • We present a novel approach to EF prediction that leverages features classified in free-text respiratory therapy notes using a large language model and logistic regression pipeline.

  • Applied to a patient cohort from University of Washington Medicine, our method identifies clinically meaningful EF-related features that improve EF prediction performance when included alongside structured patient data.

    • EN Highlights:
      • arXiv:2609.17532v1 Announce Type: new
      • Abstract: Invasive mechanical ventilation is a lifesaving therapy, but timely, safe discontinuation is essential to preventing extubation failure (EF) and relat…
      • We present a novel approach to EF prediction that leverages features classified in free-text respiratory therapy notes using a large language model and logistic…
      • Applied to a patient cohort from University of Washington Medicine, our method identifies clinically meaningful EF-related features that improve EF prediction p…

  • Faking Good and Faking Bad in LLMs: Response Distortion Across Dark Triad Personality Traits

    • Publication Time: 2026-09-17 12:00 Beijing Time
    • Abstract: [To be translated] - arXiv:2609.17534v1 Announce Type: new.
      • Abstract: Social desirability and impression management are pervasive sources of response distortion in human personality assessment, yet their effects on Large Language Models (LLMs) remain underexplored.
      • This study investigates whether contemporary LLMs systematically modulate the expression of Dark Triad traits (Machiavellianism, narcissism, and psychopathy) under fake-good and fake-bad conditions.
      • Seven state-of-the-art models were evaluated across two ecologically relevant contexts: employment selection and forensic evaluation, in which socially desirable or undesirable incentives were conveyed through contextual framing.
    • EN Highlights:
      • arXiv:2609.17534v1 Announce Type: new
      • Abstract: Social desirability and impression management are pervasive sources of response distortion in human personality assessment, yet their effects on Large…

  • This study investigates whether contemporary LLMs systematically modulate the expression of Dark Triad traits (Machiavellianism, narcissism, and psychopathy) un…

  • Seven state-of-the-art models were evaluated across two ecologically relevant contexts: employment selection and forensic evaluation, in which socially desirabl…

  • DANTINOX: A Unified Framework for Multi-Paradigm Language Modeling

    • Publication Time: 2026-09-17 12:00 Beijing Time
    • Abstract: [TO BE TRANSLATED] - arXiv:2609.17535v1 Announce Type: new.
      • Abstract: Language generation research increasingly spans three paradigms: autoregressive decoding, discrete masked diffusion, and continuous flow-matching.
      • Comparing them is difficult because each lives in a separate codebase, so measured differences often reflect implementation details rather than the paradigms themselves.
      • We present DantinoX, an open-source JAX/Flax library in which a single modular Transformer backbone serves all three paradigms.
    • Key Points:
      • arXiv:2609.17535v1 Announce Type: new
      • Abstract: Language generation research increasingly spans three paradigms: autoregressive decoding, discrete masked diffusion, and continuous flow-matching
      • Comparing them is difficult because each lives in a separate codebase, so measured differences often reflect implementation details rather than the paradigms th…
      • We present DantinoX, an open-source JAX/Flax library in which a single modular Transformer backbone serves all three paradigms

  • Think Before You Comfort: Reflective Cognitive Alignment for Protocol-Grounded Elderly Stimulation Agents

    • Publication Time: 2026-09-17 12:00 Beijing Time
    • Abstract: [TO BE TRANSLATED] - arXiv:2609.17536v1 Announce Type: new.

  • Abstract: Cognitive Stimulation Therapy (CST) offers non-pharmacological support for elders with cognitive impairment, yet scalability remains constrained by reliance on trained facilitators and severe data scarcity, particularly for privacy-sensitive, low-resource languages such as Cantonese.

  • While Large Language Models (LLMs) show promise for automated companionship, they often struggle to balance empathetic engagement with adherence to cognitive stimulation guidelines.

  • We propose a framework addressing these challenges along two complementary axes.

    • EN Key Points:
      • arXiv:2609.17536v1 Announce Type: new
      • Abstract: Cognitive Stimulation Therapy (CST) offers non-pharmacological support for elders with cognitive impairment, yet scalability remains constrained by re…
      • While Large Language Models (LLMs) show promise for automated companionship, they often struggle to balance empathetic engagement with adherence to cognitive st…
      • We propose a framework addressing these challenges along two complementary axes

  • Relation Before Entity: Deferred Commitment in Language Model Factual Recall

    • Publication Time: 2026-09-17 12:00 Beijing Time
    • Abstract: arXiv:2609.17537v1 Announce Type: new.
      • Abstract: We ask whether relation-type information (e.g., capital-of) and entity-specific information (e.g., France to Paris) become causally active at the final-token position at the same depth during recall.
      • Using four complementary causal diagnostics across four decoder-only models and eight prompt families, we find a robust temporal asymmetry: relation information becomes generation-controlling before entity information does.
      • Relation onset precedes entity onset by 10-16 tested layers (31-44% of network depth) at threshold 0.4, with the ordering holding across all 16 model-threshold combinations for thresholds 0.2-0.5.
    • EN Key Points:
      • arXiv:2609.17537v1 Announce Type: new

  • Abstract: We ask whether relation-type information (e.g., capital-of) and entity-specific information (e.g., France to Paris) become causally active at the fina…

  • Using four complementary causal diagnostics across four decoder-only models and eight prompt families, we find a robust temporal asymmetry: relation information…

  • Relation onset precedes entity onset by 10-16 tested layers (31-44% of network depth) at threshold 0.4, with the ordering holding across all 16 model-threshold…

  • From Pixels to Pairs: A Comprehensive Benchmark of LLM-Based Key-Value Extraction in Noisy Document Settings

    • Release Time: 2026-09-17 12:00 Beijing Time
    • Summary: [Translation pending] - arXiv:2609.17538v1 Announce Type: new.
      • Abstract: Large language models (LLMs) are increasingly used for structured information extraction from documents, yet their behavior under realistic OCR noise remains poorly understood.
      • We present a systematic benchmark of open-source instruction-tuned LLMs for key-value pair (KVP) extraction under both clean-text and noisy OCR conditions.
      • We evaluate representative decoder-only models (Gemma, Mistral, Qwen2.5, LLaMA 3, and DeepSeek) on the FUNSD, CORD, and SROIE benchmarks using both Gold-text annotations and OCR outputs from PaddleOCR, EasyOCR, and Tesseract.
    • EN Highlights:
      • arXiv:2609.17538v1 Announce Type: new
      • Abstract: Large language models (LLMs) are increasingly used for structured information extraction from documents, yet their behavior under realistic OCR noise…
      • We present a systematic benchmark of open-source instruction-tuned LLMs for key-value pair (KVP) extraction under both clean-text and noisy OCR conditions
      • We evaluate representative decoder-only models (Gemma, Mistral, Qwen2.5, LLaMA 3, and DeepSeek) on the FUNSD, CORD, and SROIE benchmarks using both Gold-text an…

  • MudawanSn: A Gold-Standard Wolof-Arabic Parallel Corpus for Machine Translation

    • Release Time: 2026-09-17 12:00 Beijing Time

  • Summary: [Translation Pending] - arXiv:2609.17539v1 Announce Type: new.

    • Abstract: We present MudawanSn, a gold-standard resource of 1,271 sentence-aligned pairs manually translated from Wolof into Modern Standard Arabic (MSA).
    • The source texts are drawn from the MasakhaNER corpus and cover politics, society, religion, and sports in Senegalese news discourse.
    • Although multilingual resources such as FLORES-200 and NTREX include both Wolof and Arabic, no publicly available parallel corpus is specifically designed for the Wolof-Modern Standard Arabic language pair.
    • EN Key Points:
      • arXiv:2609.17539v1 Announce Type: new
      • Abstract: We present MudawanSn, a gold-standard resource of 1,271 sentence-aligned pairs manually translated from Wolof into Modern Standard Arabic (MSA)
      • The source texts are drawn from the MasakhaNER corpus and cover politics, society, religion, and sports in Senegalese news discourse
      • Although multilingual resources such as FLORES-200 and NTREX include both Wolof and Arabic, no publicly available parallel corpus is specifically designed for t…

  • Register Bias in Complexity-Based Large Language Model Routing

    • Publication Time: 2026-09-17 12:00 Beijing Time
    • Summary: [Translation Pending] - arXiv:2609.17542v1 Announce Type: new.
      • Abstract: Large language model services increasingly route each query to one of several models of differing capability, using a cheap estimate of query complexity to send easy queries to small models and hard queries to large ones.
      • I show that this routing step is not register neutral: text written in a non-standard English register, African American English or the English of second-language writers, is systematically assigned a lower-capacity tier than a meaning-equivalent standard-English version of the same query.

  • The effect is driven by a specific, common routing signal, input length, because non-standard registers omit function words and thus look shorter and therefore simpler; other complexity signals do not carry it.

    • EN Key Points:
      • arXiv:2609.17542v1 Announce Type: new
      • Abstract: Large language model services increasingly route each query to one of several models of differing capability, using a cheap estimate of query complexi…
      • I show that this routing step is not register neutral: text written in a non-standard English register, African American English or the English of second-langua…
      • The effect is driven by a specific, common routing signal, input length, because non-standard registers omit function words and thus look shorter and therefore…

  • Large Language Models Versus Physicians in Traditional Chinese Medicine: A Real-World Clinical Case Evaluation

    • Published: 2026-09-17 12:00 Beijing Time
    • Abstract: [Translation Pending] - arXiv:2609.17544v1 Announce Type: new.
      • Abstract: Large language models (LLMs) are increasingly being explored for clinical applications, yet their assessment for real-world traditional Chinese medicine (TCM) practice remains limited We constructed a clinical case library comprising 349 de-identified outpatient cases from 62 hospitals and evaluated 16 LLMs and a comparator cohort of 60 practicing TCM physicians using 60 representative cases selected from this library.
      • Model outputs and physician reports were anonymized and scored by five senior TCM experts across nine diagnostic and therapeutic dimensions.
      • Cutting-edge general-purpose LLMs achieved higher expert scores than the physician comparators, particularly for medical advice, treatment principles and selected diagnostic tasks.
    • EN Key Points:
      • arXiv:2609.17544v1 Announce Type: new
      • Abstract: Large language models (LLMs) are increasingly being explored for clinical applications, yet their assessment for real-world traditional Chinese medici…

  • Model outputs and physician reports were anonymized and scored by five senior TCM experts across nine diagnostic and therapeutic dimensions

  • Cutting-edge general-purpose LLMs achieved higher expert scores than the physician comparators, particularly for medical advice, treatment principles and select…

  • Legal LLM Hallucination Should Be Evaluated as Failure of Legal Warrant

    • Published: 2026-09-17 12:00 Beijing Time
    • Abstract: [Translation Pending] - arXiv:2609.17546v1 Announce Type: new.
      • Abstract: In this position paper, we argue that legal LLMs’ hallucinations should be evaluated as a failure of legal warrant rather than as factual inaccuracy or citation failure.
      • We define claim-authority warrant as the context-sensitive relation between a consequential legal claim and authority that exists, applies to the relevant jurisdiction, is current for the date of analysis, has the legal status represented by the system, and supports the proposition asserted.
      • Warranted legal generation is the broader system behavior that answers, narrows, asks, warns, corrects a false premise, or abstains according to that relation.
    • EN Highlights:
      • arXiv:2609.17546v1 Announce Type: new
      • Abstract: In this position paper, we argue that legal LLMs’ hallucinations should be evaluated as a failure of legal warrant rather than as factual inaccuracy o…
      • We define claim-authority warrant as the context-sensitive relation between a consequential legal claim and authority that exists, applies to the relevant juris…
      • Warranted legal generation is the broader system behavior that answers, narrows, asks, warns, corrects a false premise, or abstains according to that relation

ArXiv cs.LG (B_intro+search) Link to heading

  • Pay Only for Disagreement: Certified No-Regression Verdicts for Model Updates with Matching Label-Complexity Bounds

    • Published: 2026-09-17 12:00 Beijing Time
    • Abstract: [Translation Pending] - arXiv:2609.17560v1 Announce Type: new.

  • Abstract: Every production model is updated, by retraining, fine-tuning, quantization, or a silent vendor swap, and each update risks being worse than what it replaced.

    • We formalize update promotion as certified paired risk-difference auditing.
    • Our starting point is a support identity: the risk difference between two models lives on the inputs where they disagree, observable without labels.
    • EN Key Points:
      • arXiv:2609.17560v1 Announce Type: new
      • Abstract: Every production model is updated, by retraining, fine-tuning, quantization, or a silent vendor swap, and each update risks being worse than what it r…
      • We formalize update promotion as certified paired risk-difference auditing
      • Our starting point is a support identity: the risk difference between two models lives on the inputs where they disagree, observable without labels

  • Beyond Static RAG: An Adaptive, Tri-Metric Routing Framework for Efficient Long-Context Inference on Commodity GPUs

    • Published: 2026-09-17 12:00 Beijing Time
    • Abstract: [To be translated] - arXiv:2609.17564v1 Announce Type: new.
      • Abstract: Deploying retrieval-augmented generation (RAG) on commodity GPUs such as the NVIDIA T4 (16 GB VRAM) exposes a practical failure mode we call the Compression Paradox: neural prompt compression can add key-value (KV) cache contention and preprocessing latency that outweigh generation-time savings, while skipping compression can cause out-of-memory (OOM) failures on long contexts.
      • We identify two distinct failure mechanisms when a vLLM-served LLM and a PyTorch-based compressor are co-deployed under tight memory budgets, and introduce the Tri-Metric Router, a deterministic, training-free policy that selects among Raw, Neural (LLMLingua-2), and Lexical (BM25) pipelines.
      • The router uses three CPU-side signals: spatial complexity ($L$), syntactic density ($\rho_{key}$), and type-token ratio (TTR).
    • EN Key Points:
      • arXiv:2609.17564v1 Announce Type: new

  • Abstract: Deploying retrieval-augmented generation (RAG) on commodity GPUs such as the NVIDIA T4 (16 GB VRAM) exposes a practical failure mode we call the Compr…

    • We identify two distinct failure mechanisms when a vLLM-served LLM and a PyTorch-based compressor are co-deployed under tight memory budgets, and introduce the…
    • The router uses three CPU-side signals: spatial complexity ($L$), syntactic density ($\rho_{key}$), and type-token ratio (TTR)

  • Where Grokking Happens: Distributed Utility and Fourier Recoding Without a Module Switch

    • Publication Time: 2026-09-17 12:00 Beijing Time
    • Abstract: [Pending Translation] - arXiv:2609.17571v1 Announce Type: new.
      • Abstract: Where in a Transformer is the change from memorization to generalization functionally expressed?
      • We introduce Transition Games–behavior-aligned exact activation games with paired non-generalizing controls–and find distributed utility gain with a prospective block-0 attention bias; selected degree-two modes account for 67–92% of its addition contrast across replacement games, and a disjoint exact path study confirms that block-1 MLP mediates more of their effect than all other tested downstream paths in 12/12 pairs.
      • The sharper “MLP memorizes, attention generalizes” prediction instead reverses (-.331 bits/example at the memory anchor; 0/12 in the predicted direction), while routing onset, global rank collapse, and a prime-invariant architecture ridge also fail, identifying grokking here as spectral recoding of an existing distributed circuit rather than a module switch.
    • EN Highlights:
      • arXiv:2609.17571v1 Announce Type: new
      • Abstract: Where in a Transformer is the change from memorization to generalization functionally expressed
      • We introduce Transition Games–behavior-aligned exact activation games with paired non-generalizing controls–and find distributed utility gain with a prospecti…

  • The sharper “MLP memorizes, attention generalizes” prediction instead reverses (-.331 bits/example at the memory anchor; 0/12 in the predicted direction), while…

  • Disentangling Algorithmic Bias from Archival Artifacts: A Controlled Audit of Vision-Language Model Valuation in Metropolitan Museum Archives

    • Publication Date: 2026-09-17 12:00 Beijing Time
    • Abstract: arXiv:2609.17572v1 Announce Type: new.
      • Abstract: Auditing vision-language models (VLMs) for societal bias requires distinguishing direct algorithmic valuation disparities from confounders embedded within archival metadata.
      • In this study, we audit Contrastive Language-Image Pretraining (CLIP) models using historical artwork metadata from the Metropolitan Museum of Art Open Access collection (N = 1,500 total objects; N = 743 attributed works: Male n = 534, Female n = 209; n = 618 anonymous).
      • We establish a quantitative audit framework evaluating zero-shot CLIP logit differential scores across three semantic prompt pairs (masterpiece, quality, and influence).
    • Key Takeaways (EN):
      • arXiv:2609.17572v1 Announce Type: new
      • Abstract: Auditing vision-language models (VLMs) for societal bias requires distinguishing direct algorithmic valuation disparities from confounders embedded wi…
      • In this study, we audit Contrastive Language-Image Pretraining (CLIP) models using historical artwork metadata from the Metropolitan Museum of Art Open Access c…
      • We establish a quantitative audit framework evaluating zero-shot CLIP logit differential scores across three semantic prompt pairs (masterpiece, quality, and in…

  • Temperon: Full-Time SAM Quality at a Third Less Wall-Clock

    • Publication Date: 2026-09-17 12:00 Beijing Time
    • Abstract: arXiv:2609.17575v1 Announce Type: new.
      • Abstract: Sharpness-aware minimization (SAM) doubles the cost of every training step, yet its benefit concentrates where training ends.

  • When the Gradient Sees Rank: Provable Necessity, Causal Recruitment, and Composition in Trained Matrix Memories

    • Published: 2026-09-17 12:00 Beijing Time
    • Abstract: [To be translated] - arXiv:2609.17594v1 Announce Type: new.
      • Abstract: Can gradient-based training learn the rank needed to store and compose associations in a matrix memory?
      • In our earlier study, we used a matrix-augmented reasoner on a task that admits a rank-1 solution, leaving this question open.
      • We train matrix memories on $K$ fresh key-value bindings whose exact linear recovery requires $\mathrm{rank}(Z) \geq K$.
    • EN Key Points:
      • arXiv:2609.17594v1 Announce Type: new
      • Abstract: Can gradient-based training learn the rank needed to store and compose associations in a matrix memory

  • In our earlier study, we used a matrix-augmented reasoner on a task that admits a rank-1 solution, leaving this question open

  • We train matrix memories on $K$ fresh key-value bindings whose exact linear recovery requires $\mathrm{rank}(Z) \geq K$

  • Prior-Free Competitive Ratios for Improving Bandits: Scale, Curvature and Horizon Are Free, but Not Jointly Under Noise

    • Published: 2026-09-17 12:00 Beijing Time
    • Abstract: 【Translation Pending】- arXiv:2609.17595v1 Announce Type: new.
      • Abstract: In the improving multi-armed bandits problem, each of $k$ arms has an unknown nondecreasing, discretely concave reward curve $f_i$, and pulling arm $i$ for the $t$-th time yields $f_i(t)$.
      • For sufficiently long horizons, Blum and Ravichandran (ALT 2025) proved that randomized algorithms achieve an $O(\sqrt k)$ approximation to the best single arm when the scale $m=f^*(T)$ of the optimal arm is known ($T\ge2k$), and $O(\sqrt k\log k)$ when it is not ($T>4k$), against an $\Omega(\sqrt k)$ lower bound.
      • The logarithmic factor is unnecessary: a one-page \emph{probe-and-commit} algorithm achieves competitive ratio $4\sqrt3,\sqrt k$ for $T\ge2\lfloor\sqrt k\rfloor$, without any knowledge of the scale, and we determine the optimal ratio for every horizon, $\Theta(\sqrt k+k/T)$, also for unknown horizons.
    • EN Highlights:
      • arXiv:2609.17595v1 Announce Type: new
      • Abstract: In the improving multi-armed bandits problem, each of $k$ arms has an unknown nondecreasing, discretely concave reward curve $f_i$, and pulling arm $i…
      • For sufficiently long horizons, Blum and Ravichandran (ALT 2025) proved that randomized algorithms achieve an $O(\sqrt k)$ approximation to the best single arm…
      • The logarithmic factor is unnecessary: a one-page \emph{probe-and-commit} algorithm achieves competitive ratio $4\sqrt3,\sqrt k$ for $T\ge2\lfloor\sqrt k\rfloo…

  • Lecture notes on Physics Informed Neural Networks, Neural Operators, and their applications

    • Published at: 2026-09-17 12:00 Beijing Time
    • Abstract: [To be translated] - arXiv:2609.17638v1 Announce Type: new.
      • Abstract: This is the set of lecture notes for the PhD course \href{ Informed Neural Network}, held at the University of Bozen/Bolzano} in the academic year 2025/2026.
      • The goal of the course was to introduce the concept of Physics Informed Deep Neural Networks (PINN) and Neural Operators (NOs), discuss their implementation from scratch in PyTorch and using advanced ad-hoc developed open-source libraries such as NVIDia PhysicsNeMo to address real-world problems in various fields (engineering, physics, petroleum reservoir).
      • We discuss recent topics such as Mixture-of-Models, Fourier Neural Operators, Physics-Informed Kolmogorov-Arnold Networks (PIKANs) and Fourier Neural Operators.
    • EN Highlights:
      • arXiv:2609.17638v1 Announce Type: new
      • Abstract: This is the set of lecture notes for the PhD course \href{ Informed Neural Network}, held at the University of Bozen/Bolzano} in the academic year 202…
      • The goal of the course was to introduce the concept of Physics Informed Deep Neural Networks (PINN) and Neural Operators (NOs), discuss their implementation fro…
      • We discuss recent topics such as Mixture-of-Models, Fourier Neural Operators, Physics-Informed Kolmogorov-Arnold Networks (PIKANs) and Fourier Neural Operators.

  • Fathom: Per-Query Read Depth for Sparse Decoding over Offloaded KV Caches

    • Published at: 2026-09-17 12:00 Beijing Time
    • Abstract: [To be translated] - arXiv:2609.17652v1 Announce Type: new.
      • Abstract: When agentic sessions run to a million tokens with many sessions resident at once, the KV cache and the index that ranks it live in host memory, and the scan that ranks all n keys for a top-k step becomes the traffic that bounds decoding.

  • We present Fathom, a key scan in which each query decides how many bits of each key channel to read.

  • The 4-bit K cache is stored channel-major as bit planes, so a prefix of t planes is exactly the channel’s t-bit quantizer, and the query spends its bit budget by reverse water-filling over the variance-weighted importance of its channels.

  • EN Key Points:

    • arXiv:2609.17652v1 Announce Type: new
    • Abstract: When agentic sessions run to a million tokens with many sessions resident at once, the KV cache and the index that ranks it live in host memory, and t…
    • We present Fathom, a key scan in which each query decides how many bits of each key channel to read
    • The 4-bit K cache is stored channel-major as bit planes, so a prefix of t planes is exactly the channel’s t-bit quantizer, and the query spends its bit budget b…

  • Reflect, Revise, Reuse: Training-Free Skill Evolution for GUI Agents

    • Publication Time: 2026-09-17 12:00 Beijing Time
    • Abstract: [To be translated] - arXiv:2609.17653v1 Announce Type: new.
      • Abstract: GUI agents execute long-horizon tasks on dynamic graphical user interfaces, where pop-ups, delayed loads, and relocated widgets routinely invalidate plans fixed before execution.
      • Recent agent-skill frameworks encapsulate reusable procedural knowledge to mitigate this, yet existing skill designs are largely developed without targeting GUI execution dynamics and treat skills as static artifacts produced before deployment rather than living procedural knowledge that improves through it.
      • We argue that what GUI agents need is not better static skills, but skills that can be revised from execution feedback at deployment time, without additional training.
    • EN Key Points:
      • arXiv:2609.17653v1 Announce Type: new
      • Abstract: GUI agents execute long-horizon tasks on dynamic graphical user interfaces, where pop-ups, delayed loads, and relocated widgets routinely invalidate p…

  • Recent agent-skill frameworks encapsulate reusable procedural knowledge to mitigate this, yet existing skill designs are largely developed without targeting GUI…

  • We argue that what GUI agents need is not better static skills, but skills that can be revised from execution feedback at deployment time, without additional tr…