System translated (Gemini)

🤖 AI 速览

This issue focuses on two new changes in AI products: Gemini 3.8 Live promotes real-time interaction and extended thinking, while Claude’s invisible watermark explores content traceability. Meanwhile, advertising monetization, full-duplex voice, and reliability evaluation for real-world …
📋 文章元数据
发布时间
2026-09-16
类型
ai-daily
字数
7094
阅读时长
34 min

2026-09-16 AI Daily | From Real-Time Conversations to Invisible Watermarking: AI Products Begin to Tackle Commercialization and Credibility Link to heading

This issue focuses on two new developments in AI products: Gemini 3.8 Live is advancing real-time interaction and expanded thinking, while Claude’s invisible fingerprinting explores content provenance. Meanwhile, advertising monetization, full-duplex voice, and reliability benchmarks for real-world scenarios are becoming key areas of competition.

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

Three main themes are worth focusing on today. First, AI products are moving from simply “being able to answer” to real-time interaction and commercialization: Google DeepMind has released Gemini 3.8 Live and its expanded thinking capabilities, while Claude’s invisible text fingerprinting brings content provenance to the forefront. Meanwhile, discussions around ChatGPT ads and Amazon Ads reveal the monetization paths for assistant products. Second, voice models continue to evolve towards multi-language and full-duplex capabilities. Areas like low-resource speech recognition, speech-to-speech translation, and addressing real-time interaction flaws such as “interrupting” are key directions for engineering teams to watch. Third, the research community is re-evaluating the true reasoning abilities of large models, with the focus of benchmarks shifting from static Q&A to reliable, verifiable reasoning in real-world scenarios, covering everything from physics problems and network protocol state machines to redundant document comprehension and clinical time deviations.

🌐 AI Hot Topics on X Link to heading

Topic 1: OpenAI’s Agentic Factory Automates Software Development End-to-End Link to heading

  • Category: AI · News
  • Overview: Trending Time:, Related Posts: 123
  • What it is: OpenAI is reportedly building an “agentic factory” that can automate the software development process end-to-end, from requirements analysis and coding to testing and delivery.
  • Why it matters: This indicates that AI agents are moving beyond coding assistance to handling the entire software engineering lifecycle, which could significantly boost development efficiency and transform the software industry’s production methods and job structures.
  • Discussion summary: The discussion on X focuses on the practical maturity, reliability, and scalability of this end-to-end automation, as well as its potential impact on developer jobs, software quality, and corporate organizational models. Some believe it will drive autonomous software production, while others question if the claims are exaggerated.

Topic 2: AI Could Revive C Programming for Efficiency Gains Link to heading

  • Category: AI · Other
  • Overview: Trending Time:, Related Posts: 48
  • What it is: Discussions are growing about the possibility of AI driving a resurgence of the C programming language through its code generation and optimization capabilities.
  • Why it matters: C is widely used in operating systems, embedded devices, and high-performance software. If AI can lower the barriers to its development and maintenance, it could improve the efficiency of critical infrastructure and extend the lifespan of existing tech stacks.
  • Discussion summary: The debate centers on whether AI can genuinely improve the performance and development efficiency of C code, and the potential risks of memory safety, debugging, maintenance, and reliability associated with AI-generated code. Some argue this would merely enhance existing systems and does not signal a full-fledged revival for C.

Topic 3: Devin AI Now Builds and Tests iOS Apps in Cloud Mac VMs Link to heading

  • Category: AI · News
  • Overview: Trending Time: 7 hours ago, Related Posts: 1000
  • What it is: Devin AI is now able to build and test iOS applications in cloud-based Mac virtual machines.
  • Why it matters: This shows that the capabilities of AI programming agents are expanding from code generation to invoking real development environments, running tests, and completing the mobile app development workflow. This could lower the barrier to entry for iOS development and increase iteration speed.
  • Discussion summary: The discussion on X is focused on how Devin Local and Cloud work together, whether AI can reliably handle the iOS build and test processes, and practical issues such as the cost, performance, and permissions of cloud Macs, as well as the quality of the generated apps.

Topic 4: Salesforce Integrates CRM into Claude at Dreamforce Link to heading

  • Category: AI · News
  • Overview: Trending Time: 10 hours ago, Related Posts: 2700
  • What it is: Salesforce announced the integration of its Customer Relationship Management (CRM) features into Claude, allowing users to directly access and manipulate relevant enterprise data and workflows through Claude.
  • Why it matters: This marks a shift for enterprise-grade AI assistants from standalone chatbots to being deeply integrated with core business systems. This can help improve work efficiency while also impacting enterprise data access, security, and the competitive landscape of AI platforms.
  • Discussion summary: The discussion on X centers on the practical value of integrating Claude with the Salesforce ecosystem, enterprise data privacy and permission controls, and whether this move will intensify competition between Salesforce and other AI assistants and enterprise software platforms.

Topic 5: Factory Raises $200M at $5B Valuation in AI Coding Boom Link to heading

  • Category: AI · News
  • Overview: Trending Time: 6 hours ago, Related Posts: 1100
  • What happened: AI programming company Factory completes $200 million financing, reaching a valuation of $5 billion, demonstrating its rapid growth in the AI programming boom.
  • Why it matters: This indicates that the capital market remains highly optimistic about the commercial potential of AI programming tools and their automated software development, and may also intensify competition and valuation expansion in related fields.
  • Discussion overview: Discussions on X likely center on Factory’s product capabilities, commercialization prospects, and whether its $5 billion valuation is reasonable. Disagreements primarily revolve around whether AI programming can consistently deliver real productivity improvements and if there is a valuation bubble in the current market.

Topic 6:Meta’s Muse AI Agent Tops App Charts Days After Launch Link to heading

  • Category:AI · News
  • Overview:Trending time: 1 day ago, Related posts: 5800
  • What happened: Meta’s Muse AI Agent topped the app charts days after its launch.
  • Why it matters: This indicates that AI agents are rapidly moving from concepts and chat tools towards practical, mass-market applications, triggering platform-level competition.
  • Discussion overview: Discussions on X mainly focus on whether agents can form “prompt loops” through continuous task execution, whether their true utility and user growth are sustainable, and issues such as privacy and platform dependence.

Today’s AI Public Opinion Summary on X Link to heading

Today’s main public opinion trend is that AI is accelerating its progression from code generation and chat assistants to agents capable of invoking real development environments, connecting to enterprise business systems, and executing tasks end-to-end, potentially reshaping software development, mobile applications, and office workflows. The consensus is that the commercialization and application of AI programming and enterprise-grade agents are accelerating, with capital investment, platform integration, and user growth also heating up concurrently. Disagreements focus on whether the actual maturity, reliability, and sustained productivity of these products are exaggerated, and whether high valuations and expectations of “completely replacing programmers” are reasonable. Judgments like the “C language revival” are also seen more as enhancements to existing technologies rather than a complete return. Potential risks include issues with generated code safety and maintenance, cloud development costs and access control, enterprise data privacy, and the impact of agent errors on software quality, employment structures, platform competition, and user autonomy.

💡 Influencer Insights Link to heading

Today’s influencer insights are temporarily unavailable. Recommended to read in-depth Watch List content.

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

Time window: Last 3 days; Covering 22 sources; Total 36 updates

All-In Podcast (A_full) Link to heading

  • JD Vance on AI, Entitlement Fraud, Iran War, Israel, H-1B Abuse & the Midterms

    • Publish time: 2026-09-16 04:46 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) VP JD Vance joins The Besties

  • (1:40) Assessing the Trump Admin so far: Border control, re-industrialization, inflation, Iran, government debt, anti-fraud

  • (9:28) AI approach: Dealing with Doomerism, “Frankenstein,” and AI negativity

  • (16:23) H-1B abuse and the massive $1T+ of fraud in federal spending

  • Satya Nadella on the AI Doomer Slowdown, Microsoft’s Master Plan & Who Wins AI

    • Posted: 2026-09-16 01:10 Beijing Time
    • Abstract: [PENDING TRANSLATION] - 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 Key Points:
      • (0:00) Satya Nadella joins The Besties
      • (0:55) Dario’s blog, “pacing the frontier,” common sense AI safety
      • (6:28) The failure of AI CEO messaging, monitoring agents, what will a slowdown mean for new AI products
      • (14:22) Economic incentives for frontier lab doomerism, where the AI profits are

  • Elon Musk & Gwynne Shotwell on AI Risks and Peer Review, Starship, Terafab, SpaceX/Tesla Merger

    • Posted: 2026-09-15 11:42 Beijing Time
    • Abstract: [PENDING TRANSLATION] - 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 要点:
      • (0:00) SpaceX’s Gwynne Shotwell joins The Besties
      • (1:45) Gwynne’s SpaceX story, selling rockets, and working for Elon
      • (6:45) Running modern day SpaceX: AI, Starlink, Rockets, and X
      • (13:43) Direct to cell with Starlink, retiring rockets, competition

Stratechery by Ben Thompson (A_full) Link to heading

  • OpenAI Ads, Amazon Ads in ChatGPT, Walmart to Accept Apple Pay
    • Published: 2026-09-15 18:00 Beijing Time
    • Summary: [To be translated] - ChatGPT ads are working, and solve Amazon’s biggest problem with chatbots.
      • Then, Walmart finally gives in to Apple Pay, because fighting the status quo is hard.
      • $15 / month or $150 / year.
      • Substantial analysis of the news of the day delivered via three weekly emails or podcasts.
      • Stratechery Interviews.
    • EN 要点:
      • ChatGPT ads are working, and solve Amazon’s biggest problem with chatbots
      • Then, Walmart finally gives in to Apple Pay, because fighting the status quo is hard.

Google DeepMind Blog (A_full) Link to heading

  • Introducing Gemini 3.8 Live and 3.8 Live Extended Thinking
    • Published: 2026-09-16 01:05 Beijing Time
    • Summary: [To be translated] - Introducing Gemini 3.8 Live and 3.8 Live Extended Thinking.
      • This piece from Google DeepMind Blog explains how Introducing Gemini 3.8 Live and 3.8 Live Extended Thinking shapes the broader AI and infrastructure landscape.
      • It also surfaces practical implications for founders, operators, and investors following Introducing Gemini 3.8 Live and 3.8 Live Extended Thinking.
    • EN 要点:
  • Introducing Gemini 3.8 Live and 3.8 Live Extended Thinking

Two Minute Papers (B_intro+search) Link to heading

  • Claude Is Now Leaving Invisible Fingerprints In Its Text
    • Publish Time: 2026-09-15 18:12 Beijing Time
    • Summary: [TO BE TRANSLATED] - ❤️ Check out Weights & Biases and sign up for a free demo here:.
      • 📝 The papers and sources are available here:.
      • Adam Bridges, B Shang, Carlos Galarza, Christian Ahlin, Eric Tyson, Juan Benet, Lukas Biewald, Michael Tedder, Owen Skarpness, Ryan Stankye, Shawn Becker, Steef, Taras Bobrovytsky, Tazaur Sagenclaw, Tybie Fitzhugh, Ueli Gallizzi.
      • Claude Is Now Leaving Invisible Fingerprints In Its Text.
    • EN Highlights:
      • ❤️ Check out Weights & Biases and sign up for a free demo here:
      • 📝 The papers and sources are available here:
      • 🙏 We would like to thank our generous Patreon supporters who make Two Minute Papers possible:
      • Adam Bridges, B Shang, Carlos Galarza, Christian Ahlin, Eric Tyson, Juan Benet, Lukas Biewald, Michael Tedder, Owen Skarpness, Ryan Stankye, Shawn Becker, Steef…

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

  • ZGCM-1: A Fully Open and Extremely Efficient Foundation Model for Math and Agentic Search

    • Publish Time: 2026-09-15 12:00 Beijing Time
    • Summary: [TO BE TRANSLATED] - arXiv:2609.13356v1 Announce Type: new.
      • Abstract: In this work, we present ZGCM-1, a fully open 7B dense foundation model trained from scratch with extreme data, system, and algorithmic efficiency.
      • ZGCM-1 is founded on a core premise: compact models cannot passively memorize the open web, but can overcome parametric capacity limits by coupling deliberate internal thinking with active external tool use.

  • To support this paradigm across a 256K context, we develop an end-to-end, high-efficiency open training recipe: Architecture & System Co-design: interleaved gated sliding-window and full attention, and a stable FP8 Muon optimizer; Progressive Curriculum & MDP Mid-Training: context scaling across 16K, 64K, and 256K, and the reformulation of interaction traces into Markov Decision Processes.

    • EN Key Points:
      • arXiv:2609.13356v1 Announce Type: new
      • Abstract: In this work, we present ZGCM-1, a fully open 7B dense foundation model trained from scratch with extreme data, system, and algorithmic efficiency
      • ZGCM-1 is founded on a core premise: compact models cannot passively memorize the open web, but can overcome parametric capacity limits by coupling deliberate i…
      • To support this paradigm across a 256K context, we develop an end-to-end, high-efficiency open training recipe: Architecture & System Co-design: interleaved gat…

  • Converge Then Diversify: Decoupling Convergence and Diversity in Multi-Objective Bayesian Optimisation

    • Published: 2026-09-15 12:00 Beijing Time
    • Abstract: [Awaiting translation] - arXiv:2609.13396v1 Announce Type: new.
      • Abstract: Multi-objective Bayesian optimisation (MOBO) is a sample-efficient approach for optimising expensive black-box functions with multiple objectives.
      • In MOBO, the goal is to adequately approximate the Pareto front; that is, to obtain a high-quality solution set with 1) good convergence (closeness to the Pareto front) and 2) good diversity (spread across the Pareto front).
      • Existing MOBO methods typically aim to accomplish these two tasks simultaneously, i.e., driving the search towards the Pareto front while maintaining a diverse set of nondominated solutions, such that the solutions, ideally, can gradually approach the entire front.
    • EN Key Points:
      • arXiv:2609.13396v1 Announce Type: new

  • Abstract: Multi-objective Bayesian optimisation (MOBO) is a sample-efficient approach for optimising expensive black-box functions with multiple objectives

    • In MOBO, the goal is to adequately approximate the Pareto front; that is, to obtain a high-quality solution set with 1) good convergence (closeness to the Paret…
    • Existing MOBO methods typically aim to accomplish these two tasks simultaneously, i.e., driving the search towards the Pareto front while maintaining a diverse…

  • Generalized Agent Iteration: One Formal Framework for Iterative Policy Improvement and Recursive Self-Improvement

    • Release Time: 2026-09-15 12:00 Beijing Time
    • Abstract: [To be translated] - arXiv:2609.13406v1 Announce Type: new.
      • Abstract: When we speak of recursive self-improvement (RSI), are we speaking of a phenomenon, a mechanism, or a prospect?
      • Towards autonomous and evolving intelligence, RSI is being claimed at many scales, while no single framework that formally describes these emerging instances exists.
      • Its counterpart in the classical realm, iterative policy improvement, is characterized by generalized policy iteration (GPI), a framework of broad applicability with well-understood theoretical properties, but only where the update principle and the evaluation base lie outside the agent.
    • EN Highlights:
      • arXiv:2609.13406v1 Announce Type: new
      • Abstract: When we speak of recursive self-improvement (RSI), are we speaking of a phenomenon, a mechanism, or a prospect
      • Towards autonomous and evolving intelligence, RSI is being claimed at many scales, while no single framework that formally describes these emerging instances ex…
      • Its counterpart in the classical realm, iterative policy improvement, is characterized by generalized policy iteration (GPI), a framework of broad applicability…

  • Vibe Patenting: Evaluating LLM Judges for Professional Patent-Drafting Agents

    • Release Time: 2026-09-15 12:00 Beijing Time

  • Summary: [Translation pending] - arXiv:2609.13422v1 Announce Type: new.

    • Abstract: LLM judges are increasingly used to evaluate and improve AI-generated outputs, yet their reliability for complex professional work remains unclear.
    • We study this problem through Vibe Patenting, an end-to-end patent-drafting testbed for AI-agent evaluation.
    • A separately-invoked LLM judge evaluates generated patent drafts and provides structured feedback for iterative revision.
    • EN Highlights:
      • arXiv:2609.13422v1 Announce Type: new
      • Abstract: LLM judges are increasingly used to evaluate and improve AI-generated outputs, yet their reliability for complex professional work remains unclear
      • We study this problem through Vibe Patenting, an end-to-end patent-drafting testbed for AI-agent evaluation
      • A separately-invoked LLM judge evaluates generated patent drafts and provides structured feedback for iterative revision

  • Toward Self-Adaptive Physical AI: Can LLM Agents Manage Long-Horizon Physical Tasks?

    • Publication Time: 2026-09-15 12:00 Beijing Time
    • Summary: [Translation pending] - arXiv:2609.13436v1 Announce Type: new.
      • Abstract: Large Language Model (LLM) agents offer a promising path toward autonomously managing long-term physical tasks without human intervention.
      • However, physical tasks require agents to continuously observe the environment, make consequential actions, and remain effective as the environment changes.
      • Existing approaches either require substantial data and retraining, or primarily focus on agents operating in the virtual world.
    • EN Highlights:
      • arXiv:2609.13436v1 Announce Type: new
      • Abstract: Large Language Model (LLM) agents offer a promising path toward autonomously managing long-term physical tasks without human intervention
      • However, physical tasks require agents to continuously observe the environment, make consequential actions, and remain effective as the environment changes

  • Existing approaches either require substantial data and retraining, or primarily focus on agents operating in the virtual world

  • LabAgent: Customize Any Research Hubs for Scientific Discoveries Using AI Agents

    • Published: 2026-09-15 12:00 Beijing Time
    • Abstract: [To be translated] - arXiv:2609.13437v1 Announce Type: new.
      • Abstract: Scientific research is a continuous process that emphasizes inheritance.
      • Methods developed by predecessors are often expanded upon by new researchers to explore more novel and in-depth scientific questions.
      • However, the change of lab staff, such as student graduation, leads to a lack of personnel capable of replicating methods.
    • Key Points:
      • arXiv:2609.13437v1 Announce Type: new
      • Abstract: Scientific research is a continuous process that emphasizes inheritance
      • Methods developed by predecessors are often expanded upon by new researchers to explore more novel and in-depth scientific questions
      • However, the change of lab staff, such as student graduation, leads to a lack of personnel capable of replicating methods

  • TimeThink: Eliciting Compositional Reasoning in Timeseries Large Language Models

    • Published: 2026-09-15 12:00 Beijing Time
    • Abstract: [To be translated] - arXiv:2609.13457v1 Announce Type: new.
      • Abstract: Timeseries multimodal large language models (TS-MLLMs) have recently begun leveraging the reasoning capabilities of large language models (LLMs) for question-answering tasks.
      • However, these models often fail to capture dynamic temporal patterns, providing only implicit reasoning that lacks the underlying explanations critical for high-stakes applications like healthcare.
      • While reinforcement learning (RL)-based timeseries language models aim to address this, they often fall short because they are trained on narrow, in-distribution data and struggle with out-of-distribution compositional questions.
    • Key Points:
      • arXiv:2609.13457v1 Announce Type: new

  • Abstract: Timeseries multimodal large language models (TS-MLLMs) have recently begun leveraging the reasoning capabilities of large language models (LLMs) for q…

  • However, these models often fail to capture dynamic temporal patterns, providing only implicit reasoning that lacks the underlying explanations critical for hig…

  • While reinforcement learning (RL)-based timeseries language models aim to address this, they often fall short because they are trained on narrow, in-distributio…

  • Root-Cause Attribution Is a Search Problem: Continual Search for Long-Horizon Agent Failures

    • Published: 2026-09-15 12:00 Beijing Time
    • Abstract: [Translation pending] - arXiv:2609.13463v1 Announce Type: new.
      • Abstract: The increasing deployment of AI agents in long-horizon tasks yields massive execution logs.
      • Diagnosing failures within these records is crucial for reliability, as it transforms outcome-level signals into actionable interventions.
      • The sheer scale of the data renders human review impractical, driving the need for automated root-cause attribution (RCA).
    • EN Key Points:
      • arXiv:2609.13463v1 Announce Type: new
      • Abstract: The increasing deployment of AI agents in long-horizon tasks yields massive execution logs
      • Diagnosing failures within these records is crucial for reliability, as it transforms outcome-level signals into actionable interventions
      • The sheer scale of the data renders human review impractical, driving the need for automated root-cause attribution (RCA)

  • Governing at Machine Speed: An Adaptive Intelligence Architecture for Real-Time AI Policy Enforcement

    • Published: 2026-09-15 12:00 Beijing Time
    • Abstract: [Translation pending] - arXiv:2609.13466v1 Announce Type: new.
      • Abstract: Enterprise AI adoption has reached 78% of organizations globally, yet the infrastructure to govern that adoption has not kept pace.

  • This paper identifies and characterizes the attestation deficit, a structural condition in which organizations maintain governance policies but cannot produce auditable, tamper-evident evidence of enforcement within regulatory timelines.

  • Drawing on empirical data from the Stanford 2026 AI Index Report (362 documented incidents), the IBM/Ponemon 2026 Cost of a Data Breach study (USD 4.99M average cost, 92% lacking access controls), and the EY/AIUC-1 Consortium survey (38% end-to-end monitoring, 17% agent-to-agent coverage), this paper demonstrates that the governance failure is organizational and architectural rather than technical.

  • EN Key Points:

    • arXiv:2609.13466v1 Announce Type: new
    • Abstract: Enterprise AI adoption has reached 78% of organizations globally, yet the infrastructure to govern that adoption has not kept pace
    • This paper identifies and characterizes the attestation deficit, a structural condition in which organizations maintain governance policies but cannot produce a…
    • Drawing on empirical data from the Stanford 2026 AI Index Report (362 documented incidents), the IBM/Ponemon 2026 Cost of a Data Breach study (USD 4.99M average…

  • OrchSLM: Probing the Dynamics of Small Language Model Orchestration

    • Publication Time: 2026-09-15 12:00 Beijing Time
    • Abstract: [Awaiting Translation] - arXiv:2609.13470v1 Announce Type: new.
      • Abstract: Although large language models (LLMs) have demonstrated remarkable capabilities, their reliance on cloud-scale infrastructure poses fundamental challenges for deployment in agentic pipelines, including latency, privacy, connectivity, and substantial computational cost.
      • Small language models (SLMs) offer a compelling alternative: recent studies suggest that many repetitive and narrowly scoped subtasks in agentic workloads may be better served by specialized SLMs than by monolithic LLMs.

  • However, the limited capacity and context windows of SLMs can constrain long-horizon reasoning and interaction-heavy orchestration strategies such as iterative verification and debate.

    • EN Key Points:
      • arXiv:2609.13470v1 Announce Type: new
      • Abstract: Although large language models (LLMs) have demonstrated remarkable capabilities, their reliance on cloud-scale infrastructure poses fundamental challe…
      • Small language models (SLMs) offer a compelling alternative: recent studies suggest that many repetitive and narrowly scoped subtasks in agentic workloads may b…
      • However, the limited capacity and context windows of SLMs can constrain long-horizon reasoning and interaction-heavy orchestration strategies such as iterative…

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

  • Token Merging for Multilingual Speech Recognition: A Systematic Study Across Model Scale and Fine-Tuning

    • Published: 2026-09-15 12:00 Beijing Time
    • Abstract: [Translation Pending] - arXiv:2609.13151v1 Announce Type: new.
      • Abstract: Leading multilingual speech recognition models like Whisper transcribe diverse, low-resource languages without language-specific training but are computationally expensive to deploy.
      • Token merging mitigates this inefficiency by dynamically combining redundant features, shortening the sequence length during inference without requiring retraining.
      • In this paper, we systematically evaluate token merging on the Whisper model family across sixteen diverse languages and three different model sizes.
    • EN Key Points:
      • arXiv:2609.13151v1 Announce Type: new
      • Abstract: Leading multilingual speech recognition models like Whisper transcribe diverse, low-resource languages without language-specific training but are comp…
      • Token merging mitigates this inefficiency by dynamically combining redundant features, shortening the sequence length during inference without requiring retrain…

  • In this paper, we systematically evaluate token merging on the Whisper model family across sixteen diverse languages and three different model sizes

  • PhysMent: An Interactive Approach For LLM Reasoning In Physics Problems

    • Release Time: 2026-09-15 12:00 Beijing Time
    • Abstract: [Translation Pending] - arXiv:2609.13152v1 Announce Type: new.
      • Abstract: Large language models (LLMs) perform strongly on static science benchmarks, yet their ability to reason about the physical world through active experimentation remains poorly understood.
      • We introduce PhysMent, a benchmark that evaluates LLM physical reasoning via iterative, toolmediated interaction with a MuJoCo physics simulator.
      • Unlike static benchmarks that supply all quantities upfront, PhysMent requires models to discover information by applying forces, querying object states, advancing time, and modifying scene geometry before answering.
    • EN Highlights:
      • arXiv:2609.13152v1 Announce Type: new
      • Abstract: Large language models (LLMs) perform strongly on static science benchmarks, yet their ability to reason about the physical world through active experi…
      • We introduce PhysMent, a benchmark that evaluates LLM physical reasoning via iterative, toolmediated interaction with a MuJoCo physics simulator
      • Unlike static benchmarks that supply all quantities upfront, PhysMent requires models to discover information by applying forces, querying object states, advanc…

  • Lexical Prompt Compression for Large Language Models: A Training-Free, Deterministic Pipeline with Empirical Pareto Analysis Across Eleven Task Categories

    • Release Time: 2026-09-15 12:00 Beijing Time
    • Abstract: [Translation Pending] - arXiv:2609.13154v1 Announce Type: new.
      • Abstract: Recent advances in large language models (LLMs) have made prompts increasingly large and complex.

  • Techniques such as chain-of-thought reasoning (Wei et al., 2022) and in-context learning (Brown et al., 2020) frequently push real-world prompts past several thousand tokens, increasing inference cost and latency.

  • Learned compression methods such as LLMLingua (Jiang et al., 2023) and Selective Context (Li et al., 2023) achieve high compression ratios but require auxiliary language models and are non-deterministic.

    • EN Key Points:
      • arXiv:2609.13154v1 Announce Type: new
      • Abstract: Recent advances in large language models (LLMs) have made prompts increasingly large and complex
      • Techniques such as chain-of-thought reasoning (Wei et al., 2022) and in-context learning (Brown et al., 2020) frequently push real-world prompts past several th…
      • Learned compression methods such as LLMLingua (Jiang et al., 2023) and Selective Context (Li et al., 2023) achieve high compression ratios but require auxiliary…

  • TestHallVQA: Exploring LVLMs’ Document-Level Reasoning under Redundant Contexts from Scientific Exams

    • Published: 2026-09-15 12:00 Beijing Time
    • Abstract: [To be translated] - arXiv:2609.13158v1 Announce Type: new.
      • Abstract: Large Vision–Language Models (LVLMs) are increasingly expected to perform visual question answering (VQA) over planar media.
      • However, existing planar VQA benchmarks typically emphasize isolated challenges: some emphasize long-document understanding with limited reasoning depth, while others require complex visual reasoning but remain restricted to single-page, noise-free settings.
      • Moreover, through theoretical analysis, we identify the impact of irrelevant visual tokens, which leads to measurable performance degradation but has received little attention with respect to systematic quantification.
    • EN Key Points:
      • arXiv:2609.13158v1 Announce Type: new
      • Abstract: Large Vision–Language Models (LVLMs) are increasingly expected to perform visual question answering (VQA) over planar media

  • However, existing planar VQA benchmarks typically emphasize isolated challenges: some emphasize long-document understanding with limited reasoning depth, while…

  • Moreover, through theoretical analysis, we identify the impact of irrelevant visual tokens, which leads to measurable performance degradation but has received l…

  • Clinical Reasoning Under a Partially Observed Objective in Cone Beam CT Report Generation

    • Release Time: 2026-09-15 12:00 Beijing Time
    • Abstract: arXiv:2609.13238v1 Announce Type: new.
      • Abstract: Maxillofacial report generation from cone beam computed tomography is scored here by a composite objective placing 80% of its weight on a large language model judgement of factual entailment and 20% on lexical overlap, of which only the lexical fifth is visible during development.
      • The grader’s BLEU-4 and METEOR routines are reproduced in pure Python and match the reference to machine precision, and an offline entailment surrogate, which tells a report written for one patient from one written for another at an area under the curve of 0.987, makes the composite objective cheap enough to optimise directly.
      • Over the 622-case public release, a report selected against the visible lexical ranking scores 0.2909, whereas one selected against the composite objective scores 0.4122, because pursuing n-gram overlap drives entailment precision from 0.522 down to 0.266.
    • EN Key Points:
      • arXiv:2609.13238v1 Announce Type: new
      • Abstract: Maxillofacial report generation from cone beam computed tomography is scored here by a composite objective placing 80% of its weight on a large langua…
      • The grader’s BLEU-4 and METEOR routines are reproduced in pure Python and match the reference to machine precision, and an offline entailment surrogate, which t…
      • Over the 622-case public release, a report selected against the visible lexical ranking scores 0.2909, whereas one selected against the composite objective scor…

  • RFCLLM: Evaluating LLMs’ Reasoning Ability of Network Protocol State Machines

    • Publication Date: 2026-09-15 12:00 Beijing Time
    • Abstract: arXiv:2609.13389v1 Announce Type: new.
      • Abstract: Mapping textual specifications into formal representations is essential for ensuring the correctness of protocol designs and implementations.
      • LLM-generated mappings, used for networking security or testing, are assumed to capture a perfect understanding of the specification, which may not hold in practice.
      • The goal of this paper is to assess the extent to which LLMs can interpret the specification correctly.
    • Key Points:
      • arXiv:2609.13389v1 Announce Type: new
      • Abstract: Mapping textual specifications into formal representations is essential for ensuring the correctness of protocol designs and implementations
      • LLM-generated mappings, used for networking security or testing, are assumed to capture a perfect understanding of the specification, which may not hold in prac…
      • The goal of this paper is to assess the extent to which LLMs can interpret the specification correctly

  • CVSS-X: A Multilingual Speech-to-Speech Translation Corpus for 28 Languages

    • Publication Date: 2026-09-15 12:00 Beijing Time
    • Abstract: arXiv:2609.13413v1 Announce Type: new.
      • Abstract: We introduce CVSS-X, a large-scale synthetic speech-to-speech translation corpus that extends CVSS by reversing the translation direction.
      • While CVSS translates from 21 languages into English, CVSS-X enables translation from English into 28 target languages spanning 12 language families.
      • The corpus comprises approximately 240,000 parallel speech pairs per language, totaling over 16,000 hours, eight times larger than CVSS.
    • Key Points:
      • arXiv:2609.13413v1 Announce Type: new
      • Abstract: We introduce CVSS-X, a large-scale synthetic speech-to-speech translation corpus that extends CVSS by reversing the translation direction

  • While CVSS translates from 21 languages into English, CVSS-X enables translation from English into 28 target languages spanning 12 language families

  • The corpus comprises approximately 240,000 parallel speech pairs per language, totaling over 16,000 hours, eight times larger than CVSS

  • Causal Analysis and Mitigation of Spurious Onsets in Full-Duplex Speech LLMs

    • Publication Time: 2026-09-15 12:00 Beijing Time
    • Abstract: arXiv:2609.13445v1 Announce Type: new.
      • Abstract: Speech-to-speech LLMs like Moshi, and its derivative PersonaPlex, can listen and speak concurrently through full-duplex generation.
      • However, they can begin speaking inappropriately during prolonged user silence: under digital-zero input, Moshi and PersonaPlex initiate speech in 12/40 and 11/40 five-minute continuations, respectively.
      • What causes this spurious speech?
    • Key Points:
      • arXiv:2609.13445v1 Announce Type: new
      • Abstract: Speech-to-speech LLMs like Moshi, and its derivative PersonaPlex, can listen and speak concurrently through full-duplex generation
      • However, they can begin speaking inappropriately during prolonged user silence: under digital-zero input, Moshi and PersonaPlex initiate speech in 12/40 and 11/…
      • What causes this spurious speech

  • Hindsight Bias in Clinical Temporal Reasoning: How Future Data Exposure Affects Large Language Model Judgment

    • Publication Time: 2026-09-15 12:00 Beijing Time
    • Abstract: arXiv:2609.13454v1 Announce Type: new.
      • Abstract: Clinical decisions are prospective, but clinical language models are often evaluated on retrospective records that reveal the final diagnosis, treatment response, and outcome.
      • Such evaluations may reward the use of future information rather than reasoning under the uncertainty present at the decision point.
      • We introduce a paired benchmark for measuring outcome-conditioned shifts consistent with hindsight bias in clinical temporal reasoning.
    • Key Points:

  • A Hybrid Hierarchical 1D-CNN-BiLSTM Framework for Extractive Summarization of Biomedical and Clinical Text

    • Publication Time: 2026-09-15 12:00 Beijing Time
    • Abstract: [Translation Pending] - arXiv:2609.13481v1 Announce Type: new.
      • Abstract: Large language models have made abstractive summarization remarkably fluent, but generated summaries can hallucinate facts, posing serious risks in biomedical and clinical domains.
      • We address this by removing generation from the pipeline and framing summarization as extractive sentence selection.
      • Our Hybrid Hierarchical CNN-LSTM Summarizer uses stacked multi-kernel convolutions to compose sentence-level embeddings into richer inter-sentence representations, followed by a bidirectional LSTM to model long-range dependencies across the document.
    • EN Key Points:
      • arXiv:2609.13481v1 Announce Type: new
      • Abstract: Large language models have made abstractive summarization remarkably fluent, but generated summaries can hallucinate facts, posing serious risks in bi…
      • We address this by removing generation from the pipeline and framing summarization as extractive sentence selection
      • Our Hybrid Hierarchical CNN-LSTM Summarizer uses stacked multi-kernel convolutions to compose sentence-level embeddings into richer inter-sentence representatio…

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

  • BudgetBench: A Budget-Tiered Protocol and Pilot Harness for Memory Strategy Evaluation in Local Large Language Model Agents

  • Release Time: 2026-09-15 12:00 Beijing Time

    • Abstract: [Translation Pending] - arXiv:2609.13149v1 Announce Type: new.
      • Abstract: For local large language model agents, active context is a scarce resource: memory capacity, prefill latency, cache growth, and service objectives all constrain how many input tokens each call can afford.
      • We present BudgetBench, an active-budget protocol and reference harness that treats the per-call input-token budget as the independent variable when comparing memory strategies.
      • Holding the model, task, sampler, and decoding fixed, it sweeps budgets over 2K, 4K, 8K, 16K, and 32K tokens and records quality, budget utilization, latency, and, as a first-class outcome, budget-violation rates.
    • EN Highlights:
      • arXiv:2609.13149v1 Announce Type: new
      • Abstract: For local large language model agents, active context is a scarce resource: memory capacity, prefill latency, cache growth, and service objectives all…
      • We present BudgetBench, an active-budget protocol and reference harness that treats the per-call input-token budget as the independent variable when comparing m…
      • Holding the model, task, sampler, and decoding fixed, it sweeps budgets over 2K, 4K, 8K, 16K, and 32K tokens and records quality, budget utilization, latency, a…

  • A derivative-fidelity failure mode in physics-informed neural networks: strengthened benchmark evidence from function-value training

    • Release Time: 2026-09-15 12:00 Beijing Time
    • Abstract: [Translation Pending] - arXiv:2609.13171v1 Announce Type: new.
      • Abstract: Physics-informed neural networks (PINNs) use automatic differentiation to impose differential-equation residuals, but good agreement in function values does not necessarily imply accurate derivatives.
      • This paper formulates derivative fidelity as a failure mode of PINNs and tests it with one-dimensional benchmarks.

  • Multilayer perceptrons are trained only on function values for sin(x) and exp(x), while second derivatives obtained by automatic differentiation are evaluated separately.

    • EN Key Points:
      • arXiv:2609.13171v1 Announce Type: new
      • Abstract: Physics-informed neural networks (PINNs) use automatic differentiation to impose differential-equation residuals, but good agreement in function value…
      • This paper formulates derivative fidelity as a failure mode of PINNs and tests it with one-dimensional benchmarks
      • Multilayer perceptrons are trained only on function values for sin(x) and exp(x), while second derivatives obtained by automatic differentiation are evaluated s…

  • Land Art as a Big-Data Climate Sensor

    • Published: 2026-09-15 12:00 Beijing Time
    • Abstract: [To be translated] - arXiv:2609.13182v1 Announce Type: new.
      • Abstract: Robert Smithson’s 1970 land artwork Spiral Jetty, located in the north arm of Utah’s Great Salt Lake, has alternated between submergence and exposure during severe lake decline.
      • We analyze 1,744 co-registered Landsat 4-9 and Sentinel-2 image chips spanning every year and calendar month from 1984 to 2025.
      • A 14-feature complexity signature combines Shannon entropy, multiscale permutation entropy, fractal dimension, lacunarity, gray-level co-occurrence texture, intensity statistics, and ImageNet-pretrained ResNet50 features.
    • EN Key Points:
      • arXiv:2609.13182v1 Announce Type: new
      • Abstract: Robert Smithson’s 1970 land artwork Spiral Jetty, located in the north arm of Utah’s Great Salt Lake, has alternated between submergence and exposure…
      • We analyze 1,744 co-registered Landsat 4-9 and Sentinel-2 image chips spanning every year and calendar month from 1984 to 2025
      • A 14-feature complexity signature combines Shannon entropy, multiscale permutation entropy, fractal dimension, lacunarity, gray-level co-occurrence texture, int…

  • LLMs or Naive Bayes? Old Gems or New Ways

  • Published: 2026-09-15 12:00 Beijing Time

    • Abstract: [To be translated] - arXiv:2609.13185v1 Announce Type: new.
      • Abstract: Large language models (LLMs) prompt a recurring question in research computing: should classical methods like Naive Bayes (NB) be retired?
      • We benchmark Complement Naive Bayes against zero-shot and few-shot LLMs spanning four model families and a 37x range in scale (27B to a 1T-parameter mixture-of-experts) across text classification tasks.
      • LLMs dominate only in zero-data regimes (98.0% vs 88.2% on Amazon Polarity sentiment), and even that win is contamination-prone: on a low-contamination sentiment task NB beats the zero-shot LLM (81.7% vs 73.0%).
    • EN Highlights:
      • arXiv:2609.13185v1 Announce Type: new
      • Abstract: Large language models (LLMs) prompt a recurring question in research computing: should classical methods like Naive Bayes (NB) be retired
      • We benchmark Complement Naive Bayes against zero-shot and few-shot LLMs spanning four model families and a 37x range in scale (27B to a 1T-parameter mixture-of-…
      • LLMs dominate only in zero-data regimes (98.0% vs 88.2% on Amazon Polarity sentiment), and even that win is contamination-prone: on a low-contamination sentimen…

  • Early Prediction of Satellite Collision Probability Using a Hybrid TCN-Transformer Model for a CDM-Based Conjunction Analysis Framework

    • Published: 2026-09-15 12:00 Beijing Time
    • Abstract: [To be translated] - arXiv:2609.13191v1 Announce Type: new.
      • Abstract: The rapid expansion of operational satellites and orbital debris has increased the frequency of close approach events in low Earth orbit (LEO), creating a higher operational burden for satellite operators.
      • This problem is especially critical for satellites using electric propulsion, where low-thrust maneuver capability imposes additional time constraints on collision avoidance planning.

  • In current practice, Conjunction Data Messages (CDMs) provide relative state, covariance, miss distance, time of closest approach, and probability of collision (PoC) information for conjunction assessment.

    • EN Key Points:
      • arXiv:2609.13191v1 Announce Type: new
      • Abstract: The rapid expansion of operational satellites and orbital debris has increased the frequency of close approach events in low Earth orbit (LEO), creati…
      • This problem is especially critical for satellites using electric propulsion, where low-thrust maneuver capability imposes additional time constraints on collis…
      • In current practice, Conjunction Data Messages (CDMs) provide relative state, covariance, miss distance, time of closest approach, and probability of collision…

  • Evaluating LLM-Generated Rules for Heart Disease Prediction

    • Publication Date: 2026-09-15 12:00 BJT
    • Abstract: [To be translated] - arXiv:2609.13192v1 Announce Type: new.
      • Abstract: This study compares traditional machine learning models and Large Language Model (LLM)-generated rule-based systems for heart disease prediction using the UCI Heart Disease dataset.
      • Several classifiers, including Logistic Regression, K-Nearest Neighbors (KNN), Support Vector Machine (SVM), Naive Bayes, Decision Tree, and Random Forest, were evaluated alongside rule-based systems generated using GPT-4o and Claude Sonnet 4.6.
      • Model performance was assessed using accuracy, precision, recall, and F1-score metrics.
    • EN Key Points:
      • arXiv:2609.13192v1 Announce Type: new
      • Abstract: This study compares traditional machine learning models and Large Language Model (LLM)-generated rule-based systems for heart disease prediction using…
      • Several classifiers, including Logistic Regression, K-Nearest Neighbors (KNN), Support Vector Machine (SVM), Naive Bayes, Decision Tree, and Random Forest, were…
      • Model performance was assessed using accuracy, precision, recall, and F1-score metrics

  • Diagnosing Faults in Reinforcement Learning Simulators and World Models with Canonical Polynomial Invariants

    • Publication Time: 2026-09-15 12:00 Beijing Time
    • Summary: [To be translated] - arXiv:2609.13194v1 Announce Type: new.
      • Abstract: A large literature builds physical structure into learned dynamics on the premise that models respecting the underlying physics predict better.
      • We test that premise using exact polynomial invariants recovered from trajectories and canonicalised as reduced Gr"obner bases over $\mathbb{Q}$.
      • On Acrobot, exactness provides little benefit for prediction: a consistency regulariser reduces algebraic residual while leaving rollout fidelity essentially unchanged, and a shaping potential recovered from a system with a 100% mass error accelerates learning as effectively as the correct potential.
    • EN Highlights:
      • arXiv:2609.13194v1 Announce Type: new
      • Abstract: A large literature builds physical structure into learned dynamics on the premise that models respecting the underlying physics predict better
      • We test that premise using exact polynomial invariants recovered from trajectories and canonicalised as reduced Gr"obner bases over $\mathbb{Q}$
      • On Acrobot, exactness provides little benefit for prediction: a consistency regulariser reduces algebraic residual while leaving rollout fidelity essentially un…

  • Machine Unlearning for Speech Question Answering in Large Audio-Language Models

    • Publication Time: 2026-09-15 12:00 Beijing Time
    • Summary: [To be translated] - arXiv:2609.13195v1 Announce Type: new.
      • Abstract: Large Audio-Language Models (LALMs) have recently shown strong capabilities in speech understanding and question answering (QA), but they also inherit privacy risks from large-scale training data, including the unintended memorization of sensitive information.

  • In this work, we study machine unlearning for speech QA in LALMs, a setting that is more challenging than prior work on text-based Large Language Models (LLMs) or Automatic Speech Recognition (ASR) due to the tight coupling between acoustic perception and factual knowledge.

    • We present and evaluate multiple unlearning strategies, including gradient ascent, task arithmetic, and alignment-based fine-tuning methods that enforce safe refusal responses, to remove private knowledge while still preserving performance on core capabilities.
    • EN Highlights:
      • arXiv:2609.13195v1 Announce Type: new
      • Abstract: Large Audio-Language Models (LALMs) have recently shown strong capabilities in speech understanding and question answering (QA), but they also inherit…
      • In this work, we study machine unlearning for speech QA in LALMs, a setting that is more challenging than prior work on text-based Large Language Models (LLMs)…
      • We present and evaluate multiple unlearning strategies, including gradient ascent, task arithmetic, and alignment-based fine-tuning methods that enforce safe re…

  • Algorithmic Information Dynamics of Learning: A Certified, Differentiable Complexity Controller for Grokking

    • Publication Time: 2026-09-15 12:00 Beijing Time
    • Abstract: [To be translated] - arXiv:2609.13197v1 Announce Type: new.
      • Abstract: Algorithmic Information Dynamics (AID) studies systems by perturbing them and measuring changes in algorithmic complexity, but its usual estimator, the Block Decomposition Method, is piecewise constant, restricting the calculus to finite differences.
      • We use $K^{\mathrm{CDM}}_{\mathrm{s}F}$, a certified, differentiable estimator, to bring the calculus into learning dynamics: grokking, where a complexity order parameter is known but has not been made to act.

  • A\empts a transient loss kick, the estimator becomes a controller that accelerates grokking in Levin’s description-length–versus-time sense, within a data-dependent Occam boundary whose finite-size trend, $f_c\sim\ln p/p$, is consistent with a coupon-collector interpretation.

    • EN Key Points:
      • arXiv:2609.13197v1 Announce Type: new
      • Abstract: Algorithmic Information Dynamics (AID) studies systems by perturbing them and measuring changes in algorithmic complexity, but its usual estimator, th…
      • We use $K^{\mathrm{CDM}}_{\mathrm{s}F}$, a certified, differentiable estimator, to bring the calculus into learning dynamics: grokking, where a complexity order…
      • A\empts a transient loss kick, the estimator becomes a controller that accelerates grokking in Levin’s description-length–versus-time sense, within a data-depe…

  • SPICE: Simple Polysemantic Feature Interpretation via Clustering-based Explanation

    • Publication Time: 2026-09-15 12:00 Beijing Time
    • Abstract: [Awaiting Translation] - arXiv:2609.13198v1 Announce Type: new.
      • Abstract: One of the pivotal recent challenges in neural network interpretability is polysemanticity, where a single neuron is activated by multiple, often unrelated concepts, hindering clear functional understanding.
      • Although prior work has explored this phenomenon, existing approaches remain architecture-specific and depend on manual heuristics such as a fixed number of concept clusters ($K$), limiting their generality and scalability–especially for modern Transformer-based models.
      • To address these limitations, we introduce SPICE (\textbf{S}imple \textbf{P}olysemantic Feature \textbf{I}nterpretation via \textbf{C}lustering-based \textbf{E}xplanation), a generalizable framework for analyzing polysemanticity in deep vision architectures.
    • EN Key Points:
      • arXiv:2609.13198v1 Announce Type: new

  • Abstract: One of the pivotal recent challenges in neural network interpretability is polysemanticity, where a single neuron is activated by multiple, often unre…

  • Although prior work has explored this phenomenon, existing approaches remain architecture-specific and depend on manual heuristics such as a fixed number of con…

  • To address these limitations, we introduce SPICE (\textbf{S}imple \textbf{P}olysemantic Feature \textbf{I}nterpretation via \textbf{C}lustering-based \textbf{E}…