Skip to content
AI News HubLIVE

This edition’s highlights

Agents

Genie One: The AI cowork home for everyone

  • Genie One becomes a unified platform for enterprise AI cowork, combining search, directories, artifacts, and actions.
  • Genie Search blends indexed search, live MCP connectors, and SQL across 100+ integrations.
Databricks BlogIn-site articleGenie One: The AI cowork home for everyone

Bring your business context to ChatGPT, Claude, Microsoft Copilot Cowork and Grok Bot

  • Genie plugins work natively in ChatGPT, Claude, Microsoft Copilot Cowork, and Grok Bot; other agents can access Genie via the MCP server.
  • Genie Ontology provides a governed context layer with certified definitions, relationships, and source context to help agents interpret business terms correctly.
Databricks BlogIn-site articleBring your business context to ChatGPT, Claude, Microsoft Copilot Cowork and Grok Bot
Chips

Modal becomes a gVisor maintainer

  • Google donates gVisor to the CNCF, moving to maintainer-led governance; Modal becomes one of three new maintainers.
  • Modal launches over 40 million gVisor containers per day for GPU inference, multi-node training, and AI coding agents.
Modal BlogIn-site articleModal becomes a gVisor maintainer
Research

How Much of Your VLA Is kNN

  • Training-free policy: retrieving nearest-neighbor demonstrations and reusing their action chunks achieves 91.7% success on LIBERO.
  • Fine-tuning nine public VLAs mostly turns the vision-language backbone into a better retrieval key, with actions remaining close to demonstrations.
arXiv RoboticsIn-site articleHow Much of Your VLA Is kNN

Evaluation protocols for historical script classification: the cost of manuscript-level leakage and the role of soft labels

  • Page-level accuracy drops from 97% to 54% (59% at manuscript level) and 39% on blind test, a significant 40-47 pp gap.
  • Models are overconfident on unseen manuscripts; patch-level ECE rises from 0.065 to 0.444; binarization reduces page-level overconfidence.
arXiv Computer VisionIn-site articleEvaluation protocols for historical script classification: the cost of manuscript-level leakage and the role of soft labels

One Table, Many Queries: Benchmarking Generalist Tabular Models Across Generation, Imputation, and Prediction

  • Introduces a unified benchmark with aligned data splits and explicit task contracts across 56 datasets, 46 methods, and 20 capability axes.
  • Rank comparisons link imputation and prediction via numerical recovery and dependence preservation, and connect generation to both through predictive utility, while distributional fidelity follows different rankings.
arXiv Machine LearningIn-site articleOne Table, Many Queries: Benchmarking Generalist Tabular Models Across Generation, Imputation, and Prediction
Models

Capacity Allocation at the Source: Sparse Target Optimization for LLM Knowledge Editing

  • Establishes a theoretical link between parameter interference and target hidden state interference.
  • Proposes an ℓ1 regularization method that prunes redundant 'tentacles' from knowledge vectors.
arXiv Computational LinguisticsIn-site articleCapacity Allocation at the Source: Sparse Target Optimization for LLM Knowledge Editing

Brasilidade: Measuring Brazilian Identity for Large Language Models

  • Brasilidade is a standardized, four-domain assessment methodology; tasks from its first domain are now public.
  • Model size alone does not determine performance; several medium-sized models outperformed larger models.
arXiv Computational LinguisticsIn-site articleBrasilidade: Measuring Brazilian Identity for Large Language Models

You have reached the end of this edition.

Your reading list
Other updates (66)
Agents

UAEF: A framework-agnostic system for closed-loop evaluation of AI agents

The Universal Agent Evaluation Framework (UAEF) is a framework-agnostic system for evaluating AI agents. UAEF converts each framework's traces into one canonical format, scores them with 36 built-in metrics across 7 dimensions, and stores results so teams can compare agent versions and catch regressions.

AWS Machine Learning BlogIn-site articleUAEF: A framework-agnostic system for closed-loop evaluation of AI agents

Introducing usage-based pricing for Amazon SageMaker AI bidirectional streaming

AWS Marketplace now supports usage-based metering for Amazon SageMaker AI bidirectional streaming inference, allowing model providers to charge for actual consumption during live streams. Deepgram is the launch partner, bringing usage-based speech-to-text and text-to-speech billing to AWS Marketplace.

AWS Machine Learning BlogIn-site articleIntroducing usage-based pricing for Amazon SageMaker AI bidirectional streaming

Genie with Genie Ontology Delivers 1.6× the Accuracy with Less than Half the Tokens

Databricks, in collaboration with Turing, built a shareable enterprise data-agent benchmark comparing Genie with two frontier agents in a live Databricks workspace. Genie passed 91% of runs versus 45% and 58%, while using less than half the tokens and finishing 3–4× faster, thanks to business context extracted from notebooks, dashboards, and saved SQL by Genie Ontology.

Databricks BlogIn-site articleGenie with Genie Ontology Delivers 1.6× the Accuracy with Less than Half the Tokens

Announcing GA of Genie Ontology

Databricks announces the general availability of Genie Ontology, an agentic ontology built by agents and governed by people to give every AI agent a shared, governed understanding of the business. The release adds Genie Search, external BI semantic model ingestion, organizational directories, and Ontology Manager. In a benchmark built with Turing, Genie with Genie Ontology passed 91% of task runs, 1.6× the strongest frontier agent, while using less than half the tokens.

Databricks BlogIn-site articleAnnouncing GA of Genie Ontology

DOBB: Dynamic Oriented Bounding Boxes for Multi-Agent Collision Avoidance under State Uncertainty and Heterogeneous Dynamics

DOBB introduces an online geometric representation that maps state estimates, uncertainty, and agent-specific characteristics into a dynamic, direction-dependent safety margin, achieving a mean total failure rate of 2.6% across four environments and keeping collision-based failure at or below 2.3% under out-of-distribution shifts.

arXiv RoboticsIn-site articleDOBB: Dynamic Oriented Bounding Boxes for Multi-Agent Collision Avoidance under State Uncertainty and Heterogeneous Dynamics

ASPIRE: An Agentic System-Identification Platform for Iterative Reasoning and Exploration

ASPIRE is a framework in which a large language model agent selects and executes nonlinear system identification methods, producing closed-form models with parameters, estimation records, interpretations, and error evaluated on data the agent cannot access. All 30 runs across three measured systems yielded valid models at 0.96 to 3.38 USD per run; an author assessment found 132 of 150 items fully satisfied, with model-structure justification the most frequent weakness.

arXiv RoboticsIn-site articleASPIRE: An Agentic System-Identification Platform for Iterative Reasoning and Exploration

Robo-MLT: Dual-System Robotic Manipulation for Long-Horizon Medical Laboratory Tasks

Robo-MLT is a dual-system robotic framework that couples explicit workflow monitoring with learned action generation to improve reliability and recoverability in long-horizon medical laboratory tasks. It achieves 82% and 70% success rates on blood gas analysis and blood sample tube sorting, significantly outperforming baselines, and can replan and continue after human intervention.

arXiv RoboticsIn-site articleRobo-MLT: Dual-System Robotic Manipulation for Long-Horizon Medical Laboratory Tasks

Auditable Human-Agent Teams for Autonomous X-ray Experimentation

A study demonstrates autonomous synchrotron X-ray experiments on nanoparticle superlattices using coding-agent infrastructure, with full transcripts enabling audit of decision provenance and cognitive labor division.

arXiv AIIn-site articleAuditable Human-Agent Teams for Autonomous X-ray Experimentation

Introducing Kimi K3.1 Pro on Amazon Bedrock

Kimi K3.1 Pro Highspeed from Moonshot AI is now available on Amazon Bedrock with a 1-million-token context window, 2–3x faster output, OpenAI-compatible APIs, Flex tier pricing, explicit prompt caching, and OpenCode integration for coding assistance.

AWS Machine Learning BlogIn-site articleIntroducing Kimi K3.1 Pro on Amazon Bedrock

JioStar’s chief architect on AI discovery at 500M-user scale

JioStar Chief Architect Vijay Seshadri shares lessons on building an AI discovery assistant for 500M users, covering evals, guardrails, and taking AI to production.

OpenAI NewsIn-site articleJioStar’s chief architect on AI discovery at 500M-user scale

Build a daily dependency watcher with LangGraph and Amazon Bedrock AgentCore web search

This post shows how to build a daily dependency watcher as a LangGraph graph on the Amazon Bedrock AgentCore Web Search API. It researches three package ecosystems in parallel, keeps rules like date window, blocklist, and search budget in code, and checks each finding against its cited sources.

AWS Machine Learning BlogIn-site articleBuild a daily dependency watcher with LangGraph and Amazon Bedrock AgentCore web search

The agent evaluation playbook: Built-in, DeepEval, and AutoEval on Amazon Bedrock AgentCore Evaluations

This post explains how to evaluate production agents on Amazon Bedrock AgentCore Evaluations by starting from business outcomes, breaking them into observable behaviors, and mapping each behavior to built-in, DeepEval, AutoEval, or custom evaluators across on-demand, online, and batch modes.

AWS Machine Learning BlogIn-site articleThe agent evaluation playbook: Built-in, DeepEval, and AutoEval on Amazon Bedrock AgentCore Evaluations

Gemini Intelligence: New features, on more devices

Gemini Intelligence is rolling out to more devices including motorola signature 27 and Pixel 10, with new features like Create My Widget and Device Assistance, plus expanded availability of Rambler and task automation.

Google AI BlogIn-site articleGemini Intelligence: New features, on more devices

Unity Catalog Metric Views, now in Power BI: one unified, governed semantic layer for all your reporting and AI

Databricks announces that Unity Catalog metric views can now be published directly to Power BI semantic models, enabling organizations to define business metrics once in Databricks and get consistent, governed results across Power BI, AI/BI dashboards, and Genie.

Databricks BlogIn-site articleUnity Catalog Metric Views, now in Power BI: one unified, governed semantic layer for all your reporting and AI

Autonomous treasury agents: How Trovata built an AI-native solution on Amazon Bedrock and AgentCore

Trovata built autonomous AI agents for corporate treasury on Amazon Bedrock and AgentCore, using an RBAC-secured architecture and bank-branded deployment to cut TMS costs, save team time, and reduce reporting from hours to under a minute.

AWS Machine Learning BlogIn-site articleAutonomous treasury agents: How Trovata built an AI-native solution on Amazon Bedrock and AgentCore

How TCS built an agentic test orchestrator with Amazon Bedrock AgentCore

Tata Consultancy Services (TCS) built an agentic test orchestrator on Amazon Bedrock AgentCore that turns requirements into executable, governed tests for its Banking, Financial Services, and Insurance (BFSI) clients. Specialized agents, per-tenant isolation, and workload-aware execution cut testing effort by 88% in an internal benchmark.

AWS Machine Learning BlogIn-site articleHow TCS built an agentic test orchestrator with Amazon Bedrock AgentCore

Cursor Projects is now multiplayer · Cursor

Cursor upgrades Projects to support multiplayer collaboration, allowing teams to work alongside their agents with shared context and agent-to-agent communication. The beta rollout is available to all Cursor users starting today.

Cursor BlogIn-site articleCursor Projects is now multiplayer · Cursor

Making the Charts in Your Documents Searchable | Unstructured

Unstructured adds an image description enrichment to its Transform parse endpoint, letting a single flag return written descriptions of every detected figure, including values read off charts, so chart-only findings become searchable.

Unstructured BlogIn-site articleMaking the Charts in Your Documents Searchable | Unstructured
Startups

Anthropic's IPO roadshow video

Anthropic's IPO roadshow video reveals previously unreleased details about its multi-chip training strategy and compute multipliers, with co-founder Tom Brown explaining how Claude runs on AWS Trainium, Google TPUs, NVIDIA GPUs, and AMD GPUs.

Simon Willison's WeblogIn-site articleAnthropic's IPO roadshow video
Research

Experience the art and legacy of Dr. Esther Mahlangu

Google Arts & Culture launched the first online retrospective of South African artist Dr. Esther Mahlangu, featuring over 45 curated stories, virtual tours, and AI-recreated early memories to preserve Ndebele culture.

Google AI BlogIn-site articleExperience the art and legacy of Dr. Esther Mahlangu

Ground Reaction Force-Aware Model Predictive Control for Underactuated Upright Physical Assistance

This paper presents GRF-MPC, a framework that integrates live ground reaction force and center of pressure measurements into a human centroidal dynamics model to optimize assistive forces via real-time quadratic programming. Deployed on the RobUST with an underactuated 4-cable configuration, it outperformed a fully actuated 8-cable baseline, reducing peak CoM error by 19.6%, tracking error by 24.1%, net mechanical work by 28.7%, and increasing voluntary excursion area by 45.9%.

arXiv RoboticsIn-site articleGround Reaction Force-Aware Model Predictive Control for Underactuated Upright Physical Assistance

PhysMoE: Fast Spatial Grounding with Single-Token Coordinate Representation in Modality-Decoupled VLMs

PhysMoE introduces a novel architecture with modality-decoupled Mixture-of-Experts, single-token coordinate encoding, and differentiable SDF supervision, addressing bottlenecks in coordinate prediction for embodied VLMs and achieving 15.7% average gain and 8.31x throughput speedup on 9 grounding benchmarks.

arXiv Computer VisionIn-site articlePhysMoE: Fast Spatial Grounding with Single-Token Coordinate Representation in Modality-Decoupled VLMs

Selective Surgical Action-Center Localization: Retaining Localization Gains through Supervised Scoring

The paper proposes a selective localization pipeline that combines heatmap and coordinate supervision with a logistic-regression scorer for surgical action-center localization. On 5,000 frames from 50 surgical sources, at an 80% acceptance quota and a tolerance of 0.10 times the image diagonal, the method retains 46.08 correct outputs per 100 inputs versus 44.20 for the baseline, a gain of 1.88. The work argues that evaluating retained correct outputs, not just the supply of correct predictions, is necessary to tell whether a selection rule preserves or erodes localization improvements.

arXiv Computer VisionIn-site articleSelective Surgical Action-Center Localization: Retaining Localization Gains through Supervised Scoring

Partition, Refine, and Detect: Preserving Novel Category Capacity in Open-Vocabulary Object Detection

The paper proposes PR-OVD, a partition-and-refine framework that explicitly reserves capacity for novel categories and strengthens category separation, achieving an average 6.5% proportional improvement on novel categories over state-of-the-art methods and integrating seamlessly into different detector architectures.

arXiv Computer VisionIn-site articlePartition, Refine, and Detect: Preserving Novel Category Capacity in Open-Vocabulary Object Detection

URL-Sim: Unified and Regulated LiDAR Simulation for High-Fidelity Reconstruction of Dynamic Driving Scenes

Researchers propose URL-Sim, a unified and regulated dynamic LiDAR simulation framework that uses a dynamic Gaussian scene graph and a regulated Gaussian densification strategy to overcome the limitations of existing 3DGS methods, achieving higher quantitative accuracy and visual quality on Waymo and KITTI-360 datasets.

arXiv Computer VisionIn-site articleURL-Sim: Unified and Regulated LiDAR Simulation for High-Fidelity Reconstruction of Dynamic Driving Scenes

Uncovering the Blind Spots of OKS-mAP: A Unified Protocol for Occlusion, Joint-Angle, and Temporal Evaluation of 2D Pose Estimators

The paper identifies three blind spots in OKS-mAP, the dominant metric for 2D human pose estimation: it does not distinguish visible from occluded keypoints, does not measure joint-angle error, and ignores temporal stability of tracking-free per-frame inference. The authors propose a three-module supplementary protocol and evaluate it on four datasets and four models, finding that aggregate mAP overestimates occluded-scene performance in every combination and that angle and temporal metrics provide independent information.

arXiv Computer VisionIn-site articleUncovering the Blind Spots of OKS-mAP: A Unified Protocol for Occlusion, Joint-Angle, and Temporal Evaluation of 2D Pose Estimators

IntentTwin: Deconfounding Candidate Support and the Oracle-to-Deployment Gap in Offline Web Behavior Prediction

This paper reveals a candidate-support confound in offline web-behavior benchmarks where inserting the demonstrated action into a finite candidate set allows support membership alone to identify the gold action. It introduces support-identifiability rate (SIR) to quantify this effect, constructs support-balanced candidate sets, and evaluates a target-free dual-path proposer on OPeRA and Mind2Web, showing that controlling support composition is essential when using offline results to reason about deployment.

arXiv Computational LinguisticsIn-site articleIntentTwin: Deconfounding Candidate Support and the Oracle-to-Deployment Gap in Offline Web Behavior Prediction

SELD: Detecting Silent Evidence Loss in Context Compression beyond the Answer

SELD is a training-free and label-free method that detects silent evidence loss in context compression by comparing a reader model's preferences over unspoken answers before and after compression, achieving 83% identification rate on a benchmark.

arXiv Computational LinguisticsIn-site articleSELD: Detecting Silent Evidence Loss in Context Compression beyond the Answer

From Relevant Posts to Attributable Spans: Explicit Symptom Evidence for Depression Questionnaire Inference

The paper introduces SREE, a pipeline that extracts first-person, symptom-specific text spans and selects a representative evidence subset via a facility location-inspired objective before evidence-gated inference, improving the precision and interpretability of depression questionnaire inference from social media.

arXiv Computational LinguisticsIn-site articleFrom Relevant Posts to Attributable Spans: Explicit Symptom Evidence for Depression Questionnaire Inference

Third Eye: Beyond Immediate Gains in Recursive Self-Improvement

The arXiv paper 'Third Eye' argues that evaluating persistent self-updates requires looking beyond immediate performance to outcomes after further native updates. Across five 3-8B backbones on math and code, immediate-selected utility is near zero or negative, and continuation reduces it further; the authors recommend treating persistent self-updates as adaptation transitions with independent scoring and a no-update baseline.

arXiv Machine LearningIn-site articleThird Eye: Beyond Immediate Gains in Recursive Self-Improvement

MADSNet: A Multi-Scale Axial-Attentive Dense Spiking Network for Spike Camera Image Reconstruction

Researchers propose MADSNet, a multi-scale axial-attentive dense spiking neural network for spike camera image reconstruction. The hierarchical spiking encoder-decoder outperforms SSIR by 4.3 dB PSNR and Spk2ImgNet by 1.94 dB on SREDS, while achieving about 15× lower latency and using only about 1/40 of the estimated energy compared with Spk2ImgNet.

arXiv Machine LearningIn-site articleMADSNet: A Multi-Scale Axial-Attentive Dense Spiking Network for Spike Camera Image Reconstruction

Exploration via Exploitation: The Blessing of Reward Diversity in Personalized Federated RL

This paper reframes reward heterogeneity in personalized federated reinforcement learning as a structural asset that catalyzes exploration rather than a hurdle to consensus. The authors propose PF-UCBVI, which achieves optimal linear speedup by decoupling shared dynamics from personalized objectives, and then introduce PF-EFVI, a purely greedy algorithm that leverages reward diversity to ensure state-action coverage. They prove that PF-EFVI attains logarithmic regret without explicit exploration bonuses under sufficient reward diversity.

arXiv Machine LearningIn-site articleExploration via Exploitation: The Blessing of Reward Diversity in Personalized Federated RL

Predictive Surprise as Self-Grounding Concept Bottleneck for Interpretable Time Series

ConceptTime uses a frozen probabilistic forecaster's predictive surprise to simultaneously locate segment boundaries, build a vocabulary of dynamical-regime concepts, and derive reliability, matching black-box accuracy while beating interpretable baselines across 17 benchmarks.

arXiv Machine LearningIn-site articlePredictive Surprise as Self-Grounding Concept Bottleneck for Interpretable Time Series

Tabular Foundation Models vs Tuned Gradient Boosting for Telecom Churn: A Profit- and Calibration-Aware Evaluation

This arXiv paper re-evaluates TabPFN models and TabICL v2 against Optuna-tuned gradient-boosted trees on five public telecom churn datasets, scoring expected maximum profit and calibration as well as ranking. TabPFN-3.5 earns significantly more profit on Cell2Cell and Iranian, but not on Orange; a larger tuning budget narrows the earlier 6.1-point PR-AUC gap to 4.0 points without closing it, and the TabPFN-3.5 weight licence, not scoring cost, is the main barrier to production.

arXiv AIIn-site articleTabular Foundation Models vs Tuned Gradient Boosting for Telecom Churn: A Profit- and Calibration-Aware Evaluation

Reinforcement Learning based control of Cancer Chemotherapy treatments

This master's thesis by Chamani Shiranthika examines reinforcement learning for cancer chemotherapy, first simulating treatment via traditional Q-learning and then proposing a model-based offline RL approach that combines conservative Q-learning with model-based offline policy optimization on real human colon cancer clinical data. Experiments indicate the Chemo RL model handles chemotherapy scheduling well and can support junior oncologists in treatment decisions. The work received the IEEE Taipei Section Best Master's Thesis Award in 2021.

arXiv AIIn-site articleReinforcement Learning based control of Cancer Chemotherapy treatments
Tools

ChatGPT Plus for veterans and servicemembers

More than 20,000 veterans and transitioning service members have claimed a free year of ChatGPT Plus. They use it to prepare for civilian careers, and OpenAI Academy and partners have hosted trainings.

OpenAI NewsIn-site articleChatGPT Plus for veterans and servicemembers
Models

Incorporating World Models in Decision Making for Next Best View Selection in Active Sensing

The paper presents a deployment adaptation method that learns a sensor-return adapter from limited calibration data while keeping a pretrained world model fixed, estimating measurement dropout to guide next-best-view selection. In real-robot experiments with a structured-light sensor, the adapted method achieved higher surface coverage with 60% fewer captures and recovered a region the nominal method repeatedly failed to measure.

arXiv RoboticsIn-site articleIncorporating World Models in Decision Making for Next Best View Selection in Active Sensing

Evidence-Grounded Iterative Reasoning for the Diagnosis of Diffuse Cystic Lung Diseases

The paper proposes DCLDs-EIR, an evidence-grounded iterative reasoning framework that couples diagnostic hypothesis refinement with multimodal evidence acquisition and verification for the differential diagnosis of diffuse cystic lung diseases, combining multiview renderings, axial CT inspection, and disease-specific knowledge graph retrieval in a closed feedback loop. Multicenter evaluation shows consistent accuracy gains over baselines and traceable differential diagnosis.

arXiv Computer VisionIn-site articleEvidence-Grounded Iterative Reasoning for the Diagnosis of Diffuse Cystic Lung Diseases

Trajectory-Regularized Distribution Distillation for One-Step Generation

The paper proposes Trajectory-Regularized Distribution Distillation (TRD), which combines distribution-level and trajectory-level supervision via a momentum-mixture policy predicted in a single student forward pass. On SANA-1.6B and Stable Diffusion 3.5, TRD outperforms consistency and distribution matching baselines in generation quality and training efficiency for one-step and few-step generation, with a trade-off in sample diversity.

arXiv Computer VisionIn-site articleTrajectory-Regularized Distribution Distillation for One-Step Generation

Attribute-Evidence Learning with Multimodal Prompt for Generalized Category Discovery

AptM-GCD introduces Attribute-Evidence Learning (AEL) and Multimodal Prompt Learning (MPL) to treat attributes as localized visual evidence, avoiding misleading global alignment and achieving state-of-the-art results on six datasets.

arXiv Computer VisionIn-site articleAttribute-Evidence Learning with Multimodal Prompt for Generalized Category Discovery

The Indic Token Tax in 2026: Measuring Tokenizer Disparity and Its Cost for the Languages of India

A new study measures tokenizer disparity across nine public tokenizers for 25 Indian languages, finding that GPT-4o's tokenizer requires 1.57 to 5.01 times more tokens than English, and 13.7 times for Santali in Ol Chiki script, directly inflating API costs.

arXiv Computational LinguisticsIn-site articleThe Indic Token Tax in 2026: Measuring Tokenizer Disparity and Its Cost for the Languages of India

khammoon: From a Linguistic Guideline to a Thai Morpheme Segmenter

Thai researchers propose segmenting Thai at the morpheme (khammoon) level rather than the word level to eliminate compounding ambiguity. They create a formal guideline with near-perfect human agreement, scale annotation with an LLM, and train a 1M-parameter convolutional BiLSTM segmenter that achieves 0.992 boundary F1 on out-of-domain text while maintaining or exceeding word-level baselines on downstream tasks.

arXiv Computational LinguisticsIn-site articlekhammoon: From a Linguistic Guideline to a Thai Morpheme Segmenter

Actionable Feedback: Improving Model Capabilities from User Edit Data on Clinical Documentation

An arXiv preprint studies how clinician edits to AI-generated clinical documentation can serve as scalable training supervision. The authors find that over 45% of edited outputs in real deployment data contain user-added information unsupported by model inputs, so naive fine-tuning risks teaching hallucination; they build a taxonomy of editing behaviors, generate semi-synthetic edit data on MIMIC-III discharge summaries, and show that progressively filtering toward grounded edits improves clinician-alignment metrics while preserving faithfulness to the input context.

arXiv Computational LinguisticsIn-site articleActionable Feedback: Improving Model Capabilities from User Edit Data on Clinical Documentation

TimeOmni-o1: Bidirectionally Coupling Time Series Foundation Models and LLMs for Multimodal Reasoning and Forecasting

TimeOmni-o1 bidirectionally couples a pretrained time series foundation model (TSFM) with an LLM: the TSFM encodes series in compressed form for the LLM to reason over with text, while the LLM's reasoning conditions the TSFM's forecast. Trained on the new MMTR dataset (12 domains, 500k+ samples), it outperforms existing models on understanding and forecasting across most domains and three open benchmarks, as well as its LLM and TSFM components used individually.

arXiv Computational LinguisticsIn-site articleTimeOmni-o1: Bidirectionally Coupling Time Series Foundation Models and LLMs for Multimodal Reasoning and Forecasting

Rethinking in Spikes: Mitigating Hallucinations in MDLMs with Step-Aware Decoding

The paper proposes StepRefine, a plug-and-play decoding strategy that detects entropy spikes during iterative denoising and switches between discrete denoising and latent refinement to mitigate hallucinations in multimodal diffusion language models (MDLMs).

arXiv Computational LinguisticsIn-site articleRethinking in Spikes: Mitigating Hallucinations in MDLMs with Step-Aware Decoding

Linear Probes to Bayesian Manifolds: Geometry of Arithmetic in Large Language Models

A new paper proposes a four-stage Bayesian pipeline to separately test decodability, shape, ownership, and causal use, offering a stricter evaluation of whether geometric structures found by linear probes actually participate in model computation. Applied to two-digit addition and multiplication in GPT-J 6B, Llama 3.1 8B, and Pythia 6.9B, results show nearly all concepts are decodable, but strongly supported shapes are ribbons or tori, 81% of shapes are inherited, and intervening on shapes retains about 61% of the effect. The authors recommend reporting probe results as verdicts on all four claims rather than a single accuracy.

arXiv Machine LearningIn-site articleLinear Probes to Bayesian Manifolds: Geometry of Arithmetic in Large Language Models

Multimodal Graph Learning: A Survey of Models, Benchmarks, and Frontiers

This survey provides an algorithm-centric and evaluation-grounded review of multimodal attributed graph learning, introducing a unified message-passing formulation across five integration levels, reviewing benchmarks and evaluation protocols, and connecting the field to AI agents and graph engineering.

arXiv Machine LearningIn-site articleMultimodal Graph Learning: A Survey of Models, Benchmarks, and Frontiers

StatPrompt-LLM: A Lightweight Statistical Refinement for Foundation-Style Time-Series Forecasting

The paper proposes StatPrompt-LLM, which augments Time-LLM prompting with dataset-level statistics and a zero-initialised autoregressive bypass, reporting average relative reductions of 40.2% MSE and 21.4% MAE under a 10% reduced-prefix setting, while full-prefix results remain dataset-dependent and computational overhead increases.

arXiv Machine LearningIn-site articleStatPrompt-LLM: A Lightweight Statistical Refinement for Foundation-Style Time-Series Forecasting

ADIndex: Layer-Adaptive Sparse Attention Indexing via Group-Level Distillation

ADIndex is an adaptive, distillation-driven indexing method for token-level sparse attention that combines layer-wise calibration, group-level distillation, and low-latency inference to significantly reduce indexer latency in long-context inference, while maintaining near-lossless task performance on DeepSeek-V3.2 and GLM-5.1.

arXiv Machine LearningIn-site articleADIndex: Layer-Adaptive Sparse Attention Indexing via Group-Level Distillation

SetPrint: Black-Box LLM Fingerprinting via Setwise Correspondence of Response Changes

SetPrint is a training-free black-box LLM fingerprinting method that jointly exploits within-set relations and between-set correspondence of response changes, outperforming existing training-free approaches on LeaFBench.

arXiv AIIn-site articleSetPrint: Black-Box LLM Fingerprinting via Setwise Correspondence of Response Changes

Graph Visualizations in Prompts for Knowledge Graph Question Answering

A controlled study evaluates when and how adding graph visualizations to prompts improves knowledge graph question answering in vision-language models, finding an average 1.85 F1 gain driven by low-complexity questions but highly sensitive to graph layout.

arXiv AIIn-site articleGraph Visualizations in Prompts for Knowledge Graph Question Answering

Tracing Psychometric Inference in Large Language Models

This study uses a cross-persona paradigm across 14 model variants from Llama 3 and Qwen 2.5 to test whether LLMs internally represent psychological constructs. Larger instruct models reproduce and amplify human cross-scale correlations; internal geometry aligns with the human matrix and correlates with behavioral amplification, but a probe-behavior gap termed readout attenuation remains, and the measured directions are not shown to mediate outputs.

arXiv AIIn-site articleTracing Psychometric Inference in Large Language Models

LandslideFM: A Foundation Model Towards Unified Landslide Perception and Understanding

Researchers propose LandslideFM, the first unified foundation model that connects landslide visual perception with contextual reasoning. It uses semi-supervised learning to adapt to multiple visual tasks and leverages predicted regions to guide a vision-language model, alongside the release of LandslideVL, the first benchmark for landslide analysis.

arXiv AIIn-site articleLandslideFM: A Foundation Model Towards Unified Landslide Perception and Understanding

Train on prefill, monitor at decode: Using sparse autoencoders for topic monitoring

arXiv paper introduces TopicScope, a framework for training causal topic classifiers on SAE features that assign a topic to every token during generation, generalizing from prefill to decode on Gemma-3 1B and 4B with strong robustness.

arXiv AIIn-site articleTrain on prefill, monitor at decode: Using sparse autoencoders for topic monitoring

EarthRecord: From Pixels to Land Change with Multimodal Agents over Four Decades of Satellite Imagery

EarthRecord is a new benchmark built on the raw record of 56 sites spanning four decades of Landsat, MODIS, VIIRS, and Sentinel-2 data, with 238 land-change questions across ten tasks that require agents to handle cloud masking and cross-sensor harmonization themselves. Evaluations of twelve multimodal agents show that leading agents complete most analyses end-to-end but falter when answers require mapping from imagery; instrument-pair controls reveal that half of correct answers follow altered satellite data, indicating agents have not yet grounded answers in real ground change.

arXiv AIIn-site articleEarthRecord: From Pixels to Land Change with Multimodal Agents over Four Decades of Satellite Imagery

llm-gemini 0.36

Simon Willison releases llm-gemini 0.36, adding support for Google's Gemini 4 Carbon, now available to all. The model is priced at $2/$10 per million tokens until December 1st, when it doubles. Willison tests it with pelican prompts and a Three.js mech animation, plus a community demo of a 3D game built with Antigravity's /swarm.

Simon Willison's WeblogIn-site articlellm-gemini 0.36

Pinecone integrated embedding now supports OpenAI's text-embedding-3 models

Pinecone now supports OpenAI's text-embedding-3-small and text-embedding-3-large models in its integrated embedding feature, letting teams run semantic search without managing an OpenAI account or API key. Tokens are billed on the Pinecone invoice at OpenAI's list price.

Pinecone BlogIn-site articlePinecone integrated embedding now supports OpenAI's text-embedding-3 models
Policy

Latent Barrier Steering: Hierarchical Safety for Generative Planning

This paper introduces Latent Barrier Steering (LBS), a hierarchical semantic-to-path framework for safe flow-matching-based generative planning. LBS first steers the generator's behavior latent toward rollouts with larger safety margin, then applies a path-space control barrier function quadratic program only for residual violations, improving goal reaching and reducing correction burden while preserving safety.

arXiv RoboticsIn-site articleLatent Barrier Steering: Hierarchical Safety for Generative Planning

AI Governance Should Prioritize Control and Knowledge Boundaries Over Limiting Intelligence

A NeurIPS 2026 position paper introduces the 'competence circuit' framework, decomposing an AI agent's real-world impact into intelligence-mediated discovery, procedural know-how, and direct control. Through simulation analysis, the paper argues that intelligence-driven speedup is capped by the ratio of competing intelligence levels, and that a large speedup multiplier need not imply rapid progress. Consequently, AI governance should prioritize restricting control and enforcing knowledge boundaries rather than limiting intelligence alone, and suggests intentionally 'putting physics in the loop' in system architectures.

arXiv AIIn-site articleAI Governance Should Prioritize Control and Knowledge Boundaries Over Limiting Intelligence

Amazon Bedrock tokenomics data lake

This post builds a centralized observability data lake that aggregates Amazon Bedrock invocation logs from every linked account into one queryable table, so you can see prompts, token counts, model choice, and per-invocation cost across your organization.

AWS Machine Learning BlogIn-site articleAmazon Bedrock tokenomics data lake
Robotics

RoboACE: Adversarial Co-Evolution for Robot Self-Improvement

RoboACE jointly evolves a curriculum harness and an execution harness around frozen foundation models, reaching 74.0% success on LIBERO-PRO, a 15.0-point gain over the initial deployment.

arXiv RoboticsIn-site articleRoboACE: Adversarial Co-Evolution for Robot Self-Improvement

CAFE: Closed-Loop Autonomous Task Formulation and Execution without Instance-Specific Instructions

This paper introduces CAFE, a framework that enables scene-driven robot autonomy by formulating and executing tasks from observations under predefined categories (cleaning, safety, storage/organization) without instance-specific goals or action sequences. CAFE maintains online memory of objects, functional zones, and execution states, and uses a rule-guided Trigger-LLM with Draft, Complete, and Audit LLMs to propose goals, resolve details, and check dependencies. Outcomes update memory for correction and discovery. Evaluated in 12 AI2-THOR scenes, Tri-LLM achieves higher GCR scores than baselines across three backbones and task-complexity tiers. Controls validate components, and LimoCobot traces illustrate storage, cleanup, and recovery from a failed grasp.

arXiv RoboticsIn-site articleCAFE: Closed-Loop Autonomous Task Formulation and Execution without Instance-Specific Instructions

Maze Runner: Autonomous Maze Navigation with Line-Based Localization and Control Barrier Functions

Maze Runner is an open-source C++/ROS 2 autonomy stack that enables end-to-end maze navigation on a simulated Husarion ROSbot XL with a 2D LiDAR. It extracts wall lines with RANSAC and total least squares, supports EKF localization and EKF SLAM, plans with A* or Hybrid A*, and controls with NMPC plus discrete-time control barrier functions. In simulation, the robot completes a 135 m route with 1.0 cm localization RMSE (known map) and 24 cm (SLAM), planning in 6 ms and NMPC solving in 11.7 ms on average within an 80 ms budget.

arXiv RoboticsIn-site articleMaze Runner: Autonomous Maze Navigation with Line-Based Localization and Control Barrier Functions

How AI Helps Recyclers Turn Mixed Waste Into Reusable Material

At COP31, with a focus on zero waste, NVIDIA highlights four startups using vision AI and robotics to sort mixed waste more precisely and keep valuable materials in circulation.

NVIDIA BlogIn-site articleHow AI Helps Recyclers Turn Mixed Waste Into Reusable Material