AI news archive.
10,661 articles filed under Research, newest first, from the outlets listed on the sources page.
- arXiv cs.AIresearchVPTwin: Real-Sim-Real Video Prediction for Robotic Manipulation Planning
arXiv:2609.33104v1 Announce Type: cross Abstract: While action-conditioned video prediction provides an intuitive world model for robotics, purely data-driven predictors often suffer from compounding errors and physicall…
- arXiv cs.AIresearchToward Comprehensive 3D Grounding: Orientation Grounding through Vision-Language Models
arXiv:2609.33109v1 Announce Type: cross Abstract: Grounding is a core capability of spatial vision-language models, yet most existing work focuses only on where a referred object is. Many 3D tasks also require knowing ho…
- arXiv cs.AIresearchD-JEPA: Design-Recoverable JEPA Representation with Swappable Physics Decoders
arXiv:2609.33110v1 Announce Type: cross Abstract: Joint-Embedding Predictive Architectures (JEPAs) provide a framework for learning compact representations without directly reconstructing high-dimensional observations. H…
- arXiv cs.AIresearchMedRouter: Demystifying Knowledge Differences Across Medical LLMs for Routing-Based Reasoning
arXiv:2609.33119v1 Announce Type: cross Abstract: Medical question answering spans diverse specialties and modalities, and individual medical large language models (LLMs) exhibit distinct strengths across tasks and domai…
- arXiv cs.AIresearchPolicy Plasticity Matters in Offline-to-Online Reinforcement Learning: Refitting Offline Policies for Online Adaptation
arXiv:2609.33127v1 Announce Type: cross Abstract: Offline-to-Online Reinforcement Learning (O2O RL) has emerged as a practical paradigm that pre-trains the policy using static offline datasets and subsequently adapts the…
- arXiv cs.AIresearchFlow-Matching-Based Protein Structure Tokenizer Made Efficient and Easy
arXiv:2609.33129v1 Announce Type: cross Abstract: As the bridge between protein modality and discrete modeling, protein structure tokenization still largely relies on heavily engineered training objectives tailored to sp…
- arXiv cs.AIresearchQuantum Monte Carlo Tree Search with Fixed Confidence
arXiv:2609.33132v1 Announce Type: cross Abstract: Recent advances in quantum computing are opening new opportunities for computationally intensive decision problems. This paper studies how quantum computing can improve M…
- arXiv cs.AIresearchBeyond the Training Horizon: Mechanisms and Limits of Length Generalization in Looped Transformers
arXiv:2609.33144v1 Announce Type: cross Abstract: Looped Transformers can generalize to reasoning chains longer than those encountered during training, but the computations enabling this behavior and limiting its extent…
- arXiv cs.AIresearchFOCUS: Benchmarking Retinal Model Generalization from Foundation Vision Encoders to Multimodal LLMs
arXiv:2609.33158v2 Announce Type: replace-cross Abstract: Progress in AI-based retinal image analysis has accelerated with the adoption of foundation models, yet evaluating their real-world reliability remains challengin…
- arXiv cs.AIresearchDynamic Manipulation with World-Action Models via Counterfactual Planning
arXiv:2609.33172v1 Announce Type: cross Abstract: World-Action models (WAMs) trained on static demonstrations often fail to manipulate moving targets even when they possess the required manipulation skills. We attribute…
- arXiv cs.AIresearchScoring the Wrong Question: Readout Failures in Constrained-Option Evaluation
arXiv:2609.33179v1 Announce Type: cross Abstract: Constrained-option scoring reads a model's probabilities for a fixed set of permitted answers, so it returns a score even when the model is about to write something else.…
- arXiv cs.AIresearchSMORE: Stability-Promoting Mesh-Agnostic Model Reduction for Time-Dependent PDEs
arXiv:2609.33205v1 Announce Type: cross Abstract: High-fidelity simulations of time-dependent partial differential equations (PDEs) are computationally expensive, motivating data-driven reduced-order surrogates for many-…
- arXiv cs.AIresearchBeyond Calibration: Do a Typed-Decision Model's Probabilities Obey the Probability Axioms?
arXiv:2609.33209v1 Announce Type: cross Abstract: Typed-decision models such as TypeSafe's Jev answer a declared yes/no or multiple-choice question about a state with a probability instead of text, and their evaluations…
- arXiv cs.AIresearchScope-WM: Scoped Computation for Efficient Visual World Models
arXiv:2609.33218v1 Announce Type: cross Abstract: Visual world models enable robotic planning by predicting future observations, but dense latent-state propagation and sample-intensive trajectory optimization incur high…
- arXiv cs.AIresearchWhen Do Models Admit They Are Wrong? Failure Disclosure Is Unstable Under Reinforcement Learning
arXiv:2609.33220v1 Announce Type: cross Abstract: Outcome-based reinforcement learning can produce models with similar task performance but very different ways of communicating about their mistakes. We study failure disc…
- arXiv cs.AIresearchVGGT-Diff: Visual Geometry Meets Diffusion for Sparse-View Novel View Synthesis
arXiv:2609.33253v2 Announce Type: replace-cross Abstract: We present VGGT-Diff, a geometry-routed multi-view diffusion model for sparse-view novel view synthesis. Existing novel view synthesis (NVS) methods face a fundam…
- arXiv cs.AIresearchBERT4DTI : BERT-based Model for Predicting Drug-Protein Interactions
arXiv:2609.33254v1 Announce Type: cross Abstract: Understanding how drugs interact with protein targets is fundamental to drug discovery, drug repurposing and the early identification of promising therapeutic candidates…
- arXiv cs.AIresearchSCISSOR: Score-Conditioned Instrument Source Separation for Orchestral Recordings
arXiv:2609.33265v1 Announce Type: cross Abstract: Orchestral separation recovers instrument sections from mixtures in which shared pitches, harmonics, and timbres obscure source identity. An aligned score provides instru…
- arXiv cs.AIresearchDomain Generalization under Sampling Pattern Shifts in Irregular Time Series
arXiv:2609.33279v1 Announce Type: cross Abstract: Irregularly sampled multivariate time series (ISMTS) are prevalent in real-world applications, where both observation times and available measurements can vary substantia…
- arXiv cs.AIresearchHesitation-Aware On-Policy Distillation for Diffusion Language Models
arXiv:2609.33301v1 Announce Type: cross Abstract: Diffusion large language models (dLLMs) generate text by iterative unmasking. At each denoising step, a dLLM proposes a token at every masked position, but the decoder co…
- arXiv cs.AIresearchBITS: Rethinking Fair and Comprehensive Evaluation for Irregular Time Series Forecasting
arXiv:2609.33303v1 Announce Type: cross Abstract: Despite recent progress in irregular time series forecasting, the field still lacks a unified benchmark for fair and comprehensive evaluation. Existing evaluations are of…
- arXiv cs.AIresearchZeroGAR: Benchmarking the Adversarial Robustness of Zero-Shot Graph Models
arXiv:2609.33314v1 Announce Type: cross Abstract: Zero-shot graph models (ZGMs), which learn transferable knowledge from source graphs and directly apply to unseen target graphs without any adaptation, have achieved prom…
- arXiv cs.AIresearchBeyond Conservatism: Recoverability-Conditioned Exploration for Model-Based Imitation Learning
arXiv:2609.33336v1 Announce Type: cross Abstract: Model-based imitation learning (MBIL) improves real-environment interaction efficiency by optimizing policies on imagined rollouts from a learned world model. However, th…
- arXiv cs.AIresearchCHI: A Composite Hallucination Index Unifying Entity, Relation, and Quantity Dimensions for Summarization Evaluation
arXiv:2609.33343v1 Announce Type: cross Abstract: Faithfulness evaluation of abstractive summaries remains an open challenge, with existing metrics addressing only isolated hallucination types: factual entity errors, rel…
- arXiv cs.AIresearchKoopCell: Koopman-Based Generative Model for Learning Single-Cell Dynamics from Distribution Snapshots
arXiv:2609.33350v1 Announce Type: cross Abstract: Learning population dynamics from temporally sparse, unpaired distribution snapshots is a fundamental challenge in developmental biology. Recent approaches based on neura…
- arXiv cs.AIresearchWhat Survives the Codec Shift: Pooled No-Vocals Residuals for Speech Deepfake Detection
arXiv:2609.33375v1 Announce Type: cross Abstract: The transition from vocoder-based to neural-codec speech synthesis makes generalization more difficult for speech deepfake detectors, particularly those relying on speech…
- arXiv cs.AIresearchFrom Position Risks to Block Survival: Faster Generation for Diffusion Language Models
arXiv:2609.33390v1 Announce Type: cross Abstract: Diffusion language models (DLMs) can accelerate generation by predicting multiple tokens in parallel, but there is a mismatch between how these tokens are predicted and h…
- arXiv cs.AIresearchAutoHGNN: Robust and Efficient Neural Architecture Search for Hypergraph Neural Networks
arXiv:2609.33392v1 Announce Type: cross Abstract: Hypergraph neural networks have achieved significant success in recent years. However, manual architecture crafting is labor-intensive and often fails to capture complex,…
- arXiv cs.AIresearchDecoupling Token Roles in Autoregressive Pretraining
arXiv:2609.33405v1 Announce Type: cross Abstract: Autoregressive pretraining increasingly draws on heterogeneous data, making it important to understand how a model learns from an individual token. The next-token predict…
- arXiv cs.AIresearchResolving State-Representation Mismatch: State-Space Visual Reasoning for Open-Loop VLA Planning
arXiv:2609.33412v1 Announce Type: cross Abstract: Despite rapid progress in vision-language-action (VLA) models, existing reasoning paradigms still face a fundamental \emph{state-representation mismatch} in open-loop pla…
- arXiv cs.AIresearchTTRSD: Test-Time Reinforcement Learning with Self-Distillation for Vision-Language Models
arXiv:2609.33414v1 Announce Type: cross Abstract: Test-time reinforcement learning enables vision-language models (VLMs) to adapt using unlabeled inputs. However, repeated sampling under fixed visual conditions can reinf…
- arXiv cs.AIresearchInvestigating the Effect of k-NN Preprocessing on Developing Graph Neural Networks: A Fairness-Based Perspective
arXiv:2609.33416v1 Announce Type: cross Abstract: In this paper, a methodology to design fair graph convolutional neural networks (GCNs) is developed and tested over several application data sets. The graphs that are use…
- arXiv cs.AIresearchExplainable Deep Learning of Resting-State Functional Connectomes Reveals Network Biomarkers of Adolescent Intelligence
arXiv:2609.33422v1 Announce Type: cross Abstract: Mapping resting-state brain organization to individual differences in cognitive ability remains a major challenge in population neuroinformatics. Although deep learning e…
- arXiv cs.AIresearchA Light Bilevel Refinement Aligns Self-Supervised Representations for Stronger Task-Specific Learning
arXiv:2609.33424v1 Announce Type: cross Abstract: Self-supervised pretraining learns representations that are broadly transferable across downstream tasks, yet direct fine-tuning can be suboptimal due to misalignment bet…
- arXiv cs.AIresearchProtected Cores Are Not Enough: Certifying AI-Proposed Revisions of Temporal Specifications
arXiv:2609.33461v1 Announce Type: cross Abstract: Runtime monitoring traditionally evaluates a specification that is fixed before execution or externally modified when requirements change. In learning-enabled and data-in…
- arXiv cs.AIresearchA Cheap Verifier is Good Enough: LLM Post-training is Robust to Erroneous Rewards
arXiv:2609.33467v1 Announce Type: cross Abstract: When post-training large language models on tasks with semi-verifiable rewards, there are many factors (training steps, base model size, training order, data quality, ver…
- arXiv cs.AIresearchChameleon: Dynamic Format Adapter for Efficient Diffusion
arXiv:2609.33496v1 Announce Type: cross Abstract: Post-training quantization (PTQ) is the standard way to run modern diffusion models on memory-constrained accelerators, yet every existing diffusion PTQ scheme fixes the…
- arXiv cs.AIresearchSource Anchoring for Physical Consistency in Flow Matching Models
arXiv:2609.33510v1 Announce Type: cross Abstract: Deep generative models are used to solve partial differential equations and model distributions of physical system states, but ensuring that the generated samples satisfy…
- arXiv cs.AIresearchE-CONAN (Entailment, CONtradition And Neutral) Diagnostics Dataset Investigating Linguistic Phenomena in Arabic Natural Language Understanding
arXiv:2609.33530v1 Announce Type: cross Abstract: Natural Language Understanding (NLU) plays a crucial role in various applications, yet its performance suffers from weaknesses in handling the complexities of human langu…
- arXiv cs.AIresearchTerMeZO: Ternary Sparse Zeroth-Order Optimization for Fine-tuning BitNet Models at the Edge
arXiv:2609.33548v1 Announce Type: cross Abstract: Fine-tuning anguage models (LLMs) with first-order optimizers requires a memory several times larger than that required for inference. Memory-efficient zeroth-order optim…
- arXiv cs.AIresearchOOD Generalization as a Bifurcation Problem
arXiv:2609.33562v1 Announce Type: cross Abstract: Systematic out-of-distribution (OOD) generation remains a critical bottleneck for continuous-time generative models. While standard joint classifier-free guidance (CFG) r…
- arXiv cs.AIresearchCompressing Value Predictions for Learning-Augmented Metrical Task Systems
arXiv:2609.33580v1 Announce Type: cross Abstract: Learning-augmented algorithms for metrical task systems (MTS) can exploit predictions of canonical dual values, but existing formulations typically require a prediction f…
- arXiv cs.AIresearchViCoR: Reliable Molecular Structure Extraction via Spatially Aligned Verification and Executable Revision
arXiv:2609.33603v1 Announce Type: cross Abstract: Reliable optical chemical structure recognition (OCSR) is essential for building high-quality chemical data from scientific literature, yet even small recognition errors…
- arXiv cs.AIresearchLearning Transferable Reaction Mechanisms from Visual Chemical Knowledge
arXiv:2609.33608v1 Announce Type: cross Abstract: Reaction mechanisms describe the step-by-step transformations underlying chemical reactions and are central to reaction analysis and synthesis. Learning-based models have…
- arXiv cs.AIresearchQuizzing the Translation: A Prover-Grounded Evaluation Metric for NL$\rightarrow$FOL
arXiv:2609.33612v1 Announce Type: cross Abstract: A standard pipeline for symbolic reasoning over natural-language problems translates them into first-order logic and invokes a theorem prover. The translation step is the…
- arXiv cs.AIresearchSpatialSpeak: QA-Native Reconstruction with Local and Global Context for Spatial Chain-of-Thought Reasoning
arXiv:2609.33616v1 Announce Type: cross Abstract: Vision-language models (VLMs) can benefit from geometric priors for multi-view spatial reasoning, yet answer-only training does not directly supervise the intermediate ge…
- arXiv cs.AIresearchLearning Dynamics of Continual Learning: A Unified View of Data Attribution, Forgetting, and Plasticity Loss
arXiv:2609.33620v1 Announce Type: cross Abstract: Modern language models are likely to be updated throughout their lifetime rather than trained once and frozen. Each update therefore participates in a recurring cycle: de…
- arXiv cs.AIresearchSafety Reconstructed: Generative Modeling via Masked Diffusion Builds Strong Safety Guardrails
arXiv:2609.33634v1 Announce Type: cross Abstract: Guard models are the last line of defense between a language model and a harmful output, yet their training objective is surprisingly narrow. Existing guards learn to pre…
- arXiv cs.AIresearchLearning to Learn from Context: Synthetic Training from Perturbed Public Documents
arXiv:2609.33642v1 Announce Type: cross Abstract: Real-world tasks often require large language models (LLMs) to learn from complex task-specific context rather than pretrained parametric knowledge. This capability remai…
- arXiv cs.AIresearchFrom Distributions to Stochastic Processes: Neural Approximation of Measure-Valued Maps
arXiv:2609.33649v1 Announce Type: cross Abstract: Learning mappings between probability distributions arises naturally when inputs and outputs are represented by populations of samples rather than individual observations…