Search: “routing”

Certified Interface Aliases: Exact Collisions in Vision-Language Preprocessing, and When They Exist

arXiv:2609.33003v1 Announce Type: new Abstract: Vision-language verifiers and routers must distinguish errors repairable by more reasoning from those caused by visual evidence never reaching the langu

Content type

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

Safety

COGNIT-Guard: Calibrated Standalone Direct-Decision Guardrails with Heterogeneous CPU-NPU Confidence Cascading under Explicit Latency and False-Positive Constraints

DGX agent

arXiv:2609.33671v1 Announce Type: cross Abstract: When must a foundation-model safety gateway generate tokens, and when should it directly output a calibrated decision? We study calibrated standalone

safetyarxiv-cs-lg

29 Sept 2026

Collaborative Principle Evolution via Evidence Transfer for Scientific Discovery

arXiv:2609.35315v1 Announce Type: new Abstract: Large Language Model (LLM)-based agents promise to automate scientific discovery, yet exploring the vast hypothesis space remains costly. Existing princ

Local Ai

Cross-Rollout Bellman Closure for Long-Horizon Agentic Reinforcement Learning

arXiv:2609.35082v1 Announce Type: new Abstract: Group-based reinforcement learning such as GRPO trains LLM agents by comparing rollouts sampled for each task, without a learned critic. In long-horizon

local-aiarxiv-cs-lg

Agents

DEALS: Decentralized Expertise-Aware Load Serving for Multi-Agent LLM Systems

arXiv:2609.33768v1 Announce Type: new Abstract: Multi-agent systems (MAS) have recently emerged as an effective approach for coordinating large language model (LLM)-based agents to solve complex tasks

agentsarxiv-cs-lg

Disentangle and Drop: Robust Universal Removal of Image Watermarks via Reconstructive Grayscale Residual Decomposition

arXiv:2609.31693v1 Announce Type: new Abstract: Invisible image watermarks are commonly evaluated against benign postprocessing operations such as compression, resizing, blur, and color changes. These

Do We Really Need KL Divergence for On-Policy Distillation of Large Language Models?

arXiv:2609.33791v1 Announce Type: new Abstract: Since the advent of knowledge distillation, KL divergence has been the standard loss in distillation. Recently, on-policy distillation (OPD) has emerged

Tutorials

Downstream-Aware Context Selection for Online In-Context Reinforcement Learning

arXiv:2609.33166v1 Announce Type: new Abstract: In-context reinforcement learning (ICRL) enables large language model agents to adapt to new environments using their interaction history without updati

tutorialsarxiv-cs-lg

Evaluating Hierarchy-Aware Deep Learning for the Recognition of Tironian Notes

arXiv:2609.35277v1 Announce Type: new Abstract: Tironian notes are generally regarded as the first Latin shorthand system and are notable for their large, fine-grained symbol inventory. Their high vis

From Static to Dynamic: On-Policy Distillation from Image to Video Diffusion Models

arXiv:2609.34371v1 Announce Type: new Abstract: On-policy distillation (OPD) specializes pretrained video diffusion models through teacher supervision along the student's own generation trajectory. Al

Local Ai

HamiFormer: Dual-Expert Diffusion Fields with Affine Symplectic Maps

arXiv:2609.32838v1 Announce Type: new Abstract: Predicting smooth dynamics and collisions requires modeling continuous evolution and abrupt state changes. We introduce HamiFormer, a dual-expert diffus

local-aiarxiv-cs-lg

HARMONIA: Interpretable Graph Learning through Mixtures of Neural Bases

arXiv:2609.33972v1 Announce Type: cross Abstract: Existing interpretable graph additive models still face limitations in either computational scalability or modeling flexibility. In terms of structura

Tutorials

How Far Are We from Removing the Visual Encoder? Scaling Laws for Encoder-Free Multimodal Pretraining

arXiv:2609.35457v1 Announce Type: new Abstract: Most modern multimodal large language models (MLLMs) build on a pretrained visual encoder that provides a strong visual prior. Encoder-free MLLMs instea

tutorialsarxiv-cs-cv

K-OPSD: Verifiable On-Policy Self-Distillation for Post-Training Vision-Language Models on AEC Drawings

arXiv:2609.34082v1 Announce Type: cross Abstract: Interpreting architecture, engineering, and construction (AEC) drawings is hard for general Multimodal Large Language Models (MLLMs) and vision-langua

Agents

Kafila: Serving Large Language Models on a Trusted Set of Heterogeneous Commodity Machines

arXiv:2609.34045v1 Announce Type: cross Abstract: Between them, the members of a research group or a circle of friends own several consumer computers, none large enough to run a capable large language

agentsarxiv-cs-lg

KV-Lingo: Learning KV-Cache Translators with Distillation

arXiv:2609.32610v1 Announce Type: cross Abstract: Large language models represent context with a key-value (KV) cache. Caches are model-specific: for the same text, models with different architectures

Less Sycophancy, Stronger Refusal? Lessons for AI Safety from Mechanistic Interpretability

arXiv:2609.35544v1 Announce Type: cross Abstract: Reliable refusal of harmful requests is essential to the safe deployment of language models. Because excessive eagerness to please users may undermine

Let CSP Be Your ANCHOR: Adaptive Crystal Search over Frozen Structure Priors

arXiv:2609.33407v1 Announce Type: cross Abstract: De novo crystal generation (DNG) models decide where to search in composition space and how to generate structures with one set of weights. We argue t

LRC-JEPA: Disentangling Dynamics and Residual Context for Efficient World Models

arXiv:2609.34375v1 Announce Type: new Abstract: Compact JEPA world models enable efficient latent-space planning, but low-dimensional representation trained under reward-free self-supervision must enc

Maintaining Benchmarks Against Increasingly Capable Agents: Detection and Remediation of Unearned Passes

arXiv:2609.34262v1 Announce Type: cross Abstract: Agentic benchmarks guide model selection and training. Yet an agent can pass a task without demonstrating the intended capability. Such outcomes const

Hardware

MarsLab: A Martian Rover Simulator for Planetary Rover Autonomous Navigation

arXiv:2609.34702v1 Announce Type: new Abstract: Future Mars missions will require rover autonomy that can operate across unstructured terrain, changing illumination, atmospheric dust, and limited comm

hardwarearxiv-cs-ro

Mend the Measurement Gap: Latent User Preference Modeling for Short-Form Video Recommendation

arXiv:2609.32839v1 Announce Type: cross Abstract: Recommender systems rely heavily on heterogeneous behavioral feedback to infer user preference. Although abundant, these signals are imperfect measure

Tutorials

Metro-WM: Long-Horizon Latent Planning with Realisable Sub-Goals

arXiv:2609.31868v1 Announce Type: cross Abstract: Model-predictive control with Joint-Embedding Predictive Architectures (JEPAs) provides a strong zero-shot goal-reaching planner, but it is only effec

tutorialsarxiv-cs-ro

Native Association: Confidence-Aware Human Perception in the Wild with a Foundation VLM

arXiv:2609.33450v1 Announce Type: new Abstract: Extracting who is where, on which team, wearing which number from a broadcast frame is typically done by stitching a detector, an OCR engine, and classi

NutriVision: Ingredient-Conditioned Fusion and Prediction for Single-Image Food Nutrition Estimation

arXiv:2609.33076v1 Announce Type: new Abstract: Nutrition estimation is a fundamental task in consumer diet tracking, clinical dietetics, chronic disease management, sports and hospital nutrition, and

OMP-MoE: Efficient Expert Pruning for Mixture-of-Experts LLMs via Orthogonal Matching Pursuit

arXiv:2609.31631v1 Announce Type: new Abstract: Mixture-of-Experts (MoE) models enable efficient scaling of large language models but face critical deployment challenges due to massive memory requirem

On the Pitfalls of Verbalized Confidence Priors for Calibrating Large Reasoning Models

arXiv:2609.32470v1 Announce Type: new Abstract: Large reasoning models (LRMs) often suffer from overconfidence when expressing their uncertainty. Confidence-aware reinforcement learning (RL) offers a

One Perception, All Maneuvers: Directional Traffic Signal Understanding for Maneuver-Level Signal Intent Prediction

arXiv:2609.32316v1 Announce Type: new Abstract: Traffic lights are a key regulatory signal for autonomous driving at urban intersections, yet existing traffic signal perception is still predominantly

Oracle Gaps in Reliability Coverage: Sampling Noise or Policy Specialization?

arXiv:2609.32996v1 Announce Type: new Abstract: Policies trained from the same base model can appear to solve different problems. An oracle that chooses the best policy for each problem may therefore

PARSEE-VAD: Efficient Training-Free Online Video Anomaly Detection via Proposition-Aware Reasoning and Streaming Evidence Escalation

arXiv:2609.33236v1 Announce Type: new Abstract: Training-free online video anomaly detection (VAD) with frozen multimodal language models faces two coupled challenges: extracting reliable current-wind

PEEL-DDPM: Physics-Enabled Evidential Learning for the Denoising Diffusion Probabilistic Model

arXiv:2609.31742v1 Announce Type: new Abstract: Normal-inverse-gamma (NIG) regression is not identifiable from its marginal Student-t likelihood: three combinations of four NIG parameters are determin

PlanGuard: A Guardrail for Multi-Step Plan Safety in Embodied Agents

arXiv:2609.32801v1 Announce Type: cross Abstract: Embodied task planners may produce multi-step plans whose subtask dependencies and interactions with the environment create physical risks during exec

Agents

PrefillShare: A Shared Prefill Module for KV Reuse in Multi-LLM Disaggregated Serving

arXiv:2602.12029v2 Announce Type: replace Abstract: Multi-agent systems increasingly orchestrate multiple specialized language models to solve complex real-world problems, often invoking them over a s

agentsarxiv-cs-lg

Reduced Cartesian Kinetostatics for Tendon-Driven Continuum Robots: Residual-Stabilized Full-Shape Propagation

arXiv:2609.31771v1 Announce Type: new Abstract: Many planning and control tasks for tendon-driven continuum robots (TDCRs) require the complete Cartesian backbone geometry. We present a reduced Cartes

safetyarxiv-cs-ro

Relational Compression: A Framework for Relational Fidelity in Constrained Representations

arXiv:2609.31816v1 Announce Type: new Abstract: What should a compressed representation preserve when the information of interest lies in relationships among elements rather than in the elements thems

Repair Before You Fuse: Frozen-Host Adaptation for Corrupted-but-Present Sensors

arXiv:2609.35264v1 Announce Type: new Abstract: Camera-LiDAR detectors can continue to consume unreliable features even when both sensors remain present, synchronized, and calibrated. We introduce Bou

researcharxiv-cs-ro

Rethinking Contextualization by Reinterpreting Attention Head Channels

arXiv:2609.33851v1 Announce Type: new Abstract: Contextualization, the core operation of language modeling, transmits information across words to build sentence-specific word representations. Prior wo

Revisiting Risky Tackle Detection with Vision Transformers

arXiv:2609.35562v1 Announce Type: new Abstract: This paper is a Track 2 reproducibility companion to an ICPR 2026 study on risky tackle detection in American football prac- tice videos. The original w

Safe Score Matching: Diffusion Policies with Hamilton-Jacobi Reachability for Online Safe Reinforcement Learning

arXiv:2609.33337v1 Announce Type: new Abstract: Online safe reinforcement learning (RL) seeks policies that maximize reward while satisfying safety constraints. A popular line of research in safe RL r

Safety-Constrained Cascade Inference for Robust Malaria Cell Classification Under Field Corruptions

arXiv:2609.33005v1 Announce Type: new Abstract: Automated malaria diagnosis from thin blood-smear microscopy could meaningfully reduce the burden on under-resourced laboratories, but a model that maxi

SEES: A Self-Evolving Embodied System via Failure-Guided VLA Policy Adaptation

arXiv:2609.32698v1 Announce Type: new Abstract: Recent vision-language-action (VLA) policies demonstrate promising generalization across diverse short-horizon tasks. However, they remain unreliable on

safetyarxiv-cs-ro

Simulation-Based Quantum System Inference with Neural Posterior Estimation

arXiv:2609.34995v1 Announce Type: cross Abstract: Models of quantum systems faithfully map system parameters to observations, but the inverse problem of parameter inference from measurement data prese

SolveEdit: Benchmarking Visual Problem Solving in Generative Models

arXiv:2609.35504v1 Announce Type: new Abstract: Machine intelligence is often evaluated through abstract reasoning problems, yet many real-world problems are visual, such as arranging objects, repairi

Source-Lifted Flow Matching for Intervenable Multimodal Imitation

model-releasesarxiv-cs-ro

arXiv:2607.10206v2 Announce Type: replace Abstract: Flow-matching policies are promising for imitation learning because they model complex multimodal action distributions. However, their stochasticity

SpatialSkill: Self-Evolving Skills for Cross-View Spatial Reasoning

arXiv:2609.34124v1 Announce Type: new Abstract: Cross-view spatial reasoning requires a model to align different viewpoints into a coherent spatial representation, yet this ability remains challenging

Local Ai

Structured Residual Connectivity Matters for Diffusion Transformers

arXiv:2609.33203v1 Announce Type: new Abstract: Diffusion Transformers (DiTs) have established themselves as a scalable backbone for high-fidelity image synthesis. However, unlike U-Net based diffusio

local-aiarxiv-cs-cv

The Judge Is Not Its Twin: Post-training makes a model's writing more predictable but barely moves its taste, as a judge, toward predictable writing

arXiv:2609.32196v1 Announce Type: cross Abstract: Language models are now routinely graded by other language models. If post-training makes a model's own writing more predictable, it may also teach th

The Key Handoff: Retrieval in Hybrid Language Models

arXiv:2609.32942v1 Announce Type: cross Abstract: A two-hop question makes a language model retrieve twice: once to produce a bridge entity, and once to retrieve with it. Transformers resolve that ent