Search: “pretraining”

Representation-Guided Discrete Molecular Graph Retrosynthesis

arXiv:2605.24428v1 Announce Type: new Abstract: Stochastic process-based molecular graph generators have become the state of the art for template-free single-step retrosynthesis. However, these models

Content type

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

SliceWorld: A Predictive and Controllable World-State Model for CT Report Generation

DGX agent

arXiv:2605.24371v1 Announce Type: cross Abstract: CT report generation (CTRG) requires models to summarize three-dimensional anatomical context and pathological findings from hundreds of axial slices.

safetyarxiv-cs-cl

26 May 2026

Test-Time Self-Adaptive Conditioning for Stable Audio-Driven Talking-Head Generation

arXiv:2605.25488v1 Announce Type: cross Abstract: Audio-driven talking-head generation has achieved remarkable progress with recent models such as AniTalker, FLOAT, and Sonic. Despite their success, m

Thaka at KSAA-2026 Task 2: Regularized Fine-Tuning for Arabic Speech Diacritization

arXiv:2605.25928v1 Announce Type: new Abstract: We describe the winning system for Task 2 of the KSAA-2026 Shared Task on Arabic Speech Dictation with Automatic Diacritization. The task requires produ

researcharxiv-cs-cl

Tiny Brains, Giant Impact: Uncovering the Keystone Neurons of LLM with Just a Few Prompts

arXiv:2605.24846v1 Announce Type: cross Abstract: Large language models (LLMs) display strong comprehensive abilities, yet the internal mechanisms that support these behaviors remain insufficiently un

TinyFormer: Preserving Tiny Objects in YOLO-DETRHybridReal-time Detectors

arXiv:2605.25046v1 Announce Type: cross Abstract: YOLO-series and DETR-based detectors struggle with tiny-object detection. YOLO-style models benefit from efficient dense prediction, but their large-s

researcharxiv-cs-ai

Tutorials

UNATE: UNsupervised ATomic Embedding for crystal structures property prediction

arXiv:2605.25866v1 Announce Type: new Abstract: Accurately predicting crystal properties is critical for accelerating materials discovery, but it is often limited by scarce labeled data and costly the

tutorialsarxiv-cs-lg

ViroBench: Benchmarking Nucleotide Foundation Models on Viral Genomics Tasks

arXiv:2605.25388v1 Announce Type: new Abstract: Nucleotide sequences constitute the fundamental genetic basis of biological systems, rendering viral genomic analysis critical for biomedical advancemen

When In-Distribution Gains Fail: Evaluating Weak-to-Strong Reward Models under Preference Shift

arXiv:2605.25629v1 Announce Type: new Abstract: Weak-to-strong (W2S) generalization is a promising framework for scalable oversight, yet existing evaluations often test students under matched train--t

researcharxiv-cs-cl

Tutorials

A mathematical theory of balancing relational generalization and memorization

arXiv:2605.22972v1 Announce Type: cross Abstract: Humans, animals, and modern machine learning models exhibit impressive abilities to learn complex behaviors and generalize these behaviors to unseen s

tutorialsarxiv-cs-ai

Beyond VLM-Based Rewards: Diffusion-Native Latent Reward Modeling

arXiv:2602.11146v2 Announce Type: replace-cross Abstract: Preference optimization for diffusion and flow-matching models relies on reward functions that are both discriminatively robust and computatio

safetyarxiv-cs-ai

BURMESE-SAN: Burmese NLP Benchmark for Evaluating Large Language Models

model-releasesarxiv-cs-cl

arXiv:2602.18788v3 Announce Type: replace Abstract: We introduce BURMESE-SAN, the first holistic benchmark that systematically evaluates large language models (LLMs) for Burmese across three core NLP

Complete-muE: Optimal Hyperparameter Transfer and Scaling for MoE Models

arXiv:2605.23893v1 Announce Type: new Abstract: We propose Complete-muE, a framework which targets hyperparameter transfer across dense FFN and any Mixture-of-Experts (MoE) setups in transformer block

Composing People Together: Iterative Pose-Image Generation for Multi-Person Interaction Scenes

arXiv:2605.23178v1 Announce Type: new Abstract: Despite recent progress, text-to-image models still struggle to generate semantically diverse and compositionally accurate multi-person interaction scen

safetyarxiv-cs-cv

EM-Vid: Training-Free Entity-Centric Memory for Efficient and Consistent Multi-Shot Video Generation

arXiv:2605.23610v1 Announce Type: cross Abstract: Multi-shot video generation requires maintaining a consistent appearance of recurring entities across shots while remaining faithful to shot-specific

researcharxiv-cs-ai

Exploiting Longitudinal Context in Clinician-Verified Interactive Lesion Tracking

arXiv:2605.23118v1 Announce Type: cross Abstract: Tracking tumor lesions across serial CT scans is essential for oncological response assessment. Existing automated methods face a fundamental trade-of

FAST-ME: Foundation-aware Adaptive Stopping for Motion Estimation for Efficient IoT Video Analysis

arXiv:2605.23428v1 Announce Type: new Abstract: In modern multimedia systems, efficient video processing is critical, especially in resource-constrained environments such as IoT-based camera networks,

Applications

GEM-4D: Geometry-Enhanced Video World Models for Robot Manipulation

arXiv:2605.22882v1 Announce Type: new Abstract: Video world models can generate realistic futures from a single instruction, but they often fail to preserve consistent point-level motion over time. As

applicationsarxiv-cs-cv

Geo-Align: Video Generation Alignment via Metric Geometry Reward

arXiv:2605.23903v1 Announce Type: new Abstract: Camera-controlled video generation has achieved remarkable progress in recent years. However, existing video-to-video re-rendering methods primarily rel

safetyarxiv-cs-cv

Multimodal Distribution Matching for Vision-Language Dataset Distillation

arXiv:2605.23482v1 Announce Type: cross Abstract: Dataset distillation compresses large training sets into compact synthetic datasets while preserving downstream performance. As modern systems increas

safetyarxiv-cs-ai

NP-LoRA: Null Space Projection for Subject-Style LoRA Fusion

arXiv:2511.11051v3 Announce Type: replace Abstract: Low-Rank Adaptation (LoRA) fusion enables the composition of subject and style representations for controllable generation without retraining. Howev

Scalable Heterogeneous Graph Foundation Models for Data-Driven Optimal Power Flow in Smart Grids

arXiv:2605.23194v1 Announce Type: cross Abstract: Fast and reliable optimal power flow (OPF) approximation is essential for reliable smart-grid operation, yet many learning-based surrogates either fla

researcharxiv-cs-ai

Local Ai

SCOPE: Simulating Cross-game Operations in Playable Environments for FPS World Models

arXiv:2605.23345v1 Announce Type: new Abstract: Interactive world models for first-person shooter (FPS) games must resolve high-frequency overlapping control signals at every frame without disrupting

local-aiarxiv-cs-cv

Training-Free Looped Transformers

arXiv:2605.23872v1 Announce Type: new Abstract: We introduce training-free looped transformers, in which a lightweight inference-time wrapper loops a contiguous mid-stack block of layers of a frozen c

V-VLAPS: Value-Guided Planning for Vision-Language-Action Models

arXiv:2601.00969v2 Announce Type: replace-cross Abstract: Vision-language-action (VLA) models provide strong action priors for robotic manipulation, but their reactive behavior can fail under distribu

safetyarxiv-cs-ai

VINS-120K: Ultra High-Resolution Image Editing with A Large-Scale Dataset

arXiv:2605.23518v1 Announce Type: new Abstract: Directly editing ultra-high-resolution (UHR) images is valuable but underexplored, primarily due to the lack of high-quality data and the challenge in m

A Boundary-Layer Mechanism for One-Third Scaling in Online Softmax Classification

arXiv:2605.22341v1 Announce Type: new Abstract: Hard-label classification is usually trained with smooth surrogate losses, most prominently softmax cross-entropy. We isolate an asymptotic mechanism by

AMUSE: Anytime Muon with Stable Gradient Evaluation

arXiv:2605.22432v1 Announce Type: new Abstract: Modern deep learning commonly relies on AdamW with prescribed learning rate schedules, but recent works challenge both components: Schedule-Free optimiz

Bringing Stability to Diffusion: Decomposing and Reducing Variance of Training Masked Diffusion Models

arXiv:2511.18159v2 Announce Type: replace Abstract: Masked diffusion models (MDMs) are a promising alternative to autoregressive models (ARMs), but they suffer from inherently much higher training var

Local Ai

Can Transformers Learn to Verify During Backtracking Search?

arXiv:2605.22221v1 Announce Type: new Abstract: Backtracking search underlies classical constraint solvers, planners, and theorem provers. Recent transformer-based reasoning systems explore search tre

local-aiarxiv-cs-lg

Cross-Species RSA Reveals Conserved Early Visual Alignment but Divergent Higher-Area Rankings Across Human fMRI and Macaque Electrophysiology

arXiv:2605.22401v1 Announce Type: new Abstract: Does the relationship between learning rules and brain alignment generalize across species? We extend our prior finding that untrained CNNs match backpr

safetyarxiv-cs-lg

Local Ai

Disentanglement Beyond Generative Models with Riemannian ICA

arXiv:2605.22531v1 Announce Type: new Abstract: There is a gap between the theoretical foundations of disentanglement and the practice of modern representation learning. Existing theoretical framework

local-aiarxiv-cs-lg

Applications

Physics-Informed Generative Solver: Bridging Data-Driven Priors and Conservation Laws for Stable Spatiotemporal Field Reconstruction

arXiv:2605.22338v1 Announce Type: new Abstract: Reconstructing continuous physical fields from sparse measurements is a central inverse problem, but data-driven generative models can produce states th

applicationsarxiv-cs-lg

Rethinking Forward Processes for Score-Based Nonlinear Data Assimilation in High Dimensions

arXiv:2604.02889v2 Announce Type: replace-cross Abstract: Data assimilation is the process of estimating the state of a dynamical system over time by combining model predictions with measurements. Thi

Action with Visual Primitives

arXiv:2605.22183v1 Announce Type: new Abstract: Vision-Language-Action (VLA) models have emerged as a promising paradigm for generalist robotic manipulation. A common design in current architectures m

researcharxiv-cs-ro

Applications

BodyReLux: Temporally Consistent Full-Body Video Relighting

arXiv:2605.21766v1 Announce Type: new Abstract: Being able to relight human performance is a fundamental task for post production and content creation. We present BodyReLux, a subject-specific video d

applicationsarxiv-cs-cv

Check Your LLM's Secret Dictionary! Five Lines of Code Reveal What Your LLM Learned (Including What It Shouldn't Have)

model-releasesarxiv-cs-cl

arXiv:2605.22005v1 Announce Type: cross Abstract: We show that singular value decomposition of the lm_head} weight matrix of a transformer-based large language model -- requiring only five lines of Py

Tutorials

Conceptualizing Embeddings: Sparse Disentanglement for Vision-Language Models

arXiv:2605.22679v1 Announce Type: new Abstract: Vision-language models learn powerful multimodal embeddings, yet their internal semantics remain opaque. While sparse autoencoders (SAEs) can extract in

tutorialsarxiv-cs-cv

CrossVLA: Cross-Paradigm Post-Training and Inference Optimization for Vision-Language-Action Models

arXiv:2605.21854v1 Announce Type: new Abstract: Vision-Language-Action (VLA) models have rapidly converged on a small set of architectural patterns: discrete-token autoregression (e.g. OpenVLA) and co

Declarative Data Services: Structured Agentic Discovery for Composing Data Systems

arXiv:2605.20690v1 Announce Type: new Abstract: Agentic discovery has shown that LLM-driven search can find novel algorithms, designs, and code under benchmark conditions. Translating the paradigm to

DecQ: Detail-Condensing Queries for Enhanced Reconstruction and Generation in Representation Autoencoders

arXiv:2605.22777v1 Announce Type: new Abstract: Representation Autoencoders (RAEs) leverage frozen vision foundation models (VFMs) as tokenizer encoders, providing robust high-level representations th

researcharxiv-cs-cv

Agents

Diffusion-guided Generalizable Enhancer for Urban Scene Reconstruction

arXiv:2605.22420v1 Announce Type: new Abstract: Urban scene reconstruction from real-world observations has emerged as a powerful tool for self-driving development and testing. While current neural re

agentsarxiv-cs-cv

Do Vision Models Encode Object-Level Semantic Relatedness? A Cognitive Psychology-Inspired Benchmark

arXiv:1709.03806v2 Announce Type: replace Abstract: Modern vision models have achieved strong object-recognition performance, yet it remains unclear whether their representations encode object-level s

Efficient Agentic Reasoning Through Self-Regulated Simulative Planning

model-releasesarxiv-cs-cl

arXiv:2605.22138v1 Announce Type: cross Abstract: How should an agent decide when and how to plan? A dominant approach builds agents as reactive policies with adaptive computation (e.g., chain-of-thou

Agents

GA-VLN: Geometry-Aware BEV Representation for Efficient Vision-Language Navigation

arXiv:2605.22036v1 Announce Type: new Abstract: Despite significant progress in Vision-Language Navigation (VLN), existing approaches still rely on dense RGB videos that produce excessive patch tokens

agentsarxiv-cs-cv

InfVSR: Breaking Length Limits of Generic Video Super-Resolution

arXiv:2510.00948v2 Announce Type: replace Abstract: Real-world videos often extend over thousands of frames. Existing generative video super-resolution (VSR) approaches, however, face two persistent c

LACO: Adaptive Latent Communication for Collaborative Driving

arXiv:2605.22504v1 Announce Type: cross Abstract: Collaborative driving aims to improve safety and efficiency by enabling connected vehicles to coordinate under partial observability. Recent approache

safetyarxiv-cs-cv

LVDrive: Latent Visual Representation Enhanced Vision-Language-Action Autonomous Driving Model

arXiv:2605.22089v1 Announce Type: new Abstract: Vision-Language-Action (VLA) models have emerged as a promising framework for end-to-end autonomous driving. However, existing VLAs typically rely on sp