Search: “pretraining”

arXiv:2604.06832v2 Announce Type: replace Abstract: Vision-language models (VLMs) predominantly rely on autoregressive decoding, which generates tokens one at a time and fundamentally limits inference

agentsarxiv-cs-cl

Content type

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

Hardware

Fast Model-guided Instance-wise Adaptation Framework for Real-world Pansharpening with Fidelity Constraints

DGX agent

arXiv:2604.08903v1 Announce Type: new Abstract: Pansharpening aims to generate high-resolution multispectral (HRMS) images by fusing low-resolution multispectral (LRMS) and high-resolution panchromati

hardwarearxiv-cs-cv

13 Apr 2026

Fine-tuning is Not Enough: A Parallel Framework for Collaborative Imitation and Reinforcement Learning in End-to-end Autonomous Driving

model-releasesarxiv-cs-ai

arXiv:2603.13842v3 Announce Type: replace-cross Abstract: End-to-end autonomous driving is typically built upon imitation learning (IL), yet its performance is constrained by the quality of human demo

InstrAct: Towards Action-Centric Understanding in Instructional Videos

arXiv:2604.08762v1 Announce Type: cross Abstract: Understanding instructional videos requires recognizing fine-grained actions and modeling their temporal relations, which remains challenging for curr

safetyarxiv-cs-ai

Learning Vision-Language-Action World Models for Autonomous Driving

arXiv:2604.09059v1 Announce Type: cross Abstract: Vision-Language-Action (VLA) models have recently achieved notable progress in end-to-end autonomous driving by integrating perception, reasoning, and

safetyarxiv-cs-ai

Low-Data Supervised Adaptation Outperforms Prompting for Cloud Segmentation Under Domain Shift

arXiv:2604.08956v1 Announce Type: new Abstract: Adapting vision-language models to remote sensing imagery presents a fundamental challenge: both the visual and linguistic distributions of satellite da

OASIS: Online Activation Subspace Learning for Memory-Efficient Training

arXiv:2604.09406v1 Announce Type: new Abstract: Training large language models (LLMs) is constrained by memory requirements, with activations accounting for a substantial fraction of the total footpri

On the Limits of Layer Pruning for Generative Reasoning in Large Language Models

arXiv:2602.01997v2 Announce Type: replace-cross Abstract: Recent work has shown that layer pruning can effectively compress large language models (LLMs) while retaining strong performance on classific

Local Ai

Online3R: Online Learning for Consistent Sequential Reconstruction Based on Geometry Foundation Model

arXiv:2604.09480v1 Announce Type: new Abstract: We present Online3R, a new sequential reconstruction framework that is capable of adapting to new scenes through online learning, effectively resolving

local-aiarxiv-cs-cv

Tutorials

Physically Grounded 3D Generative Reconstruction under Hand Occlusion using Proprioception and Multi-Contact Touch

arXiv:2604.09100v1 Announce Type: new Abstract: We propose a multimodal, physically grounded approach for metric-scale amodal object reconstruction and pose estimation under severe hand occlusion. Unl

tutorialsarxiv-cs-cv

Reflection of Episodes: Learning to Play Game from Expert and Self Experiences

arXiv:2502.13388v4 Announce Type: replace Abstract: StarCraft II is a complex and dynamic real-time strategy (RTS) game environment, which is very suitable for artificial intelligence and reinforcemen

Seeing is Believing: Robust Vision-Guided Cross-Modal Prompt Learning under Label Noise

model-releasesarxiv-cs-ai

arXiv:2604.09532v1 Announce Type: cross Abstract: Prompt learning is a parameter-efficient approach for vision-language models, yet its robustness under label noise is less investigated. Visual conten

UniSemAlign: Text-Prototype Alignment with a Foundation Encoder for Semi-Supervised Histopathology Segmentation

arXiv:2604.09169v1 Announce Type: new Abstract: Semi-supervised semantic segmentation in computational pathology remains challenging due to scarce pixel-level annotations and unreliable pseudo-label s

safetyarxiv-cs-cv

VAG: Dual-Stream Video-Action Generation for Embodied Data Synthesis

arXiv:2604.09330v1 Announce Type: cross Abstract: Recent advances in robot foundation models trained on large-scale human teleoperation data have enabled robots to perform increasingly complex real-wo

safetyarxiv-cs-cv

WAND: Windowed Attention and Knowledge Distillation for Efficient Autoregressive Text-to-Speech Models

arXiv:2604.08558v1 Announce Type: cross Abstract: Recent decoder-only autoregressive text-to-speech (AR-TTS) models produce high-fidelity speech, but their memory and compute costs scale quadratically

Adapting Foundation Models for Annotation-Efficient Adnexal Mass Segmentation in Cine Images

arXiv:2604.08045v1 Announce Type: new Abstract: Adnexal mass evaluation via ultrasound is a challenging clinical task, often hindered by subjective interpretation and significant inter-observer variab

AnomalyVFM -- Transforming Vision Foundation Models into Zero-Shot Anomaly Detectors

arXiv:2601.20524v2 Announce Type: replace Abstract: Zero-shot anomaly detection aims to detect and localise abnormal regions in the image without access to any in-domain training images. While recent

Are Face Embeddings Compatible Across Deep Neural Network Models?

arXiv:2604.07282v1 Announce Type: cross Abstract: Automated face recognition has made rapid strides over the past decade due to the unprecedented rise of deep neural network (DNN) models that can be t

safetyarxiv-cs-lg

Behavior-Aware Item Modeling via Dynamic Procedural Solution Representations for Knowledge Tracing

arXiv:2604.08260v1 Announce Type: new Abstract: Knowledge Tracing (KT) aims to predict learners' future performance from past interactions. While recent KT approaches have improved via learning item r

researcharxiv-cs-cl

Component-Adaptive and Lesion-Level Supervision for Improved Small Structure Segmentation in Brain MRI

arXiv:2604.08015v1 Announce Type: new Abstract: We propose a unified objective function, termed CATMIL, that augments the base segmentation loss with two auxiliary supervision terms operating at diffe

Tutorials

Cross-Modal Emotion Transfer for Emotion Editing in Talking Face Video

arXiv:2604.07786v1 Announce Type: new Abstract: Talking face generation has gained significant attention as a core application of generative models. To enhance the expressiveness and realism of synthe

tutorialsarxiv-cs-cv

DeCo: Frequency-Decoupled Pixel Diffusion for End-to-End Image Generation

arXiv:2511.19365v2 Announce Type: replace-cross Abstract: Pixel diffusion aims to generate images directly in pixel space in an end-to-end fashion. This approach avoids the limitations of VAE in the t

Discrete Flow Matching Policy Optimization

arXiv:2604.06491v1 Announce Type: cross Abstract: We introduce Discrete flow Matching policy Optimization (DoMinO), a unified framework for Reinforcement Learning (RL) fine-tuning Discrete Flow Matchi

safetyarxiv-cs-ai

Distilling Specialized Orders for Visual Generation

arXiv:2504.17069v2 Announce Type: replace Abstract: Autoregressive (AR) image generators are becoming increasingly popular due to their ability to produce high-quality images and their scalability. Ty

Drift-Based Policy Optimization: Native One-Step Policy Learning for Online Robot Control

arXiv:2604.03540v2 Announce Type: replace Abstract: Although multi-step generative policies achieve strong performance in robotic manipulation by modeling multimodal action distributions, they require

safetyarxiv-cs-ro

Flux Attention: Context-Aware Hybrid Attention for Efficient LLMs Inference

model-releasesarxiv-cs-cl

arXiv:2604.07394v1 Announce Type: cross Abstract: The quadratic computational complexity of standard attention mechanisms presents a severe scalability bottleneck for LLMs in long-context scenarios. W

Hardware

Graph Neural ODE Digital Twins for Control-Oriented Reactor Thermal-Hydraulic Forecasting Under Partial Observability

arXiv:2604.07292v1 Announce Type: new Abstract: Real-time supervisory control of advanced reactors requires accurate forecasting of plant-wide thermal-hydraulic states, including locations where physi

hardwarearxiv-cs-lg

GroundingAnomaly: Spatially-Grounded Diffusion for Few-Shot Anomaly Synthesis

arXiv:2604.08301v1 Announce Type: new Abstract: The performance of visual anomaly inspection in industrial quality control is often constrained by the scarcity of real anomalous samples. Consequently,

How Independent are Large Language Models? A Statistical Framework for Auditing Behavioral Entanglement and Reweighting Verifier Ensembles

arXiv:2604.07650v1 Announce Type: cross Abstract: The rapid growth of the large language model (LLM) ecosystem raises a critical question: are seemingly diverse models truly independent? Shared pretra

safetyarxiv-cs-cl

ImVideoEdit: Image-learning Video Editing via 2D Spatial Difference Attention Blocks

arXiv:2604.07958v1 Announce Type: new Abstract: Current video editing models often rely on expensive paired video data, which limits their practical scalability. In essence, most video editing tasks c

JoyAI-LLM Flash: Advancing Mid-Scale LLMs with Token Efficiency

model-releasesarxiv-cs-ai

arXiv:2604.03044v2 Announce Type: replace-cross Abstract: We introduce JoyAI-LLM Flash, an efficient Mixture-of-Experts (MoE) language model designed to redefine the trade-off between strong performan

Low-Rank Key Value Attention

arXiv:2601.11471v3 Announce Type: replace Abstract: The key-value (KV) cache is a primary memory bottleneck in Transformers. We propose Low-Rank Key-Value (LRKV) attention, which reduces KV cache memo

Monocular Depth Estimation From the Perspective of Feature Restoration: A Diffusion Enhanced Depth Restoration Approach

arXiv:2604.07664v1 Announce Type: new Abstract: Monocular Depth Estimation (MDE) is a fundamental computer vision task with important applications in 3D vision. The current mainstream MDE methods empl

Non-identifiability of Explanations from Model Behavior in Deep Networks of Image Authenticity Judgments

arXiv:2604.07254v1 Announce Type: cross Abstract: Deep neural networks can predict human judgments, but this does not imply that they rely on human-like information or reveal the cues underlying those

Local Ai

PixelCAM: Pixel Class Activation Mapping for Histology Image Classification and ROI Localization

arXiv:2503.24135v3 Announce Type: replace Abstract: Weakly supervised object localization (WSOL) methods allow training models to classify images and localize ROIs. WSOL only requires low-cost image-c

local-aiarxiv-cs-cv

Reinforcement-Guided Synthetic Data Generation for Privacy-Sensitive Identity Recognition

arXiv:2604.07884v1 Announce Type: new Abstract: High-fidelity generative models are increasingly needed in privacy-sensitive scenarios, where access to data is severely restricted due to regulatory an

Rethinking Entropy Allocation in LLM-based ASR: Understanding the Dynamics between Speech Encoders and LLMs

model-releasesarxiv-cs-cl

arXiv:2604.08003v1 Announce Type: cross Abstract: Integrating large language models (LLMs) into automatic speech recognition (ASR) has become a dominant paradigm. Although recent LLM-based ASR models

Semantic Noise Reduction via Teacher-Guided Dual-Path Audio-Visual Representation Learning

arXiv:2604.08147v1 Announce Type: cross Abstract: Recent advances in audio-visual representation learning have shown the value of combining contrastive alignment with masked reconstruction. However, j

safetyarxiv-cs-cv

Stress Estimation in Elderly Oncology Patients Using Visual Wearable Representations and Multi-Instance Learning

arXiv:2604.06990v1 Announce Type: cross Abstract: Psychological stress is clinically relevant in cardio-oncology, yet it is typically assessed only through patient-reported outcome measures (PROMs) an

The Illusion of Superposition? A Principled Analysis of Latent Thinking in Language Models

arXiv:2604.06374v1 Announce Type: cross Abstract: Latent reasoning via continuous chain-of-thoughts (Latent CoT) has emerged as a promising alternative to discrete CoT reasoning. Operating in continuo

UniversalVTG: A Universal and Lightweight Foundation Model for Video Temporal Grounding

arXiv:2604.08522v1 Announce Type: new Abstract: Video temporal grounding (VTG) is typically tackled with dataset-specific models that transfer poorly across domains and query styles. Recent efforts to

Weakly-Supervised Lung Nodule Segmentation via Training-Free Guidance of 3D Rectified Flow

arXiv:2604.08313v1 Announce Type: new Abstract: Dense annotations, such as segmentation masks, are expensive and time-consuming to obtain, especially for 3D medical images where expert voxel-wise labe

You Point, I Learn: Online Adaptation of Interactive Segmentation Models for Handling Distribution Shifts in Medical Imaging

arXiv:2503.06717v3 Announce Type: replace Abstract: Interactive segmentation uses real-time user inputs, such as mouse clicks, to iteratively refine model predictions. Although not originally designed