Wei Lin

30

Papers

1,261

Total Citations

Papers (30)

MME: A Comprehensive Evaluation Benchmark for Multimodal Large Language Models

Semantic Convergence: Harmonizing Recommender Systems via Two-Stage Alignment and Behavioral Semantic Tokenization

LiveXiv - A Multi-Modal live benchmark based on Arxiv papers content

Point-to-Region Loss for Semi-Supervised Point-Based Crowd Counting

How to Trade Off the Quantity and Capacity of Teacher Ensemble: Learning Categorical Distribution to Stochastically Employ a Teacher for Distillation

Efficient Long Context Fine-tuning with Chunk Flow

Proximal Mapping Loss: Understanding Loss Functions in Crowd Counting & Localization

PortraitBooth: A Versatile Portrait Model for Fast Identity-preserved Personalization

Towards Robust Learning to Optimize with Theoretical Guarantees

From Fourier to Neural ODEs: Flow Matching for Modeling Complex Systems

FESSNC: Fast Exponentially Stable and Safe Neural Controller

A Statistical Theory of Regularization-Based Continual Learning

Switched Flow Matching: Eliminating Singularities via Switching ODEs

Learning From Synthetic Data for Crowd Counting in the Wild

SwapText: Image Based Texts Transfer in Scenes

Cross-View Cross-Scene Multi-View Crowd Counting

Optimal Transport Minimization: Crowd Localization on Density Maps for Semi-Supervised Counting

Video Test-Time Adaptation for Action Recognition

ActMAD: Activation Matching To Align Distributions for Test-Time-Training

MAtch, eXpand and Improve: Unsupervised Finetuning for Zero-Shot Action Recognition with Language Knowledge

MATE: Masked Autoencoders are Online 3D Test-Time Learners

CycDA: Unsupervised Cycle Domain Adaptation to Learn from Image to Video

PerLA: Perceptive 3D Language Assistant

KOEnsAttack: Towards Efficient Data-Free Black-Box Adversarial Attacks via Knowledge-Orthogonalized Substitute Ensembles

Robust 3D Object Detection using Probabilistic Point Clouds from Single-Photon LiDARs

Zero-Sum vs. Positive-Sum: Effects of Inter-team Competition Modes and Haptic Feedback on Team Flow in Multi-team VR

Adversarial-Inspired Backdoor Defense via Bridging Backdoor and Adversarial Attacks

Arithmetic Feature Interaction Is Necessary for Deep Tabular Learning

Hypergraph Neural Architecture Search

LaFTer: Label-Free Tuning of Zero-shot Classifier using Language and Unlabeled Image Collections