Zhangyang Wang

111

Papers

638

Total Citations

Papers (111)

Feature 3DGS: Supercharging 3D Gaussian Splatting to Enable Distilled Feature Fields

StreamingT2V: Consistent, Dynamic, and Extendable Long Video Generation from Text

OpenBias: Open-set Bias Detection in Text-to-Image Generative Models

PAIR Diffusion: A Comprehensive Multimodal Object-Level Image Editor

From Low Rank Gradient Subspace Stabilization to Low-Rank Weights: Observations, Theories, and Applications

Lift3D: Zero-Shot Lifting of Any 2D Vision Model to 3D

Zero-Painter: Training-Free Layout Control for Text-to-Image Synthesis

REPA Works Until It Doesn’t: Early-Stopped, Holistic Alignment Supercharges Diffusion Training

Social Reward: Evaluating and Enhancing Generative AI through Million-User Feedback from an Online Creative Community

Understanding and Mitigating Bottlenecks of State Space Models through the Lens of Recency and Over-smoothing

Drag-and-Drop LLMs: Zero-Shot Prompt-to-Weights

FlexGS: Train Once, Deploy Everywhere with Many-in-One Flexible 3D Gaussian Splatting

Principled Architecture-aware Scaling of Hyperparameters

On the Provable Separation of Scales in Maximal Update Parameterization

PIPA: Preference Alignment as Prior-Informed Statistical Estimation

Troubleshooting Blind Image Quality Models in the Wild

DiSparse: Disentangled Sparsification for Multitask Model Compression

VideoINR: Learning Video Implicit Neural Representation for Continuous Space-Time Super-Resolution

EI-CLIP: Entity-Aware Interventional Contrastive Learning for E-Commerce Cross-Modal Retrieval

Quarantine: Sparsity Can Uncover the Trojan Attack Trigger for Free

The Principle of Diversity: Training Stronger Vision Transformers Calls for Reducing All Levels of Redundancy

CADTransformer: Panoptic Symbol Spotting Transformer for CAD Drawings

Aug-NeRF: Training Stronger Neural Radiance Fields With Triple-Level Physically-Grounded Augmentations

MMG-Ego4D: Multimodal Generalization in Egocentric Action Recognition

Specialist Diffusion: Plug-and-Play Sample-Efficient Fine-Tuning of Text-to-Image Diffusion Models To Learn Any Unseen Style

NeuralLift-360: Lifting an In-the-Wild 2D Photo to a 3D Object With 360deg Views

AligNeRF: High-Fidelity Neural Radiance Fields via Alignment-Aware Training

Robust Video Super-Resolution With Learned Temporal Dynamics

AOD-Net: All-In-One Dehazing Network

Delving Into Robust Object Detection From Unmanned Aerial Vehicles: A Deep Nuisance Disentanglement Approach

AutoGAN: Neural Architecture Search for Generative Adversarial Networks

Controllable Artistic Text Style Transfer via Shape-Matching GAN

ABD-Net: Attentive but Diverse Person Re-Identification

DeblurGAN-v2: Deblurring (Orders-of-Magnitude) Faster and Better

Searching for Two-Stream Models in Multivariate Space for Video Recognition

SSH: A Self-Supervised Framework for Image Harmonization

GDP: Stabilized Neural Network Pruning via Gates With Differentiable Polarization

StegaNeRF: Embedding Invisible Information within Neural Radiance Fields

Enhancing NeRF akin to Enhancing LLMs: Generalizable NeRF Transformer with Mixture-of-View-Experts

Versatile Diffusion: Text, Images and Variations All in One Diffusion Model

Robust Mixture-of-Expert Training for Convolutional Neural Networks

AdaMV-MoE: Adaptive Multi-Task Vision Mixture-of-Experts

Physics-Driven Turbulence Image Restoration with Stochastic Refinement

Vision HGNN: An Image is More than a Graph of Nodes

Text2Video-Zero: Text-to-Image Diffusion Models are Zero-Shot Video Generators

GAN Slimming: All-in-One GAN Compression by A Unified Optimization Framework

HALO: Hardware-Aware Learning to Optimize

Deep Plastic Surgery: Robust and Controllable Image Editing with Human-Drawn Sketches

Point Cloud Domain Adaptation via Masked Local 3D Structure Prediction

A Simple Single-Scale Vision Transformer for Object Detection and Instance Segmentation

Unified Implicit Neural Stylization

Fast and High Quality Image Denoising via Malleable Convolution

Single Frame Atmospheric Turbulence Mitigation: A Benchmark Study and a New Physics-Inspired Transformer Model

DNA: Improving Few-Shot Transfer Learning with Low-Rank Decomposition and Alignment

SinNeRF: Training Neural Radiance Fields on Complex Scenes from a Single Image

Scalable Learning to Optimize: A Learned Optimizer Can Train Big Models

L2-GCN: Layer-Wise and Learned Efficient Training of Graph Convolutional Networks

Feature4X: Bridging Any Monocular Video to 4D Agentic AI with Versatile Gaussian Feature Fields

Steepest Descent Density Control for Compact 3D Gaussian Splatting

Sparse Transfer Learning Accelerates and Enhances Certified Robustness: A Comprehensive Study

Prompt-Free Diffusion: Taking “Text” out of Text-to-Image Diffusion Models

Taming Mode Collapse in Score Distillation for Text-to-3D Generation

D3: Deep Dual-Domain Based Fast Restoration of JPEG-Compressed Images

Studying Very Low Resolution Recognition Using Deep Networks

Single Image Deraining: A Comprehensive Benchmark Analysis

Collaborative Global-Local Networks for Memory-Efficient Segmentation of Ultra-High Resolution Images

Adversarial Robustness: From Self-Supervised Pre-Training to Fine-Tuning

A Stitch in Time Saves Nine: Small VLM is a Precise Guidance for Accelerating Large VLMs

Peek-a-Boo: Occlusion Reasoning in Indoor Scenes With Plane Representations

Debiased Subjective Assessment of Real-World Image Enhancement

The Lottery Tickets Hypothesis for Supervised and Self-Supervised Pre-Training in Computer Vision Models

Can We Gain More from Orthogonality Regularizations in Training Deep Networks?

Theoretical Linear Convergence of Unfolded ISTA and Its Practical Weights and Thresholds

Model Compression with Adversarial Robustness: A Unified Optimization Framework

E2-Train: Training State-of-the-art CNNs with Over 80% Energy Savings

Learning to Optimize in Swarms

ShiftAddNet: A Hardware-Inspired Deep Network

Graph Contrastive Learning with Augmentations

Training Stronger Baselines for Learning to Optimize

Once-for-All Adversarial Training: In-Situ Tradeoff between Robustness and Accuracy for Free

MATE: Plugging in Model Awareness to Task Embedding for Meta Learning

FracTrain: Fractionally Squeezing Bit Savings Both Temporally and Spatially for Efficient DNN Training

The Lottery Ticket Hypothesis for Pre-trained BERT Networks

Robust Pre-Training by Adversarial Contrastive Learning

AugMax: Adversarial Composition of Random Augmentations for Robust Training

You are caught stealing my winning lottery ticket! Making a lottery ticket claim its ownership

Improving Contrastive Learning on Imbalanced Data via Open-World Sampling

Sparse Training via Boosting Pruning Plasticity with Neuroregeneration

Hyperparameter Tuning is All You Need for LISTA

Delayed Propagation Transformer: A Universal Computation Engine towards Practical Control in Cyber-Physical Systems

Sanity Checks for Lottery Tickets: Does Your Winning Ticket Really Win the Jackpot?

TransGAN: Two Pure Transformers Can Make One Strong GAN, and That Can Scale Up

Chasing Sparsity in Vision Transformers: An End-to-End Exploration

Data-Efficient GAN Training Beyond (Just) Augmentations: A Lottery Ticket Perspective

IA-RED$^2$: Interpretability-Aware Redundancy Reduction for Vision Transformers

The Elastic Lottery Ticket Hypothesis

Stronger NAS with Weaker Predictors

Augmentations in Hypergraph Contrastive Learning: Fabricated and Generative

Sparse Winning Tickets are Data-Efficient Image Recognizers

A Comprehensive Study on Large-Scale Graph Training: Benchmarking and Rethinking

Old can be Gold: Better Gradient Flow can Make Vanilla-GCNs Great Again

Symbolic Distillation for Learned TCP Congestion Control

Signal Processing for Implicit Neural Representations

M³ViT: Mixture-of-Experts Vision Transformer for Efficient Multi-task Learning with Model-Accelerator Co-design

Back Razor: Memory-Efficient Transfer Learning by Self-Sparsified Backpropagation

Randomized Channel Shuffling: Minimal-Overhead Backdoor Attack Detection without Clean Datasets

Deep Architecture Connectivity Matters for Its Convergence: A Fine-Grained Analysis

Trap and Replace: Defending Backdoor Attacks by Trapping Them into an Easy-to-Replace Subnetwork

Scaling Multimodal Pre-Training via Cross-Modality Gradient Harmonization

Deep k-Means: Re-Training and Parameter Sharing with Harder Cluster Assignments for Compressing Deep Convolutions

Plug-and-Play Methods Provably Converge with Properly Trained Denoisers