Jindong Gu

30

Papers

492

Total Citations

1

Affiliations

Affiliations

University of Oxford & Google DeepMind

Papers (30)

MM-SafetyBench: A Benchmark for Safety Evaluation of Multimodal Large Language Models

FedDAT: An Approach for Foundation Model Finetuning in Multi-Modal Heterogeneous Federated Learning

Improved Techniques for Optimization-Based Jailbreaking on Large Language Models

Latent Guard: a Safety Framework for Text-to-image Generation

Dataset Distillation by Automatic Training Trajectories

Does Few-Shot Learning Suffer from Backdoor Attacks?

FedBiP: Heterogeneous One-Shot Federated Learning with Personalized Latent Diffusion Models

Which Model Generated This Image? A Model-Agnostic Approach for Origin Attribution

ReVisionLLM: Recursive Vision-Language Model for Temporal Grounding in Hour-Long Videos

Discretization-Induced Dirichlet Posterior for Robust Uncertainty Quantification on Regression

Fair Generation without Unfair Distortions: Debiasing Text-to-Image Generation with Entanglement-Free Attention

Do DALL-E and Flamingo Understand Each Other?

FRAug: Tackling Federated Learning with Non-IID Features via Representation Augmentation

Multi-Event Video-Text Retrieval

Are Vision Transformers Robust to Patch Perturbations?

Towards Efficient Adversarial Training on Vision Transformers

Watermark Vaccine: Adversarial Attacks to Prevent Watermark Removal

SegPGD: An Effective and Efficient Adversarial Attack for Evaluating and Boosting Segmentation Robustness

Hide in Thicket: Generating Imperceptible and Rational Adversarial Perturbations on 3D Point Clouds

Localizing Events in Videos with Multimodal Queries

Not Just Text: Uncovering Vision Modality Typographic Threats in Image Generation Models

SafetyDPO: Scalable Safety Alignment for Text-to-Image Generation

FedPop: Federated Population-based Hyperparameter Tuning

Self-Discovering Interpretable Diffusion Latent Directions for Responsible Text-to-Image Generation

Initialization Matters for Adversarial Transfer Learning

Provably Better Explanations with Optimized Aggregation of Feature Attributions

Improving the Robustness of Capsule Networks to Image Affine Transformations

Capsule Network Is Not More Robust Than Convolutional Network

Backdoor Defense via Adaptively Splitting Poisoned Dataset

Benchmarking Robustness of Adaptation Methods on Pre-trained Vision-Language Models