Related papers: SafeRedir: Prompt Embedding Redirection for Robust…

Federated Prompt Learning has emerged as a communication-efficient and privacy-preserving paradigm for adapting large vision-language models like CLIP across decentralized clients. However, the security implications of this setup remain…

Machine learning models are known to be vulnerable to adversarial evasion attacks as illustrated by image classification models. Thoroughly understanding such attacks is critical in order to ensure the safety and robustness of critical AI…

Prompt engineering is still the primary way for users of generative text-to-image models to manipulate generated images in a targeted way. Based on treating the model as a continuous function and by passing gradients between the image space…

Despite impressive recent advances in text-to-image diffusion models, obtaining high-quality images often requires prompt engineering by humans who have developed expertise in using them. In this work, we present NeuroPrompts, an adaptive…

Recent advancements in image editing have utilized large-scale multimodal models to enable intuitive, natural instruction-driven interactions. However, conventional methods still face significant challenges, particularly in spatial…

Feature selection aims to identify the optimal feature subset for enhancing downstream models. Effective feature selection can remove redundant features, save computational resources, accelerate the model learning process, and improve the…

Generative vision-language models like Stable Diffusion demonstrate remarkable capabilities in creative media synthesis, but they also pose substantial risks of producing unsafe, offensive, or culturally inappropriate content when prompted…

Multimodal Diffusion Transformers (MMDiTs) for text-to-image generation maintain separate text and image branches, with bidirectional information flow between text tokens and visual latents throughout denoising. In this setting, we observe…

Internal Safety Collapse (ISC) is a failure mode in which frontier LLMs, when executing legitimate professional tasks whose correct completion structurally requires harmful content, spontaneously generate that content with safety failure…

Deep generative models can create remarkably photorealistic fake images while raising concerns about misinformation and copyright infringement, known as deepfake threats. Deepfake detection technique is developed to distinguish between real…

Owing to the growing concerns about privacy and regulatory compliance, it is desirable to regulate the output of generative models. To that end, the objective of this work is to prevent the generation of outputs containing undesired…

Recent advances in text-to-image diffusion models have achieved impressive image generation capabilities. However, it remains challenging to control the generation process with desired properties (e.g., aesthetic quality, user intention),…

Text-to-motion (T2M) generation with diffusion backbones achieves strong realism and alignment. Safety concerns in T2M methods have been raised in recent years; existing methods replace discrete VQ-VAE codebook entries to steer the model…

Image restoration involves recovering a high-quality clean image from its degraded version. Deep learning-based methods have significantly improved image restoration performance, however, they have limited generalization ability to…

The pre-trained foundation models (PFMs) have become essential for facilitating large-scale multimodal learning. Researchers have effectively employed the ``pre-train, prompt, and predict'' paradigm through prompt learning to induce…

Web agents have emerged as an effective paradigm for automating interactions with complex web environments, yet remain vulnerable to prompt injection attacks that embed malicious instructions into webpage content to induce unintended…

Recent advances in large-scale text-to-image diffusion models have heightened concerns about their potential misuse, especially in generating harmful or misleading content. This underscores the urgent need for effective machine unlearning,…

Modern text-to-image generative models can inadvertently reproduce copyrighted content memorized in their training data, raising serious concerns about potential copyright infringement. We introduce Guardians of Generation, a model agnostic…

Prompt learning is an effective method to customize Vision-Language Models (VLMs) for various downstream tasks, involving tuning very few parameters of input prompt tokens. Recently, prompt pretraining in large-scale dataset (e.g.,…

In this work, we demonstrate that certain machine unlearning methods may fail under straightforward prompt attacks. We systematically evaluate eight unlearning techniques across three model families using output-based, logit-based, and…