Microsoft

A one-prompt attack that breaks LLM safety alignment

As LLMs and diffusion models power more applications, their safety alignment becomes critical. Our research shows that even minimal downstream fine‑tuning can weaken safeguards, raising a key question ...
News Watchman

Two Sites, One Team: Portsmouth, Paducah trade lessons learned in Golden Era

Sharing knowledge of the best approaches to deactivating process buildings is one of the ways the two Portsmouth Paducah ...
GitHub

PLAID (Protein Latent Induced Diffusion)

PLAID is a multimodal generative model that can generate protein sequence and all-atom structure based on conditional function and taxonomic prompts. Please see our ...
IEEE

Diffusion Process-Based Model for Network Trajectory Propagation: Formulation, Data Mining, and Cross-Comparative Analysis

Abstract: This paper introduces a diffusion process-based approach to address trajectory generation and vehicle propagation within traffic networks. This approach presents dynamic transition ...
GitHub

Using Powerful Prior Knowledge of Diffusion Model in Deep Unfolding Networks for Image Compressive Sensing

1 School of Electronic and Information Engineering, Beijing Jiaotong University, China 2 School of Automation and Intelligence, Beijing Jiaotong University, China Recently, Deep Unfolding Networks ...