New ‘Test-Time Training’ method lets AI keep learning without exploding inference costs

By allowing models to actively update their weights during inference, Test-Time Training (TTT) creates a "compressed memory" ...

Convert Any Image or Photo into Editable 3D Models for Makers & Creators : SAM 3D

Try SAM 3D to create editable 3D models and meshes from images, with manual scale and rotate tools, helping beginners turn ...
GitHub

Towards Robust and Controllable Text-to-Motion via Masked Autoregressive Diffusion (ACM MM 2025)

@inproceedings{zhang2025towards, author = {Zhang, Zongye and Kong, Bohan and Liu, Qingjie and Wang, Yunhong}, title = {Towards Robust and Controllable Text-to-Motion via Masked Autoregressive ...
IEEE

DADSR: Degradation-Aware Diffusion Super-Resolution Model for Object-Level SAR Image

Abstract: Synthetic aperture radar (SAR) plays a vital role in remote sensing applications but suffers from coupled degradation problems, including resolution deterioration, speckle noise, and ...
IEEE

EdgeDiff: Energy-Efficient Multi-Modal Few-Step Diffusion Model Accelerator Using Mixed-Precision and Reordered Group Quantization

Abstract: Recent advances in diffusion models (DMs)—such as few-step denoising and multi-modal conditioning—have significantly improved computational efficiency and functional flexibility, but they ...
GitHub

ab0626/Financial-Time-Series-Generation-with-GANs-Diff

A comprehensive implementation of state-of-the-art generative models for creating synthetic financial time series data that preserves real-world statistical properties, temporal dynamics, and market ...