Blog(4)

• Blind Face Video Restoration with Temporal Consistent Generative Prior and Degradation-Aware Prompt

  In real-world scenarios, both face images and videos may suffer from unknown and varied types of degradation, such as down-sampling, noise, blur, and compression.

• Self-Attention Decomposition for Training Free Diffusion Editing

  Diffusion models have established themselves as the state-of-the-art for generative modeling, producing high-fidelity images through iterative denoising.

• LAFUFU: Latent Acoustic Features for Ultra-Fast Utterance Restoration

  Speech (utterance) restoration is the task of recreating high-fidelity speech from imperfect recordings degraded by noise, reverberation, and other distortions. Such problems have been addressed by classical and neural signal processing methods [21], but their reliance on fixed statistical assumptions limits generalization to varied acoustic conditions.

View More

Research Areas(0)

Publications(14)

• WhiFlash-Token-Level-Cross-Paradigm-Routing-for-Speculative-Decoding

  AuthorYoung D. Kwon, Miles Williams, Rui Li, Alexandros Kouris, Stylianos I. Venieris

  PublishedEMNLP 2026 Empirical Methods in Natural Language Processing

  Date2026-10-24

• NanoFLUX: Distillation-Driven Compression of Large Text-to-Image Generation Models for Mobile Devices

  AuthorRuchika Chavhan, Malcolm Chadwick, Alberto Gil Ramos, Luca Morreale, Mehdi Noroozi

  PublishedICML 2026 (International Conference on Machine Learning)

  Date2026-07-07

• Weight-Space Learning for Certifiable Few-shot Transfer Learning

  AuthorFady Rezk, Royson Lee, Timothy Hospedales, Minyoung Kim

  PublishedICML 2026 (International Conference on Machine Learning)

  Date2026-07-07

View More

News(5)

• [Blog] Self-Attention Decomposition for Training Free Diffusion Editing

  Diffusion models have established themselves as the state-of-the-art for generative modeling, producing high-fidelity images through iterative denoising.

• [Blog] GalaxyEdit: Large Scale Image Editing Dataset with Enhanced Diffusion Adapter

  Have you ever wished you could easily add or remove objects in your photos—like adding a candle on the top of your birthday cake, or removing that background person from your photo?

• [Blog] Edge-SD-SR: Low Latency and Parameter Efficient On-device Super-Resolution with Stable Diffusion via Bidirectional Conditioning

  Stable Diffusion [1] for Super Resolution (i.e. SD-SR), has been shown to produce steep improvements compared to previous SR approaches.

View More

Others(0)