Diffusion Model Connect Session - 22nd December 2025

ResearchTrend.AI Diffusion Model Connect Session: Dynamic Guidance & Protein Flows!

We are excited to announce our upcoming biweekly Diffusion Model (DiffM) Connect Session on ResearchTrend.AI!

This virtual session 💻 features two essential presentations from leading researchers 🧑‍🔬, focusing on optimizing guidance for conditional generation and revolutionary physics-based protein modeling.

Agenda (UTC) - Monday, December 22nd

08:00 - 08:30: Alexandr G.(Google DeepMind)
📄 Paper: Learn to Guide Your Diffusion Model
💡 Abstract: Classifier-free guidance (CFG) is the industry standard for improving sample quality, but static guidance weights often lead to distributional mismatch. Alexandre introduces a novel framework to learn dynamic guidance weights that are continuous functions of conditioning and time. This approach significantly improves Fréchet Inception Distance (FID) and image-prompt alignment by minimizing the gap between the true and guided diffusion processes.

08:30 - 09:00: Yogesh Verma (Aalto University)
📄 Paper: Let Physics Guide Your Protein Flows: Topology-aware Unfolding and Generation
💡 Abstract: Standard protein diffusion models often ignore physical constraints, leading to steric clashes and broken bonds. Yogesh presents PhysFlow, a physics-inspired non-linear noising process grounded in classical mechanics. By unfolding proteins into secondary structures while preserving topological integrity and integrating with flow-matching on SE(3), this model achieves state-of-the-art performance in producing designable and novel protein structures.

🌟 This is a fantastic opportunity to engage directly with research that enhances the mathematical precision of guidance and the physical realism of biological generative models.

🗓️ Time: 8:00 AM - 9:00 AM UTC
📍 Location: Virtual
👉 Register for this event here: https://lnkd.in/gDbXfTnb
Don't miss our future sessions! 📆 Find out more about upcoming events: https://lnkd.in/etXjzcUq