• Home
    • Technology
    • Gaming
    • Entertainment
    • World & Business
    • Science
    • Sports
    • AI
    HomeTechnologyGamingEntertainmentWorld & BusinessScienceSportsAI
    AI
    Announcement

    LLM-to-Flow Distillation aims to steer scientific generators with natural-language goals

    Its creators say iterative DPO on LLM-generated sub-goals adapts generators without task-specific proxy rewards.

    2 Sources, 8h ago, first seen 8h ago

    TLDR

    The creators introduced LLM-to-Flow Distillation (LFD) as a way to adapt scientific generators to goals such as designing peptides with promising anticancer activity or molecules that bind the D2 dopamine receptor. They say it uses iterative DPO on LLM-generated sub-goals, needs no task-specific proxy rewards, works with any frontier LLM through API access, and outperforms FlowGRPO and DiffusionNFT in held-out computational evaluations.

    Combined views

    —

    2 Sources, first seen 8h ago

    — likes— comments— saves— reposts

    Combined views

    —

    2 Sources, first seen 8h ago

    — likes— comments— saves— reposts
    Image: @r_de_santi on X

    Sentiment

    Positive——Negative

    Summary

    Not enough discussion yet.

    No sentiment analysis available yet.

    Featured Source

    Sentiment

    Positive——Negative

    Summary

    Not enough discussion yet.

    No sentiment analysis available yet.

    2 Sources

    Riccardo De Santi@r_de_santiWe’re bringing natural-language control to any scientific generator. We are very excited to introduce LLM-to-Flow Distillation (LFD), a method to adapt scientific generators to high-level natural-language discovery goals. Starting from textual goals like “design peptides with promising anticancer activity” or “design small molecules binding to the D2 dopamine receptor”, LFD distills high-level goal-related knowledge from LLMs into scientific generators through iterative DPO on LLM-generated sub-goals. 🔥 LFD does not need task-specific proxy rewards, works with any frontier LLM via API-access, and outperforms state-of-the-art reward-guided methods such as FlowGRPO and DiffusionNFT on held-out computational evaluations. (1/6)8h
    Bharath Ramsundar@rbhar90RT @r_de_santi: We’re bringing natural-language control to any scientific generator. We are very excited to introduce LLM-to-Flow Distilla…2h
    • Home
    • Technology
    • Gaming
    • Entertainment
    • World & Business
    • Science
    • Sports
    • AI

    2 Sources

    Riccardo De Santi@r_de_santiWe’re bringing natural-language control to any scientific generator. We are very excited to introduce LLM-to-Flow Distillation (LFD), a method to adapt scientific generators to high-level natural-language discovery goals. Starting from textual goals like “design peptides with promising anticancer activity” or “design small molecules binding to the D2 dopamine receptor”, LFD distills high-level goal-related knowledge from LLMs into scientific generators through iterative DPO on LLM-generated sub-goals. 🔥 LFD does not need task-specific proxy rewards, works with any frontier LLM via API-access, and outperforms state-of-the-art reward-guided methods such as FlowGRPO and DiffusionNFT on held-out computational evaluations. (1/6)8h
    Bharath Ramsundar@rbhar90RT @r_de_santi: We’re bringing natural-language control to any scientific generator. We are very excited to introduce LLM-to-Flow Distilla…2h
    Today's Rank

    —

    Not ranked yet

    Today's Rank

    —

    Not ranked yet