Skip navigation
Introducing CLP: coupled learning and planning from limited experience and delayed outcomes.
0:00 / 0:57

Introducing CLP: coupled learning and planning from limited experience and delayed outcomes.

Rei

@rei_labs16K followers

Visit on X
215 likesDownloadOriginal post

13,367 views

Introducing CLP: coupled learning and planning from limited experience and delayed outcomes. Learning guides trials; planning and execution generate evidence. We study a response to inevitable training-data scarcity, starting with adversarial decision-making and machine control.

25 Comments

Comments from the original post on X (Twitter)

  • Rei's profile picture

    @rei_labs

    We’ve documented our exploration of how retained values and routines make later planning more useful, and how planning supports further learning. The full article includes the experiments and results.

    38 likes

  • Rei's profile picture

    @rei_labs

    Training data is finite. Progress may require experience that does not exist yet. Adapt-1, our general learner, uses retained values and routines to guide trials, then learns from planning assessments and observed outcomes without demonstrations or expert action labels.

    31 likes

  • Rei's profile picture

    @rei_labs

    The chess experiment began with 800 games. Records contained twelve piece counts and an eventual outcome; 77.1% followed random moves. No expert move labels or engine evaluations were supplied. The first fit had little direct guidance about good decisions.

    18 likes

  • Rei's profile picture

    @rei_labs

    After an initial value fit, one-ply lookahead reconsidered stored positions. Search assessments helped form later targets, while observed outcomes kept a separate contribution. Revised values then changed decisions under the same planning settings.

    19 likes

  • Rei's profile picture

    @rei_labs

    In Machina, search compares four complete eight-command routines. The selected routine runs separately; its commands and measured outcome enter learning. Each acquisition run generated 2,048 proposals and admitted 512 executions.

    19 likes

  • Rei's profile picture

    @rei_labs

    Four candidates per choice can matter across hundreds of rounds: a selected execution changes the basis of later proposals. In the recorded trace, a new parent changed all four next routines, even with the same requested seed.

    18 likes

  • Rei's profile picture

    @rei_labs

    The result is a retained skill: an acquired routine plus a learned correction for the starting observation. Both stay fixed during use, while commands vary with the start. The skill can execute without another search.

    19 likes

  • Rei's profile picture

    @rei_labs

    After acquisition and refinement, the frozen docking skill succeeded on 34/64 new starts through direct execution, or 52/64 with evaluation search. No learning took place during evaluation. API details and examples:

    24 likes

  • reitern's profile picture

    @0xreitern

    crude learning: learning from scratch without demonstration, engine label or evaluation retained learning: success in applying previously learned skill in new initial environment

    14 likes

  • Guidø's profile picture

    @guid_o

    look at me, i'm reitard now!

    9 likes

  • Ruimtepak 🔮's profile picture

    @Ruimtepak_

    greitness

    7 likes

  • 0xFex (rei/acc)'s profile picture

    @ifexito1

    LFG More gigabrain tech I’ll never fully comprehend

    6 likes

  • omoi's profile picture

    @0xacrnm

    technically speaking the data you provided adapt-1 is not just noisy it’s bad but it still got something out of it

    6 likes

  • Knox's profile picture

    @kn0xhound

    👀👀👀

    6 likes

  • Lojom's profile picture

    @Lojom53

    Interesting direction. Coupling learning and planning via retained state makes sense with scarce experience and delayed feedback. Chess setup is clean: piece counts + outcomes (mostly random), then lookahead improves targets. Curious how far this scales. Grei

    6 likes

  • 8480_Howard's profile picture

    @8480Howard

    Grei

    5 likes

  • S's profile picture

    @abdominoux

    Tech is Greit

    5 likes

  • m0r31ra's profile picture

    @moreir92

    Grei

    2 likes

  • kali's profile picture

    @kaliXBT

    👀

    1 like

  • Hova's profile picture

    @Hovavayo

    Hey @grok sıfırdan yaradıblar yoxsa fine tune ediblər

    1 like