{"id":2056,"date":"2026-07-18T06:50:14","date_gmt":"2026-07-18T13:50:14","guid":{"rendered":"http:\/\/macdaddy4sure.ai\/?p=2056"},"modified":"2026-07-18T06:50:14","modified_gmt":"2026-07-18T13:50:14","slug":"reinforcement-learning-in-_augmentedintelligence","status":"publish","type":"post","link":"http:\/\/macdaddy4sure.ai\/index.php\/2026\/07\/18\/reinforcement-learning-in-_augmentedintelligence\/","title":{"rendered":"Reinforcement Learning in _AugmentedIntelligence"},"content":{"rendered":"\n

This is the operator hub<\/strong> for RL: what exists, how pieces fit together, and<\/p>\n\n\n\n

which docs\/commands to use first.<\/p>\n\n\n\n

Related maps:<\/p>\n\n\n\n

Doc<\/strong><\/td>Focus<\/strong><\/td><\/tr>
AugmentedIntelligenceSystems.md (AugmentedIntelligenceSystems.md)<\/td>Whole-app system map<\/td><\/tr>
RLCurriculum.md (RLCurriculum.md)<\/td>Domain catalog, smoke tests, deployment ladder<\/td><\/tr>
RLDomains.md (RLDomains.md)<\/td>Expanded domain catalog detail<\/td><\/tr>
MathNN.md (MathNN.md)<\/td>In-process neural nets + rl net \/ learners<\/td><\/tr>
LifeRL.md (LifeRL.md)<\/td>Advisory multi-domain \u201clife\u201d RL<\/td><\/tr>
GamingRL.md (GamingRL.md) \/ GamingRLFPS.md (GamingRLFPS.md)<\/td>FPS gaming RL<\/td><\/tr>
DrivingRL.md (DrivingRL.md)<\/td>Driving RL<\/td><\/tr>
KnowledgeRL.md (KnowledgeRL.md)<\/td>Knowledge \/ claim labeling RL<\/td><\/tr>
DesktopIconVision.md (DesktopIconVision.md)<\/td>Desktop icon classifier (TF SavedModel + LiteRT) for Computer RL<\/td><\/tr>
MinecraftPluginTrainingGuide.md (MinecraftPluginTrainingGuide.md)<\/td>Minecraft DQN + MathNN + executive plans<\/td><\/tr>
CHILD_PSYCHOLOGY_PARENTING_RL.md (CHILD_PSYCHOLOGY_PARENTING_RL.md)<\/td>Safety-bounded parenting coaching<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n

Architecture (three layers)<\/h2>\n\n\n\n

[text]
<\/em>\u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510
\u2502  Typed commands \/ Lua \/ bridges (operator surface)          \u2502
\u2502  rl \u2026 \u00b7 life rl \u2026 \u00b7 minecraft dqn \u2026 \u00b7 gaming rl \u2026 \u00b7 nn \u2026    \u2502
\u2514\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u252c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2518
                            \u2502
\u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u25bc\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510
\u2502  Domain agents (bespoke loops)                              \u2502
\u2502  MinecraftDQN \u00b7 GamingRL \u00b7 DrivingRL \u00b7 Knowledge \u00b7 LifeRL   \u2502
\u2502  + Executive Functions plans (high-level steps)             \u2502
\u2514\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u252c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2518
                            \u2502
        \u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u253c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510
        \u25bc                   \u25bc                   \u25bc
\u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510   \u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510   \u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510
\u2502 RLCore \/      \u2502   \u2502 RLAgentNet    \u2502   \u2502 Remote \/ deep    \u2502
\u2502 Algorithms \/  \u2502   \u2502 MathNN        \u2502   \u2502 trainer (optional)\u2502
\u2502 Curriculum    \u2502   \u2502 RLNeural      \u2502   \u2502 RLDistributed    \u2502
\u2502 smoke envs    \u2502   \u2502 local DQN\u2026    \u2502   \u2502                  \u2502
\u2514\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2518   \u2514\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2518   \u2514\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2518<\/p>\n\n\n\n

    \n
  1. Curriculum \/ smoke layer<\/strong> \u2014 safe synthetic RLEnvs; no live actuation.<\/li>\n\n\n\n
  2. Bespoke domain agents<\/strong> \u2014 real observations (Spigot, client, game mod, HUD).<\/li>\n\n\n\n
  3. MathNN \/ RLAgentNet<\/strong> \u2014 opt-in in-process neural policies (DQN, dueling,<\/li>\n<\/ol>\n\n\n\n

       actor-critic, bandit, hierarchical, multi-head, world-model, reward-model).<\/p>\n\n\n\n

    Command families (quick map)<\/h2>\n\n\n\n
    Prefix<\/strong><\/td>Role<\/strong><\/td><\/tr>
    rl curriculum \u2026 \/ rl sim \u2026<\/td>Domain catalog + synthetic smoke<\/td><\/tr>
    rl real \u2026<\/td>Deployment stages (sim \u2192 offline \u2192 shadow \u2192 gated \u2192 live)<\/td><\/tr>
    rl net \u2026<\/td>Per-agent MathNN policy registry<\/td><\/tr>
    rl cartpole \/ rl dqn<\/td>CartPole ODE + MathNN smoke (optional shield\/tripwire)<\/td><\/tr>
    rl life \u2026 \/ life rl \u2026<\/td>Advisory Life RL<\/td><\/tr>
    nn \u2026<\/td>MathNN demos, math tutoring RL, architectures<\/td><\/tr>
    minecraft dqn \u2026<\/td>Minecraft agent (observe\/train, net, plan, goals)<\/td><\/tr>
    gaming rl \u2026<\/td>FPS gaming RL<\/td><\/tr>
    driving rl \u2026<\/td>Driving RL<\/td><\/tr>
    knowledge \u2026<\/td>Knowledge acquisition RL<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n

    Full routing: rl dispatches through _RLFeatures (curriculum, real, net,<\/p>\n\n\n\n

    cartpole, life). Prefer dry-run \/ smoke<\/strong> before live control.<\/p>\n\n\n\n

    In-process neural RL (MathNN)<\/h2>\n\n\n\n

    Enable a neural policy on an agent:<\/p>\n\n\n\n

    [text]
    <\/em>nn demo
    rl net modes
    rl net enable minecraft
    rl net mode minecraft dueling-dqn
    rl net wm minecraft on
    rl net curiosity minecraft 0.05
    rl net reward minecraft good
    rl net q minecraft
    rl net loss minecraft
    rl net save minecraft checkpoints\/minecraft_mathnn.txt<\/p>\n\n\n\n

    Modes:<\/strong> dqn, dueling-dqn, actor-critic, bandit, hierarchical,<\/p>\n\n\n\n

    multihead-dueling, worldmodel (dueling + curiosity), reward-model.<\/p>\n\n\n\n

    Auxiliary (combinable):<\/strong><\/p>\n\n\n\n

    [text]
    <\/em>rl net rm <agent> on|off
    rl net blend <agent> <alpha>      # \u03b1 on env reward; (1\u2212\u03b1) on reward model
    rl net wm <agent> on|off
    rl net curiosity <agent> <scale><\/p>\n\n\n\n

    CartPole (ODE dynamics + MathNN):<\/p>\n\n\n\n

    [text]
    <\/em>rl cartpole 50
    rl cartpole 50 dueling curiosity shield tripwire<\/p>\n\n\n\n

    Details: MathNN.md (MathNN.md).<\/p>\n\n\n\n

    Domain agents<\/h2>\n\n\n\n

    Minecraft DQN<\/h3>\n\n\n\n
      \n
    • Observations: Spigot bridge (preferred) or RCON fallback; optional client keys.<\/li>\n\n\n\n
    • Neural: minecraft dqn net on \/ rl net enable minecraft.<\/li>\n\n\n\n
    • Goals: reach \/ mine \/ gather \/ earn.<\/li>\n\n\n\n
    • Executive Functions:<\/strong> high-level plan steps (wall-guard, net on, status)<\/li>\n<\/ul>\n\n\n\n

        advance each train\/eval step when enabled.<\/p>\n\n\n\n

      [text]
      <\/em>minecraft bridge start
      minecraft dqn plan on
      minecraft dqn goal mine iron_ore
      minecraft dqn plan detail
      minecraft dqn net on
      minecraft dqn dry-run
      minecraft dqn train 1000
      minecraft dqn plan status<\/p>\n\n\n\n

      See MinecraftPluginTrainingGuide.md (MinecraftPluginTrainingGuide.md).<\/p>\n\n\n\n

      Gaming FPS \/ Strategy<\/h3>\n\n\n\n

      [text]
      <\/em>gaming rl init
      gaming rl dry-run on
      gaming rl algorithm dqn
      rl net enable gaming_fps
      rl net mode gaming_fps actor-critic<\/p>\n\n\n\n

      See GamingRL.md (GamingRL.md), GameModBridges.md (GameModBridges.md).<\/p>\n\n\n\n

      Driving<\/h3>\n\n\n\n

      [text]
      <\/em>driving rl status
      rl net enable driving
      rl net mode driving dueling-dqn<\/p>\n\n\n\n

      See DrivingRL.md (DrivingRL.md).<\/p>\n\n\n\n

      Life RL (advisory)<\/h3>\n\n\n\n

      [text]
      <\/em>life rl domains
      life rl smoke all 8
      life rl net on
      life rl suggest focus_blocks
      life rl reward good<\/p>\n\n\n\n

      See LifeRL.md (LifeRL.md). No external side effects<\/strong> from the policy.<\/p>\n\n\n\n

      Knowledge \/ parenting \/ political<\/h3>\n\n\n\n

      Specialized modules with stronger safety text. Prefer synthetic smoke and human<\/p>\n\n\n\n

      review before anything that could affect real users.<\/p>\n\n\n\n

      Executive Functions + RL<\/h2>\n\n\n\n

      Executive planning is not<\/strong> a second motor policy. It runs high-level<\/strong><\/p>\n\n\n\n

      setup\/coaching steps (status, safety toggles, memory\/awareness) while domain RL<\/p>\n\n\n\n

      handles frame-to-frame actions.<\/p>\n\n\n\n

      Surface<\/strong><\/td>Commands<\/strong><\/td> <\/td> <\/td> <\/td> <\/td><\/tr>
      Generic<\/td>plan create \u2026, plan next, plan detail<\/td> <\/td> <\/td> <\/td> <\/td><\/tr>
      RSI<\/td>rsi agent plan \u2026 (sync\/advance with RSI goals)<\/td> <\/td> <\/td> <\/td> <\/td><\/tr>
      Minecraft<\/td>`minecraft dqn plan on\\<\/td>off\\<\/td>sync\\<\/td>next\\<\/td>status`<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n

      Nested minecraft dqn train\/step\/observe from executive steps is blocked<\/strong><\/p>\n\n\n\n

      to avoid re-entrancy during auto-train.<\/p>\n\n\n\n

      Safe training ladder<\/h2>\n\n\n\n
        \n
      • Math smoke:<\/strong> nn demo, rl cartpole 50<\/li>\n\n\n\n
      • Curriculum smoke:<\/strong> rl curriculum smoke all 8 96<\/li>\n\n\n\n
      • Domain dry-run:<\/strong> observe \/ dry-run only (no live keys or RCON write)<\/li>\n\n\n\n
      • Short train:<\/strong> small episode count; watch status\/loss<\/li>\n\n\n\n
      • Checkpoint:<\/strong> rl net save \/ domain save<\/li>\n\n\n\n
      • Eval gate:<\/strong> promote only if metrics improve<\/li>\n\n\n\n
      • Live:<\/strong> deployment stages via rl real stage \u2026<\/li>\n<\/ul>\n\n\n\n

        Never jump from cold start to live desktop, medical, legal, finance, or public<\/p>\n\n\n\n

        multiplayer bots.<\/p>\n\n\n\n

        RSI and Lua<\/h2>\n\n\n\n

        RSI scripts get MathNN + executive Lua globals, for example:<\/p>\n\n\n\n

        [lua]
        <\/em>ai_nn_modes()
        ai_nn_policy_enable(“life”, “dueling-dqn”, 16, 8)
        rsi_sync_executive_plan(false)
        rsi_advance_executive_plan(1)
        executive_status()
        ai_rl_status()<\/p>\n\n\n\n

        RSI cycles can seed and advance executive plans while the CommandRouter runs<\/p>\n\n\n\n

        perception\/reasoning commands. See RSI scripts under lua_scripts\/RSI-*.lua.<\/p>\n\n\n\n

        File layout (code)<\/h2>\n\n\n\n
        Area<\/strong><\/td>Files<\/strong><\/td><\/tr>
        Substrate<\/td>RLCore, RLAlgorithms, RLExplore, RLFeatures, RLCurriculumOrchestrator, RLDeployment, RLReal, RLRealEnvs<\/td><\/tr>
        Neural<\/td>MathNN*, RLNeural, RLAgentNet, MathNNAgentBridge, MathNNIntegration<\/td><\/tr>
        Domains<\/td>MinecraftDQN, GamingRL, DrivingRL, KnowledgeAcquisitionRL, LifeRL, ChildPsychologyParentingRL, PoliticalRL<\/td><\/tr>
        Planning<\/td>Executive Functions<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n

        Troubleshooting<\/h2>\n\n\n\n
        Symptom<\/strong><\/td>Check<\/strong><\/td><\/tr>
        rl net not recognized<\/td>Ensure rl routes via _RLFeatures (not curriculum-only).<\/td><\/tr>
        Net resets every tick<\/td>Agent must pass a stable<\/strong> rl_mode to EnsureConfigured (Minecraft uses dueling-dqn).<\/td><\/tr>
        No learning<\/td>Bad state\/reward\/terminal wiring first; then LR\/epsilon.<\/td><\/tr>
        Minecraft walks into walls<\/td>Wall-guard, bridge collision features, dry-run; not only network size.<\/td><\/tr>
        Executive flips train off<\/td>Mode-changing steps are blocked; use plan detail to inspect steps.<\/td><\/tr>
        Life RL \u201cdoes nothing real\u201d<\/td>By design \u2014 advisory only.<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n

        Version note<\/h2>\n\n\n\n

        Documentation updated for the MathNN multi-learner registry, Life RL, Minecraft<\/p>\n\n\n\n

        executive plan hooks, RSI executive integration, and unified rl command<\/p>\n\n\n\n

        routing (2026).<\/p>\n","protected":false},"excerpt":{"rendered":"

        This is the operator hub for RL: what exists, how pieces fit together, and which docs\/commands to use first. Related maps: Doc Focus AugmentedIntelligenceSystems.md (AugmentedIntelligenceSystems.md) Whole-app system map RLCurriculum.md (RLCurriculum.md) Domain catalog, smoke tests, deployment ladder RLDomains.md (RLDomains.md) Expanded domain catalog detail MathNN.md (MathNN.md) In-process neural nets + rl net \/ learners LifeRL.md (LifeRL.md) Advisory […]<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[],"class_list":["post-2056","post","type-post","status-publish","format-standard","hentry","category-uncategorized"],"_links":{"self":[{"href":"http:\/\/macdaddy4sure.ai\/index.php\/wp-json\/wp\/v2\/posts\/2056","targetHints":{"allow":["GET"]}}],"collection":[{"href":"http:\/\/macdaddy4sure.ai\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"http:\/\/macdaddy4sure.ai\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"http:\/\/macdaddy4sure.ai\/index.php\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"http:\/\/macdaddy4sure.ai\/index.php\/wp-json\/wp\/v2\/comments?post=2056"}],"version-history":[{"count":1,"href":"http:\/\/macdaddy4sure.ai\/index.php\/wp-json\/wp\/v2\/posts\/2056\/revisions"}],"predecessor-version":[{"id":2057,"href":"http:\/\/macdaddy4sure.ai\/index.php\/wp-json\/wp\/v2\/posts\/2056\/revisions\/2057"}],"wp:attachment":[{"href":"http:\/\/macdaddy4sure.ai\/index.php\/wp-json\/wp\/v2\/media?parent=2056"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"http:\/\/macdaddy4sure.ai\/index.php\/wp-json\/wp\/v2\/categories?post=2056"},{"taxonomy":"post_tag","embeddable":true,"href":"http:\/\/macdaddy4sure.ai\/index.php\/wp-json\/wp\/v2\/tags?post=2056"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}