page: title, subtitle and abstract from the final paper

index.html

@@ -3,7 +3,7 @@
 <head>
   <meta charset="utf-8">
   <meta name="viewport" content="width=device-width, initial-scale=1">
-  <title>RSIGame: Controllable Agentic Game Development</title>
+  <title>RSIGame: Autonomous Agentic Game Development</title>
   <meta name="description" content="RSIGame turns open-ended game improvement into goal-oriented, state-controlled progressive development through agentic verification.">
   <link rel="preconnect" href="https://fonts.googleapis.com">
   <link rel="preconnect" href="https://fonts.gstatic.com" crossorigin>
@@ -34,7 +34,7 @@
   <div class="hero-inner">
     <p class="eyebrow">Agentic verification · Progressive development</p>
     <h1>RSIGame</h1>
-    <p class="hero-sub">Controllable Agentic Game Development with Recursive Self-Improvement</p>
+    <p class="hero-sub">Autonomous Agentic Game Development with Recursive Self-improvement</p>
     <p class="authors">Authors to be announced</p>
     <div class="hero-cta">
       <a class="btn btn-primary" href="#compare">Watch games evolve</a>
@@ -57,29 +57,9 @@
     <div class="overview">
       <div>
         <h2>Abstract</h2>
-        <p>Game development is a long-horizon, multi-objective process that requires repeated
-           playtesting, revision, and validation rather than a single generation pass. While recent
-           language-model agents can generate playable games and iteratively refine them, reliably
-           controlling such improvement over extended development remains challenging. An autonomous
-           developer must continuously identify where meaningful improvement headroom remains, gather
-           interaction evidence, edit the game, verify the consequences, and determine whether further
-           development is still worthwhile.</p>
-        <p>We find that a multi-agent loop of active game exploration, evidence-grounded editing, and
-           independent agentic verification can substantially improve generated games. However,
-           unguided self-improvement eventually plateaus, and continued refinement beyond that point
-           can even lead to regressions. We therefore introduce <strong>RSIGame</strong>, a controllable
-           framework for recursive game self-improvement. RSIGame autonomously explores and tests the
-           executable game, grounds improvement objectives in observed behavior, edits the game, and
-           verifies each new version before continuing development. A Global Quality Monitor
-           periodically quantifies development progress, preserves the globally best checkpoint, and
-           detects emerging saturation. Rather than treating saturation as termination, RSIGame can
-           introduce sparse high-level guidance from a human or a strong model, opening new directions
-           for progressive, multi-stage game evolution.</p>
-        <p>To support both game generation and experience internalization, we further construct a
-           large-scale game-development corpus containing <b>2,213</b> generation trajectories,
-           <b>2,108</b> planning trajectories, and <b>2,013</b> verified improvement trajectories.
-           Across 140 GameCraft-Bench tasks, two game engines, and five generators, RSIGame
-           consistently improves game quality under matched development budgets.</p></div>
+        <p>Recent advances in large language models have made automatic game generation increasingly feasible, yet reliably improving generated games beyond a playable version remains challenging. Naive iterative refinement can easily overfit a small set of test cases, producing fragile games with unresolved bugs, missing behaviors, and poor generalization to broader player interactions.</p>
+        <p>We introduce <strong>RSIGame</strong>, an autonomous agentic game development framework with recursive self-improvement. RSIGame organizes development into complementary local and global loops. Concretely, a local <em>explore-diagnose-improve</em> loop broadly explores the executable game, diagnoses and prioritizes discovered issues, and performs evidence-grounded revision, where an evolving checklist continually accumulates new testing and improvement guidance. A global loop tracks overall quality, preserves the best checkpoint, and detects saturation or regression over long-horizon development. Beyond test-time improvement, RSIGame further internalizes successful development experience into the generator through training.</p>
+        <p>Across 140 GameCraft-Bench tasks, two game engines, and five generators, RSIGame consistently improves game quality under matched development budgets. Notably, iterative development with Qwen3.8-27B reaches 53.90 on Godot and 50.24 on Phaser, surpassing the one-shot performance of substantially stronger GPT-5.5-based generators.</p></div>
       <aside class="loopcard">
         <h3>How a game develops</h3>
         <section class="loopbox inner">
@@ -285,7 +265,7 @@ const PLAY_GAMES = [
   <div class="wrap">
     <header class="section-head"><h2>Citation</h2></header>
     <pre class="bibtex"><button class="copy" id="copy-bib">Copy</button><code id="bib">@inproceedings{rsigame2027,
-  title  = {RSIGame: Controllable Agentic Game Development with Recursive Self-Improvement},
+  title  = {RSIGame: Autonomous Agentic Game Development with Recursive Self-improvement},
   author = {TBA},
   year   = {2027}
 }</code></pre>