File size: 3,425 Bytes
efb9dfc | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 | {
"schema_version": 1,
"title": "Reproduction: Maximum Likelihood Reinforcement Learning",
"emoji": "🎯",
"space_id": "Umong/repro-maximum-likelihood-reinforcement-learning",
"paper": {
"arxiv_id": "2602.02710"
},
"tags": [
"icml2026-repro",
"paper-EeuLO2BjFN"
],
"updated_at": "2026-07-18T18:54:07+00:00",
"root": {
"slug": "index",
"title": "Reproduction: Maximum Likelihood Reinforcement Learning",
"file": "pages/index.md",
"children": [
{
"slug": "executive-summary",
"title": "Executive summary",
"file": "pages/executive-summary/page.md",
"children": []
},
{
"slug": "claim-1-maxrl-defines-a-compute-indexed-family-of-sample-based-objectives-that-interpolates-between-standard-rl-and-exact-maximum-likelihood-as-sampling-compute-increases-abstract",
"title": "Claim 1: MaxRL defines a compute-indexed family of sample-based objectives that interpolates between standard RL and exact maximum likelihood as sampling compute increases (Abstract).",
"file": "pages/claim-1-maxrl-defines-a-compute-indexed-family-of-sample-based-objectives-that-interpolates-between-standard-rl-and-exact-maximum-likelihood-as-sampling-compute-increases-abstract/page.md",
"children": []
},
{
"slug": "claim-2-the-maxrl-objectives-admit-a-simple-unbiased-policy-gradient-estimator-for-non-differentiable-sampling-settings-abstract",
"title": "Claim 2: The MaxRL objectives admit a simple unbiased policy-gradient estimator for non-differentiable sampling settings (Abstract).",
"file": "pages/claim-2-the-maxrl-objectives-admit-a-simple-unbiased-policy-gradient-estimator-for-non-differentiable-sampling-settings-abstract/page.md",
"children": []
},
{
"slug": "claim-3-the-paper-claims-maxrl-converges-to-maximum-likelihood-optimization-in-the-infinite-compute-limit-abstract",
"title": "Claim 3: The paper claims MaxRL converges to maximum-likelihood optimization in the infinite-compute limit (Abstract).",
"file": "pages/claim-3-the-paper-claims-maxrl-converges-to-maximum-likelihood-optimization-in-the-infinite-compute-limit-abstract/page.md",
"children": []
},
{
"slug": "claim-4-empirically-maxrl-pareto-dominates-tested-existing-methods-across-all-evaluated-models-and-tasks-abstract",
"title": "Claim 4: Empirically, MaxRL Pareto-dominates tested existing methods across all evaluated models and tasks (Abstract).",
"file": "pages/claim-4-empirically-maxrl-pareto-dominates-tested-existing-methods-across-all-evaluated-models-and-tasks-abstract/page.md",
"children": []
},
{
"slug": "claim-5-maxrl-reports-up-to-20x-test-time-scaling-efficiency-gains-compared-with-a-grpo-trained-counterpart-abstract",
"title": "Claim 5: MaxRL reports up to 20x test-time scaling efficiency gains compared with a GRPO-trained counterpart (Abstract).",
"file": "pages/claim-5-maxrl-reports-up-to-20x-test-time-scaling-efficiency-gains-compared-with-a-grpo-trained-counterpart-abstract/page.md",
"children": []
},
{
"slug": "conclusion",
"title": "Conclusion",
"file": "pages/conclusion/page.md",
"children": []
}
]
},
"agent_view_tokens": 3376,
"revision": "1784400847404442706"
} |