Section 7, paragraph 1The experiments do not show that an AlphaZero-style learner discovers the bypasses by self-play
Informal interpretation Motivation, interpretation or scope; not a formal claim.
1 Paper text
The stronger experimental goal would be to show that an AlphaZero-style learner discovers effective play through self-play once a bypass is available. We do not establish that here. The theoretical barrier concerns what the restricted network can represent, independently of training; removing it leaves open whether the required computation can be learned, with what data and under which training regime.
Generated from paper_jmlr.tex, lines 995–995.
2 Explain this
The paragraph states the limit of the experiments up front. The strong goal would be to show that an AlphaZero-style learner discovers good play by self-play once a bypass is available. That is not shown. What is reported is an intermediate step: supervised and imitation experiments that supply guidance which outcome-only self-play would have to generate for itself.
3 How this is formalised
A scope statement. The open question is recorded as Section 8.5, paragraph 3.
| Paper anchor | sec:experiments-p1 (website anchor; the manuscript has no label here) |
|---|---|
| Displayed as | Section 7, paragraph 1 |
| Claim | The experiments do not show that an AlphaZero-style learner discovers the bypasses by self-play |
| Coverage category | Informal interpretation |
| Statement source | paper_jmlr.tex lines 995–995; SHA-256 of the normalised block source 7c0f87c46de57174… |
| Explanations | explain/c-exp-goal, formal/c-exp-goal |
| Build evidence | fresh isolated rebuild, audit report |
4 Lean statement and source
No Lean statement corresponds to this claim. This is interpretation or motivation, not a formal mathematical claim.
Build evidence: fresh isolated rebuild of all 38 bundled modules with Lean (version 4.26.0, started 2026-10-10T15:21:56Z, finished 2026-10-10T15:27:00Z: 38/38 modules compiled, 1588 axiom reports, no errors. See Build for the commands and the independent audit of 2026-10-10 (dated evidence).