From a501a932237bb5079853d6a73cf7cae40c103e6b Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?=EC=A0=95=EC=8B=9C=EC=9B=90?= Date: Mon, 11 May 2026 07:03:45 +0900 Subject: [PATCH] Cycle 4: revert use_rollout_value=true, scale network 512x3 -> 768x4 MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit c3 showed use_rollout_value=false alone is not the fix: without the heuristic rollout safety net the random-init network value gives bad MCTS Q early on, the agent plays more (PA 0.10 -> 0.22+) but eats more -20 expedition penalties (score worsened from -57 to -97 avg). Mathematical intuition: in Lost Cities, opening an expedition is a 20-point commitment. Break-even requires rank-sum × (handshakes+1) >= 20. The model has to learn: - which colors to open (based on hand handshake/high-rank holdings) - when to commit vs discard - card-ordering constraints (ascending only) This is a moderately rich value function. 512x3 (~800k params, ~290 input dim) might be undersized. Test capacity hypothesis with 768x4 (~2M params) while keeping the rollout safety net so MCTS Q stays competent. Other params from c1 kept: c_puct=5, virtual_loss=5, dirichlet α=0.3/ε=0.4, parallel_simulations=64, n_simulations=50. --- configs/ismcts/default.yaml | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/configs/ismcts/default.yaml b/configs/ismcts/default.yaml index 7ef5ee1..d9b8e28 100644 --- a/configs/ismcts/default.yaml +++ b/configs/ismcts/default.yaml @@ -17,8 +17,8 @@ encoding: slot_aware_playability: true network: kind: mlp - hidden_size: 512 - num_layers: 3 + hidden_size: 768 + num_layers: 4 activation: relu mcts: n_simulations: 50 @@ -28,7 +28,7 @@ mcts: virtual_loss_value: 5.0 eval_n_simulations: 16 rollout_policy: heuristic_balanced - use_rollout_value: false + use_rollout_value: true root_dirichlet_alpha: 0.3 root_dirichlet_epsilon: 0.4 temperature: