Audit pass on long-standing research-note TODOs in `src/MATLAB/algorithms/unknown-models/SI.m`: - [ ] Line 196 (label `#1`): what is the difference between `b`, `bb`, and `B`? They appear equivalent in this formulation. - [ ] Line 251 (label `#3`): is the `ob` variable used for anything? - [ ] Line 283 (label `#4`): why is the backward smoothing useful here? - [ ] Line 295 (label `#5`): what is the meaning of the outer product in the context of this algorithm? - [ ] Line 105 (label `#6`): `b = B` — doesn't this erase any distinction set above? Is it not meant to be a uniform distribution? - [ ] Line 82 (label `#7`): `T = 27` — is this used anywhere useful, and what is it meant to be? - [ ] Line 266: is `6` here chosen simply as 6 timesteps to backward smooth over? Hyperparameter? - [ ] Line 359: why is the state-prediction-error `if` block in the planning loop when it's only relevant for tree search? Audit task — many of these may resolve to "this is correct, document why" rather than code changes.
Audit pass on long-standing research-note TODOs in
src/MATLAB/algorithms/unknown-models/SI.m:#1): what is the difference betweenb,bb, andB? They appear equivalent in this formulation.#3): is theobvariable used for anything?#4): why is the backward smoothing useful here?#5): what is the meaning of the outer product in the context of this algorithm?#6):b = B— doesn't this erase any distinction set above? Is it not meant to be a uniform distribution?#7):T = 27— is this used anywhere useful, and what is it meant to be?6here chosen simply as 6 timesteps to backward smooth over? Hyperparameter?ifblock in the planning loop when it's only relevant for tree search?Audit task — many of these may resolve to "this is correct, document why" rather than code changes.