AGENT SKILL

factorized_score_composition

Compose F-NPSE and PF-NPSE posterior scores from learned score terms and prior corrections for SBI.

License 许可
Apache-2.0
最近更新
2026-09-03
Stars
★ 351
Forks
27

01 前言 Front Matter

name
factorized_score_composition
description
Compose F-NPSE and PF-NPSE posterior scores from learned score terms and prior corrections for SBI.

02 资源文件 2

03 SKILL.md 正文

Factorized Score Composition

Use this skill when implementing the central compositional step from F-NPSE or PF-NPSE. It turns score estimates for individual observations or small observation groups into one score for a multi-observation posterior approximation.

Do not use this skill for denoising score training or for sampling. It expects a score predictor to already exist and returns a composed score that a sampler can consume.

Inputs

  • Current theta.
  • Time index t and total diffusion levels T.
  • Observations for F-NPSE or groups for PF-NPSE.
  • Prior score function, such as -theta for a standard normal prior.
  • Score predictor for one observation or one group.

Outputs

  • Composed score vector.
  • Metadata with method, observation count, group count, and prior coefficient.
  • Per-term score diagnostics that can be logged for recovery evidence.

Workflow

For F-NPSE, compute:

((1 - n) * (T - t) / T) * grad_log_prior(theta) + sum_j s(theta, t, x_j)

For PF-NPSE, partition observations into groups of size at most m, compute k groups, and use the same formula with k replacing n and group score predictions replacing single-observation scores.

Validation

Run:

python /share/project/yuyang/workspace/Paper2Skills/Distiller/skills/module-to-skill/scripts/validate_skill_tree.py <skill_dir> --run-tests

The tests verify the paper formulas, grouping behavior, and invariance to group order for deterministic score terms.

Limitations

This skill does not decide whether score estimates are accurate. It only enforces the algebraic composition contract from the paper.

04 同仓库更多 Skills

VectorSpaceLab
VectorSpaceLab/AREX-Skill

passnet-triton-opt

Make a single PassNet Triton kernel fast AND numerically faithful: performance model (when a replacement can win at all), block/grid/warp tuning, autotune policy, launch overhead, and per-op numeric recipes to pass the dtype baseline tolerances. Use when a pass matches and is correct but speedup below expected, or when correctness fails by small numeric margins.

★ 351
VectorSpaceLab
VectorSpaceLab/AREX-Skill

passnet-solve

END-TO-END playbook for solving one PassNet sample: analyze the computation graphs, decide the optimization strategy (what to fuse, what to replace, what to leave alone), drive the iteration loop, and maximize the sample score. This is the ENTRY skill — invoke it first for any PassNet optimization task; it tells you when to use passnet-pattern-fusion, passnet-triton-opt and passnet-feedback.

★ 351
VectorSpaceLab
VectorSpaceLab/AREX-Skill

passnet-skill

PassNet GPU kernel optimization via compiler passes. Design and implement Triton-based optimization passes, create pass files under ./pass_dir/, self-evaluate with pass_evaluator, and iterate to maximize GPU speedup.

★ 351
VectorSpaceLab
VectorSpaceLab/AREX-Skill

passnet-pattern-fusion

Author PassNet pass files: write patterns that actually MATCH the FX graph, pick fusion regions (which ops to absorb into one Triton kernel), and structure multi-pass submissions with the mandatory shared-dispatch architecture. Use when creating or fixing pattern()/replacement_args()/replacement_func() files, when a pass "failed to match", or when deciding how to fuse multiple kernels/ops.

★ 351
VectorSpaceLab
VectorSpaceLab/AREX-Skill

passnet-orchestrate

Round/batch planning and decision support for PassNet graph optimization. Use for multi-sample triage, worker allocation, eval-budget policy, and post-eval keep/revert decisions. For solving one concrete sample end to end, use passnet-solve as the entry skill; consult this skill only when you need broader planning or a specific decision gate.

★ 351
VectorSpaceLab
VectorSpaceLab/AREX-Skill

passnet-feedback

Fast, deterministic feedback for PassNet work: pre-flight pattern/match verification WITHOUT burning a GPU evaluation, per-node bottleneck analysis of a sample's graphs, and eval-log parsing into per-variant status + estimated score + failure classification. Use BEFORE every GPU evaluation (check_pattern), at the START of a sample (analyze_graph), and AFTER every evaluation (parse_eval_log).

★ 351