yzj623/pplab/adaptive-ntk-weighting

Claude Code and agent skill by yzj623/pplab

Trust tier 3
  • community
Source check pending

Checking that the SKILL.md file is still in the repository.

Compute Algorithm 1 PINN loss weights from NTK trace ratios with degeneracy checks and diagnostics.

Install

npx vskill@latest install yzj623/pplab/adaptive-ntk-weighting

SKILL.md

The same SKILL.md was published earlier as yzj623/pplab/adaptive-ntk-weighting.

Adaptive NTK Weighting

Use this skill after computing PINN NTK block traces. It implements the trace-ratio weighting rule proposed for calibrating convergence rates of different loss components.

Inputs

  • trace_full: trace of the combined boundary-plus-residual NTK.
  • trace_kuu: boundary-output kernel trace.
  • trace_krr: residual-operator kernel trace.
  • Optional epsilon for rejecting degenerate traces.

Outputs

  • lambda_b = trace_full / trace_kuu.
  • lambda_r = trace_full / trace_krr.
  • A diagnostic indicating which component receives the stronger weight.

Workflow

  1. Validate traces are finite and nonnegative.
  2. Reject component traces below epsilon instead of silently dividing by zero.
  3. Compute trace-ratio weights exactly as Algorithm 1 states.
  4. Preserve component names so a recovery harness applies weights to the correct losses.

Validation

Run:

bash
python tests/test_adaptive_weights.py

The tests verify the paper's ratio formulas and that the smaller-trace component receives a larger weight.

Limitations

The skill does not tune learning rates or batch sizes. It only computes the NTK-based loss weights used by a downstream training loop.

Security

  • Security scanPassed, 90 of 100Pattern checks on the skill's files.
  • Model reviewPassed

Quality evaluation

  • Task evaluationNot runNo evaluation yet. Run one in Skill Studio.

Source

Popularity

7-day trend score
5

Not refreshed yet.

Works with

Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.

  • Claude Code
  • Cursor
  • GitHub Copilot
  • Windsurf
  • Codex
  • Gemini CLI
  • Cline
  • Amp
  • Roo Code
  • Goose
  • Continue
  • Kilo Code
  • Junie
  • Trae
  • OpenHands
  • Qwen Code
  • AI Agent
  • Tabnine Chat
  • Sourcegraph Cody
  • Amazon Q
  • Firebase Genkit
  • Sweep
  • Devin
  • Phind
  • Mentat
  • Claude Desktop
  • Aider
  • GPT Pilot
  • Plandex
  • bolt.new
  • v0
  • Replit Agent
  • Anse
  • TypingMind
  • BetterChatGPT
  • LibreChat
  • Warp AI
  • Pieces
  • Fig AI
Embed badge
vskill VERIFIED badge
[![vskill VERIFIED](https://verified-skill.com/api/v1/skills/yzj623/pplab/adaptive-ntk-weighting/badge)](https://verified-skill.com/skills/yzj623/pplab/adaptive-ntk-weighting)

Something wrong with this skill or its page? Tell us.