Skip to content

[qmc] frustration-free: symmetric NQS for chiral graviton - #277

Open
jiangweiqi001 wants to merge 1 commit into
QuantumBFS:mainfrom
jiangweiqi001:challenge/qmc-chiral-graviton
Open

[qmc] frustration-free: symmetric NQS for chiral graviton#277
jiangweiqi001 wants to merge 1 commit into
QuantumBFS:mainfrom
jiangweiqi001:challenge/qmc-chiral-graviton

Conversation

@jiangweiqi001

Copy link
Copy Markdown

Summary

  • add an antisymmetric, exact-SU(2)-projected Pfaffian NQS and finite-sphere ED/VMC acceptance framework for Challenge [challenge]: Symmetric neural-network ansatz for the chiral graviton in ν = 1/3 fractional quantum Hall state #15
  • add auditable production orchestration, chiral-response operators, and Torch/DCU backend primitives
  • include a Chinese evidence-backed report with deterministic SVGs, verified N=6 CPU smoke coverage, and the successful Torch Pfaffian DCU smoke
  • clearly record that final N=6–8 scientific acceptance, multiplet certification, and chirality production results remain pending

Test plan

  • deterministic report SVG regeneration and compact-manifest/raw-artifact consistency checks
  • focused JAX, chiral-response, artifact/schema, and Torch primitive suites recorded in the report
  • Torch Pfaffian target smoke passed on one BW DCU with Torch 2.4.1/HIP 6.1
  • complete N=6–8 production VMC/exact evaluation and final scientific acceptance gates
  • resolve the Torch projector checkpoint/JVP target-runtime compatibility item

Submission note

This is a standalone Challenge #15 submission and does not modify or merge into PR #152.

Made with Cursor

Provide the projected-Pfaffian implementation, auditable production workflow, chiral-response and DCU progress, plus an evidence-backed Chinese report that clearly marks pending scientific acceptance gates.

Co-authored-by: Cursor <cursoragent@cursor.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant