AXIOM BOINC SESSION RESULTS LOG
Session Date: 2026-03-04 05:07 (America/Denver)
Task: Part 3 - Save, Git Prior-Art Timestamp, Upload, Archive

PART 1 VALIDATION / CREDIT SUMMARY (from validate_2026-03-04_0501.txt)
- Completed experiment results credited this session: 251
- Total credit awarded this session: 1965 (session cap check PASS, < 10,000)
- Residual uncredited completed rows after batch: 4 (new arrivals after cutoff)
- Bulk behavior reviewed: mostly successful ~830-900 second runs; upload JSON artifacts missing at audit time.

Credit additions by contributor:
- Amapola: +1313
- Orange Kid: +267
- ChelseaOilman: +199
- Steve Dodd: +96
- Remaining contributors combined: +90

CLEANUP / STABILITY ACTIONS
- Broken families monitored (no active rows to abort in this pass):
  1) oscillatory_roughchannel_lbm_resonance
  2) abx_cycle_hgt_delay_resonance
  3) potts_pulse_anneal_resonance
  4) spatial_pgg_delay_fatigue
- Stuck-task cleanup checks: 0 rows matched (>12h dead-host rule), 0 rows matched (>48h hard ceiling).
- No safe targeted patch applied this pass: failures remain outcome=5, exit_status=0, empty stderr signature.

PART 2 DEPLOYMENT / RESEARCH SUMMARY (from run_2026-03-03_1352.log)
- Retirement pass executed: ABORT_TOTAL=0 (no newly abortable unsent tasks in retirement candidates at that checkpoint).

CPU DEPLOYMENT
- CPU deployment pass completed.
- Hosts considered for queue fill: 81
- Hosts skipped for low RAM (<6GB): 2
- CPU workunits created: 2937
- CPU scripts used in deployment pass:
  1) wd_batchnoise_interaction.py
  2) wd_labelsmooth_interaction.py

GPU DEPLOYMENT CHECKPOINT
- Run log shows GPU deployment pass started with scripts:
  1) wd_curvature_trigger_gpu.py
  2) wd_timing_scale_gpu.py
- The captured run log ended with an interrupt (^C) before GPU summary counters were printed.
- Live checkpoint query after this session:
  - GPU hosts with queued experiment tasks: 11
  - Queued GPU experiment tasks: 36
  - Total result rows currently present for wd_curvature_trigger_gpu*: 47
  - Total result rows currently present for wd_timing_scale_gpu*: 37
  - Active queued GPU tasks observed on hosts: 355, 57, 1, 159, 299

NEW/UPDATED EXPERIMENT DESIGN + NOVELTY CHECK NOTES
- New script written and compiled in Part 2:
  1) wd_batchnoise_interaction.py (py_compile OK)
- Existing active scripts reviewed/deployed:
  1) wd_labelsmooth_interaction.py
  2) wd_curvature_trigger_gpu.py
  3) wd_timing_scale_gpu.py
- Novelty-check evidence recorded in run log via literature/web search terms:
  - "weight decay batch size interaction neural networks"
  - "arxiv weight decay batch size interaction deep learning"
  - "site:arxiv.org weight decay label smoothing interaction"
  - "Scheduled Weight Decay paper arxiv 2021"
- Novel hypothesis focus retained: interaction effects for late-WD timing versus gradient-noise/batch regime and confidence-smoothing regime, plus GPU curvature-triggered onset tests.

KEY SCIENTIFIC FINDINGS
1. High-throughput volunteer compute remains stable at runtime level (dominant successful completions in the ~13-15 minute window), supporting continued large-batch exploration.
2. Reliability bottleneck remains the primary blocker: many successful terminal rows still lack recoverable upload JSON payloads, limiting scientific extraction despite valid compute donation.
3. Four failure-prone experiment families remain effectively paused/contained this session, with no additional active rows requiring abort.
4. The active research direction is now explicitly interaction-centric (late WD x batch-noise regime, late WD x label smoothing, and GPU curvature-triggered onset), with a new CPU interaction script added and compiled.

FILES USED FOR THIS SESSION LOG
- C:\GPT3.5PrmtScpts\ExperimentResults\validate_2026-03-04_0501.txt
- C:\GPT3.5PrmtScpts\AutoReviewLogs\run_2026-03-03_1352.log
- C:\GPT3.5PrmtScpts\ExperimentResults\findings_summary.txt