{ "schema_version": 1, "builder_plan": { "scenario": "You already need task-specific training and want to reduce its memory or compute cost.", "change": "Change only the training technique on a small controlled run after checking prerequisites.", "baseline": "Your current training recipe with the same model, data split, and evaluation.", "steps": [ "Locate the paper's implementation and check its license, model support, data needs, and hardware requirements. Stop here if you cannot obtain the required artifacts.", "Save a small set of representative product tasks and expected outcomes. Keep evaluation tasks out of any training or tuning, and choose quality and cost limits before running.", "Run the baseline and the proposed change on identical inputs. Keep other settings fixed and repeat paired runs to expose variation.", "Save per-task outputs, failures, timings, and all usage costs. Review regressions before expanding the test." ], "measure": "Held-out quality, peak memory, training time, and full compute cost to the target quality.", "decision": "Keep the change only if it meets your preselected quality and latency limits and improves the metric you care about after setup and operating costs. Otherwise retain the baseline. This is an exploration plan, not a validated recommendation." }, "builder_review": null, "arxiv_id": "2609.39137", "title": "ID Balancing: Stable Training of Extremely Sparse MoE via PID-Based Load Control", "summary": "ID Balancing replaces auxiliary-load-balancing losses and prior loss-free methods (DeepSeek-style integral control, Kimi-style proportional control) with an Integral-Derivative controller that updates router bias terms during large-scale MoE pretraining. It costs negligible compute, adding only control updates to routing biases, but presupposes training MoE models with hundreds of experts across many GPUs. It fails to be relevant below foundation-model-training scale: it addresses expert load imbalance during multi-billion-parameter distributed training, a problem that does not exist when fine-tuning or serving on a single 24 GB GPU or consuming third-party APIs.", "rationale": "The technique is an MoE pretraining-stability method that requires distributed training of tens of billions of parameters, which is explicitly outside the reader's infrastructure and production decisions.", "technique": "PID-based auxiliary-loss-free MoE load balancing", "family": "treino_eficiente", "recommendation": "nao_aplica", "published": "2026-09-30", "as_of": "2026-10-18", "editorial_status": "indexed", "source_url": "https://arxiv.org/abs/2609.39137", "independent_implementations": 0, "total_implementations": 0, "stars": 0, "citations": 0, "claims": [ { "claim_id": "claim-01", "statement": "ID Balancing replaces auxiliary-load-balancing losses and prior loss-free methods (DeepSeek-style integral control, Kimi-style proportional control) with an Integral-Derivative controller that updates router bias terms during large-scale MoE pretraining. It costs negligible compute, adding only control updates to routing biases, but presupposes training MoE models with hundreds of experts across many GPUs. It fails to be relevant below foundation-model-training scale: it addresses expert load imbalance during multi-billion-parameter distributed training, a problem that does not exist when fine-tuning or serving on a single 24 GB GPU or consuming third-party APIs.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2609.39137", "source_page": null, "source_excerpt": "" }, { "claim_id": "claim-02", "statement": "Over 50% reduction in worst-case backbone MaxVio and 12% in training-average MinVio versus the best baselines at Top-3-of-768; at 18.9B to 69.9B parameters (Top-10-of-768), worst-case MaxVio stays nearly constant and is approximately 89.6% lower than the auxiliary-loss baseline.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2609.39137", "source_page": null, "source_excerpt": "" } ], "exposure_map": [ { "dimension": "quality", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "compute", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "latency", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "operations", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "compatibility", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "security", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "data_and_training", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "reproducibility", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" } ], "risks": [], "minimum_test": [], "open_questions": [], "independent_tests": [], "lab_reproductions": [], "report_available": false, "validation_tier": "unknown", "evidence_tier": "unknown", "infrastructure_basis": "unknown", "equations": [], "equations_status": "not_fetched", "core_kind": "", "equations_fetched_at": "", "equations_source_url": "", "equations_source_sha256": "", "exposures_reviewed_at": "", "exposures_source_sha256": "" }