{ "schema_version": 1, "builder_plan": { "scenario": "A self-hosted model uses more compute or memory than your workload can justify.", "change": "Check for a usable pruned checkpoint and runtime support before attempting model surgery.", "baseline": "The original unpruned model at the same precision, context, and load.", "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": "Task quality, actual GPU memory, serving throughput, and preparation cost.", "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": "2610.02951", "title": "Dynamic Expert Pruning for Multi-Agent Systems", "summary": "DEP replaces static offline-calibrated expert pruning masks for MoE models with a per-request mask predicted from the agent's system and task prompts by a lightweight predictor trained once on workflow transcripts. It costs one extra predictor forward pass per request plus the training data and pipeline for the predictor, while still requiring an MoE backbone whose active subset must fit on the accelerator. It can fail if the workload is not prompt-differentiable (homogeneous tasks gain little), if serving infrastructure does not support per-request expert loading without prohibitive latency, or if production prompts drift from the transcript distribution the predictor was trained on.", "rationale": "The technique targets MoE backbones too large for a single 24 GB GPU or API-based serving, so the reader lacks the hardware and workload control to deploy it now, though it is worth tracking for future self-hosted MoE serving.", "technique": "Prompt-conditioned dynamic expert pruning for MoE", "family": "esparsidade_e_poda", "recommendation": "observar", "published": "2026-10-02", "as_of": "2026-10-18", "editorial_status": "indexed", "source_url": "https://arxiv.org/abs/2610.02951", "independent_implementations": 0, "total_implementations": 1, "stars": 10, "citations": null, "claims": [ { "claim_id": "claim-01", "statement": "DEP replaces static offline-calibrated expert pruning masks for MoE models with a per-request mask predicted from the agent's system and task prompts by a lightweight predictor trained once on workflow transcripts. It costs one extra predictor forward pass per request plus the training data and pipeline for the predictor, while still requiring an MoE backbone whose active subset must fit on the accelerator. It can fail if the workload is not prompt-differentiable (homogeneous tasks gain little), if serving infrastructure does not support per-request expert loading without prohibitive latency, or if production prompts drift from the transcript distribution the predictor was trained on.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2610.02951", "source_page": null, "source_excerpt": "" }, { "claim_id": "claim-02", "statement": "DEP achieves better overall accuracy than static pruning and merging baselines, with the largest margin when few experts are retained, and generalizes to unseen workflows without retraining; no quantified figure is given in the abstract.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2610.02951", "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": "" }