{ "schema_version": 1, "builder_plan": { "scenario": "Your self-hosted model spends too long generating answers token by token.", "change": "Check whether a compatible draft model and serving integration exist before renting compute.", "baseline": "The same target with speculation off, then your existing drafter if you have one.", "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": "Accepted draft length, tokens per second, tail latency, output quality, and total GPU 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": "2610.20175", "title": "TreeSpark: Learning to Draft the Trees You Verify for Speculative Decoding", "summary": "TreeSpark replaces reference-conditioned drafter training in semi-autoregressive speculative decoding (DSpark) with parent-conditioned tree construction trained via survival NLL and a capped coverage hinge, reusing existing target distributions so no extra target forward passes are needed. The cost is a custom drafter training pipeline tied to a specific architecture and target model (Qwen3-4B/8B), plus added training complexity from replaying tree expansion, pruning, and packing. It can fail if the workload differs from the evaluated general and writing benchmarks, since acceptance-length gains depend on drafter-target alignment and may not transfer to other targets or to batched serving, where gains shrink to 12.6%-17.2%.", "rationale": "The throughput gain is real but modest, and adopting it requires training a custom drafter on a specific semi-autoregressive architecture, which is a significant investment for a constrained team compared with off-the-shelf speculative decoding solutions.", "technique": "Trained tree drafting for speculative decoding", "family": "decodificacao_especulativa", "recommendation": "observar", "published": "2026-10-14", "as_of": "2026-10-18", "editorial_status": "indexed", "source_url": "https://arxiv.org/abs/2610.20175", "independent_implementations": 0, "total_implementations": 0, "stars": 0, "citations": null, "claims": [ { "claim_id": "claim-01", "statement": "TreeSpark replaces reference-conditioned drafter training in semi-autoregressive speculative decoding (DSpark) with parent-conditioned tree construction trained via survival NLL and a capped coverage hinge, reusing existing target distributions so no extra target forward passes are needed. The cost is a custom drafter training pipeline tied to a specific architecture and target model (Qwen3-4B/8B), plus added training complexity from replaying tree expansion, pruning, and packing. It can fail if the workload differs from the evaluated general and writing benchmarks, since acceptance-length gains depend on drafter-target alignment and may not transfer to other targets or to batched serving, where gains shrink to 12.6%-17.2%.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2610.20175", "source_page": null, "source_excerpt": "" }, { "claim_id": "claim-02", "statement": "23.4% to 31.6% higher mean acceptance length and 15.1% to 25.0% higher single-request throughput over DSpark; the supervision alone adds about 4.2% to 5.0% throughput.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2610.20175", "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": "" }