{ "schema_version": 1, "builder_plan": { "scenario": "Long customer conversations exhaust memory or limit concurrent requests.", "change": "Change only the attention cache policy, after checking model and runtime support.", "baseline": "The same model with its ordinary attention cache.", "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": "Peak cache memory, long-context answer quality, first-token latency, and requests completed.", "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": "2610.12514", "title": "CORAL-KV: Compact Cross-Layer Prediction and Residual Coding for Autoregressive Video KV Caches", "summary": "Coral-KV replaces clustering-based video KV-cache quantization (e.g., QVG-4/QVG-Pro) with cross-layer prediction: compact block-diagonal-plus-low-rank predictors estimate keys and values in later layers from earlier cache states, and only the residuals are quantized. The cost is fitting and storing the predictors, an optional per-video refit of block-diagonal components for higher fidelity, and added implementation complexity, with latency reported as comparable to baselines. It can fail if cross-layer dependencies are weak in a different model architecture, if residual coding error still compounds over very long generations, or if the reusable predictors do not transfer to videos outside the calibration distribution.", "rationale": "The method offers a better fidelity-memory trade-off than existing video KV-cache compression, but its gains are validated only on two specific autoregressive video diffusion models, so the reader should validate it on their own workload before production adoption.", "technique": "Cross-layer residual prediction for video KV caches", "family": "cache_kv", "recommendation": "testar", "published": "2026-10-07", "as_of": "2026-10-18", "editorial_status": "indexed", "source_url": "https://arxiv.org/abs/2610.12514", "independent_implementations": 1, "total_implementations": 2, "stars": 25, "citations": 0, "claims": [ { "claim_id": "claim-01", "statement": "Coral-KV replaces clustering-based video KV-cache quantization (e.g., QVG-4/QVG-Pro) with cross-layer prediction: compact block-diagonal-plus-low-rank predictors estimate keys and values in later layers from earlier cache states, and only the residuals are quantized. The cost is fitting and storing the predictors, an optional per-video refit of block-diagonal components for higher fidelity, and added implementation complexity, with latency reported as comparable to baselines. It can fail if cross-layer dependencies are weak in a different model architecture, if residual coding error still compounds over very long generations, or if the reusable predictors do not transfer to videos outside the calibration distribution.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2610.12514", "source_page": null, "source_excerpt": "" }, { "claim_id": "claim-02", "statement": "Reusable configuration improves PSNR by 0.32 dB (HY-WorldPlay) and 4.95 dB (MAGI-1) over QVG-4 while using 48% and 37% less cache memory; video-adapted configuration improves MAGI-1 PSNR by 3.77 dB over QVG-Pro with 33.7% less memory, at comparable generation latency.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2610.12514", "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": "" }