Result

Perturbench energy mmd (gene expression space) of CPA* on Frangieh21 (PerturBench v4 co-culture split)

2.1 ± 2 × 10 − 2 Perturbench energy mmd (gene expression space)

Methods

Tested configuration: CPA*
Protocol: PerturBench v4 covariate transfer from mono-culture to co-culture on Frangieh21
Dataset subset: Frangieh21 (PerturBench v4 co-culture split)
Procedure: PerturBench v4 covariate transfer from mono-culture to co-culture on Frangieh21
Evaluation: CPA ∗ on PerturBench v4 frangieh21
Coverage: Not reported scored / Not reported eligible
Uncertainty: SD 0.02
Evidence: Independent external evaluation · source checkedPerturBench: Benchmarking Machine Learning Models for Cellular Perturbation Analysis (arXiv v4) · Table 8, row 'CPA ∗', column 'MMD'

A source-checked result verifies the numerical transcription, not every model or protocol detail. Evaluation metadata: source checked. Source checked does not mean independently reproduced.

Reproduction

Split: Frangieh21 genetic perturbations in melanoma cells; 70% of perturbations in the co-culture condition held out.
Adaptation: Trained from scratch per dataset with HPO
Scoring implementation: PerturBench evaluation module (Section 4.2)

No execution recipe has been verified for this exact configuration and evaluation. A benchmark's general instructions may use different inputs, splits or model settings.

Reproducing this published result requires matching its model configuration, data, split and scorer. Source checking or a successful smoke test does not establish score reproduction.

Evidence

Source checking verifies the cited claim or transcription. It does not establish independent reproduction.

Evidence table

Inspect claims, sources and review details

Trace each statement to its source and review. A context-only reference supports the record generally; it does not verify an individual field. Source checking does not reproduce an experiment.

One row per statement and cited source. Multiple citations are not independent evaluations. Shared locators are labelled explicitly.

1 evidence row matching the loaded filters

Claims, original sources and review scope · Release 2026-10-23-5347695500ed
Property and statement	Original source and location	Review and provenance
Reported result 2.1 ± 2 × 10 − 2 Individual claims	PerturBench: Benchmarking Machine Learning Models for Cellular Perturbation Analysis (arXiv v4) Original source ↗ Table 8, row 'CPA ∗', column 'MMD' Version: arXiv 2408.10609v4, posted 2025-10-24 (NeurIPS 2025 Datasets and Benchmarks Track version) Retrieved: 2026-10-11T21:29:35Z	source checked ["source-hash-verification","deterministic-table-parse","independent-cell-check"] · 2026-10-12 independent paper Audit details Extraction: Parsed the pinned arXiv v4 HTML table markup with extract_perturbench_v4.py: header and row counts asserted per table, math-bold digits (best-in-column marks) normalised to plain digits, '± m × 10 − k' converted to a decimal standard deviation. Table 8, row 'CPA ∗', column 'MMD'. Independent review 2026-10-12: value, uncertainty, locator, metric, unit, direction and linked configuration and protocol checked by the reviewer's own parse of all Tables 2-13 cells and a hand check of 48 values against the v4 PDF. Corrections from the review applied by the extractor as listed in review.md. Field: `attributes.printed_value` Source artifact SHA-256: `b73f2644b129ca209b69812ad467d595e15c283f6cae5c3c9bbc5a4c5870996a` Hash scope: Exact bytes of the arXiv HTML rendering of v4 as retrieved; arXiv may regenerate HTML, so the hash describes this retrieval. Inspected artifact Extraction artifact SHA-256: `b73f2644b129ca209b69812ad467d595e15c283f6cae5c3c9bbc5a4c5870996a` Extraction artifact

How to read the evidence

Sources and history

Release 2026-10-23-5347695500ed · Record review: source checked

1 source record and release history

PerturBench: Benchmarking Machine Learning Models for Cellular Perturbation Analysis (arXiv v4) · Original source · arXiv 2408.10609v4, posted 2025-10-24 (NeurIPS 2025 Datasets and Benchmarks Track version)

Download this release (gzip)

Technical metadata and extraction receipts

Stable ID: lit-cells-networks-20261011-perturbench-v4-result-frangieh21-cpa-mmd-t8

metric: perturbench-energy-mmd
metric direction: lower
unit: unit-unreported
printed value: 2.1 ± 2 × 10 − 2
source locator: Table 8, row 'CPA ∗', column 'MMD'
review: method: source-hash-verification; deterministic-table-parse; independent-cell-check; method note: Parsed the pinned arXiv v4 HTML table markup with extract_perturbench_v4.py: header and row counts asserted per table, math-bold digits (best-in-column marks) normalised to plain digits, '± m × 10 − k' converted to a decimal standard deviation. Table 8, row 'CPA ∗', column 'MMD'; reviewer: claude; reviewer note: Claude review agent reviewer-cn-2, independent of the extractor lit-cells-networks; no human review claimed; date: 2026-10-12; artifact sha256: b73f2644b129ca209b69812ad467d595e15c283f6cae5c3c9bbc5a4c5870996a; retrieval url: https://arxiv.org/html/2408.10609v4; note: Extraction: Parsed the pinned arXiv v4 HTML table markup with extract_perturbench_v4.py: header and row counts asserted per table, math-bold digits (best-in-column marks) normalised to plain digits, '± m × 10 − k' converted to a decimal standard deviation. Table 8, row 'CPA ∗', column 'MMD'. Independent review 2026-10-12: value, uncertainty, locator, metric, unit, direction and linked configuration and protocol checked by the reviewer's own parse of all Tables 2-13 cells and a hand check of 48 values against the v4 PDF. Corrections from the review applied by the extractor as listed in review.md.
numeric value: 2.1
metric qualifier: gene expression space
uncertainty: type: standard_deviation; value: 0.02; printed: 2 × 10 − 2; n: 5; scope: across 5 training seeds (Table 2 caption)

Related records

evaluation: CPA ∗ on PerturBench v4 frangieh21

Suggest a correction