Method

CNN

Baseline implemented by the Therapeutics Data Commons authors over the stated molecular featurisation.

Overview

Baseline implemented by the Therapeutics Data Commons authors over the stated molecular featurisation.

Consult the linked sources for architecture or protocol details. Missing evidence is not evidence of a missing capability.

Evaluations and results

22 evaluations · 22 results. Different protocols are not a single leaderboard.

Filter evaluations

Tested model or methodProtocolDatasetEvaluation setupMetricEvidence origin

Applied filters: All linked evaluations

Exact evaluated configurations and original reported results
Tested configuration	Protocol and dataset	Finding	Evidence and details
Method: CNN	Task: TDC ADMET benchmark group TDC-AMES: Toxicity: TDC.AMES Dataset subset: TDC.AMES (TDC ADMET benchmark group split)	0.776 ± 0.015 auroc fraction · higher Uncertainty: type: reported_plus_minus_type_unresolved; value: 0.015 Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source CNN on TDC ADMET benchmark group TDC-AMES: Toxicity: TDC.AMES Scaffold split, as the ADMET benchmark group defines it. Table 4 reports a value and a plus-or-minus spread but does not identify the spread's statistical meaning. Aggregation: Not reported Therapeutics Data Commons: Machine Learning Datasets and Tasks for Drug Discovery and Development · Table 4, block(Toxicity), row(CNN), column(TDC.AMES)
Method: CNN	Task: TDC ADMET benchmark group TDC-AQSOL: Absorption: TDC.AqSol Dataset subset: TDC.AqSol (TDC ADMET benchmark group split)	1.023 ± 0.023 mae error · lower Uncertainty: type: reported_plus_minus_type_unresolved; value: 0.023 Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source CNN on TDC ADMET benchmark group TDC-AQSOL: Absorption: TDC.AqSol Scaffold split, as the ADMET benchmark group defines it. Table 4 reports a value and a plus-or-minus spread but does not identify the spread's statistical meaning. Aggregation: Not reported Therapeutics Data Commons: Machine Learning Datasets and Tasks for Drug Discovery and Development · Table 4, block(Absorption), row(CNN), column(TDC.AqSol)
Method: CNN	Task: TDC ADMET benchmark group TDC-BBB: Distribution: TDC.BBB Dataset subset: TDC.BBB (TDC ADMET benchmark group split)	0.781 ± 0.030 auroc fraction · higher Uncertainty: type: reported_plus_minus_type_unresolved; value: 0.030 Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source CNN on TDC ADMET benchmark group TDC-BBB: Distribution: TDC.BBB Scaffold split, as the ADMET benchmark group defines it. Table 4 reports a value and a plus-or-minus spread but does not identify the spread's statistical meaning. Aggregation: Not reported Therapeutics Data Commons: Machine Learning Datasets and Tasks for Drug Discovery and Development · Table 4, block(Distribution), row(CNN), column(TDC.BBB)
Method: CNN	Task: TDC ADMET benchmark group TDC-BIOAV: Absorption: TDC.Bioav Dataset subset: TDC.Bioav (TDC ADMET benchmark group split)	0.613 ± 0.013 auroc fraction · higher Uncertainty: type: reported_plus_minus_type_unresolved; value: 0.013 Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source CNN on TDC ADMET benchmark group TDC-BIOAV: Absorption: TDC.Bioav Scaffold split, as the ADMET benchmark group defines it. Table 4 reports a value and a plus-or-minus spread but does not identify the spread's statistical meaning. Aggregation: Not reported Therapeutics Data Commons: Machine Learning Datasets and Tasks for Drug Discovery and Development · Table 4, block(Absorption), row(CNN), column(TDC.Bioav)
Method: CNN	Task: TDC ADMET benchmark group TDC-CACO2: Absorption: TDC.Caco2 Dataset subset: TDC.Caco2 (TDC ADMET benchmark group split)	0.446 ± 0.036 mae error · lower Uncertainty: type: reported_plus_minus_type_unresolved; value: 0.036 Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source CNN on TDC ADMET benchmark group TDC-CACO2: Absorption: TDC.Caco2 Scaffold split, as the ADMET benchmark group defines it. Table 4 reports a value and a plus-or-minus spread but does not identify the spread's statistical meaning. Aggregation: Not reported Therapeutics Data Commons: Machine Learning Datasets and Tasks for Drug Discovery and Development · Table 4, block(Absorption), row(CNN), column(TDC.Caco2)
Method: CNN	Task: TDC ADMET benchmark group TDC-CL-HEPA: Excretion: TDC.CL-Hepa Dataset subset: TDC.CL-Hepa (TDC ADMET benchmark group split)	0.235 ± 0.021 spearman correlation · higher Uncertainty: type: reported_plus_minus_type_unresolved; value: 0.021 Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source CNN on TDC ADMET benchmark group TDC-CL-HEPA: Excretion: TDC.CL-Hepa Scaffold split, as the ADMET benchmark group defines it. Table 4 reports a value and a plus-or-minus spread but does not identify the spread's statistical meaning. Aggregation: Not reported Therapeutics Data Commons: Machine Learning Datasets and Tasks for Drug Discovery and Development · Table 4, block(Excretion), row(CNN), column(TDC.CL-Hepa)
Method: CNN	Task: TDC ADMET benchmark group TDC-CL-MICRO: Excretion: TDC.CL-Micro Dataset subset: TDC.CL-Micro (TDC ADMET benchmark group split)	0.252 ± 0.116 spearman correlation · higher Uncertainty: type: reported_plus_minus_type_unresolved; value: 0.116 Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source CNN on TDC ADMET benchmark group TDC-CL-MICRO: Excretion: TDC.CL-Micro Scaffold split, as the ADMET benchmark group defines it. Table 4 reports a value and a plus-or-minus spread but does not identify the spread's statistical meaning. Aggregation: Not reported Therapeutics Data Commons: Machine Learning Datasets and Tasks for Drug Discovery and Development · Table 4, block(Excretion), row(CNN), column(TDC.CL-Micro)
Method: CNN	Task: TDC ADMET benchmark group TDC-CYP2C9-INHIBITION: Metabolism: TDC.CYP2C9 Inhibition Dataset subset: TDC.CYP2C9 Inhibition (TDC ADMET benchmark group split)	0.713 ± 0.006 auprc fraction · higher Uncertainty: type: reported_plus_minus_type_unresolved; value: 0.006 Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source CNN on TDC ADMET benchmark group TDC-CYP2C9-INHIBITION: Metabolism: TDC.CYP2C9 Inhibition Scaffold split, as the ADMET benchmark group defines it. Table 4 reports a value and a plus-or-minus spread but does not identify the spread's statistical meaning. Aggregation: Not reported Therapeutics Data Commons: Machine Learning Datasets and Tasks for Drug Discovery and Development · Table 4, block(Metabolism), row(CNN), column(TDC.CYP2C9 Inhibition)
Method: CNN	Task: TDC ADMET benchmark group TDC-CYP2C9-SUBSTRATE: Metabolism: TDC.CYP2C9 Substrate Dataset subset: TDC.CYP2C9 Substrate (TDC ADMET benchmark group split)	0.367 ± 0.059 auprc fraction · higher Uncertainty: type: reported_plus_minus_type_unresolved; value: 0.059 Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source CNN on TDC ADMET benchmark group TDC-CYP2C9-SUBSTRATE: Metabolism: TDC.CYP2C9 Substrate Scaffold split, as the ADMET benchmark group defines it. Table 4 reports a value and a plus-or-minus spread but does not identify the spread's statistical meaning. Aggregation: Not reported Therapeutics Data Commons: Machine Learning Datasets and Tasks for Drug Discovery and Development · Table 4, block(Metabolism), row(CNN), column(TDC.CYP2C9 Substrate)
Method: CNN	Task: TDC ADMET benchmark group TDC-CYP2D6-INHIBITION: Metabolism: TDC.CYP2D6 Inhibition Dataset subset: TDC.CYP2D6 Inhibition (TDC ADMET benchmark group split)	0.544 ± 0.053 auprc fraction · higher Uncertainty: type: reported_plus_minus_type_unresolved; value: 0.053 Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source CNN on TDC ADMET benchmark group TDC-CYP2D6-INHIBITION: Metabolism: TDC.CYP2D6 Inhibition Scaffold split, as the ADMET benchmark group defines it. Table 4 reports a value and a plus-or-minus spread but does not identify the spread's statistical meaning. Aggregation: Not reported Therapeutics Data Commons: Machine Learning Datasets and Tasks for Drug Discovery and Development · Table 4, block(Metabolism), row(CNN), column(TDC.CYP2D6 Inhibition)
Method: CNN	Task: TDC ADMET benchmark group TDC-CYP2D6-SUBSTRATE: Metabolism: TDC.CYP2D6 Substrate Dataset subset: TDC.CYP2D6 Substrate (TDC ADMET benchmark group split)	0.485 ± 0.037 auprc fraction · higher Uncertainty: type: reported_plus_minus_type_unresolved; value: 0.037 Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source CNN on TDC ADMET benchmark group TDC-CYP2D6-SUBSTRATE: Metabolism: TDC.CYP2D6 Substrate Scaffold split, as the ADMET benchmark group defines it. Table 4 reports a value and a plus-or-minus spread but does not identify the spread's statistical meaning. Aggregation: Not reported Therapeutics Data Commons: Machine Learning Datasets and Tasks for Drug Discovery and Development · Table 4, block(Metabolism), row(CNN), column(TDC.CYP2D6 Substrate)
Method: CNN	Task: TDC ADMET benchmark group TDC-CYP3A4-INHIBITION: Metabolism: TDC.CYP3A4 Inhibition Dataset subset: TDC.CYP3A4 Inhibition (TDC ADMET benchmark group split)	0.821 ± 0.003 auprc fraction · higher Uncertainty: type: reported_plus_minus_type_unresolved; value: 0.003 Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source CNN on TDC ADMET benchmark group TDC-CYP3A4-INHIBITION: Metabolism: TDC.CYP3A4 Inhibition Scaffold split, as the ADMET benchmark group defines it. Table 4 reports a value and a plus-or-minus spread but does not identify the spread's statistical meaning. Aggregation: Not reported Therapeutics Data Commons: Machine Learning Datasets and Tasks for Drug Discovery and Development · Table 4, block(Metabolism), row(CNN), column(TDC.CYP3A4 Inhibition)
Method: CNN	Task: TDC ADMET benchmark group TDC-CYP3A4-SUBSTRATE: Metabolism: TDC.CYP3A4 Substrate Dataset subset: TDC.CYP3A4 Substrate (TDC ADMET benchmark group split)	0.662 ± 0.031 auroc fraction · higher Uncertainty: type: reported_plus_minus_type_unresolved; value: 0.031 Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source CNN on TDC ADMET benchmark group TDC-CYP3A4-SUBSTRATE: Metabolism: TDC.CYP3A4 Substrate Scaffold split, as the ADMET benchmark group defines it. Table 4 reports a value and a plus-or-minus spread but does not identify the spread's statistical meaning. Aggregation: Not reported Therapeutics Data Commons: Machine Learning Datasets and Tasks for Drug Discovery and Development · Table 4, block(Metabolism), row(CNN), column(TDC.CYP3A4 Substrate)
Method: CNN	Task: TDC ADMET benchmark group TDC-DILI: Toxicity: TDC.DILI Dataset subset: TDC.DILI (TDC ADMET benchmark group split)	0.792 ± 0.016 auroc fraction · higher Uncertainty: type: reported_plus_minus_type_unresolved; value: 0.016 Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source CNN on TDC ADMET benchmark group TDC-DILI: Toxicity: TDC.DILI Scaffold split, as the ADMET benchmark group defines it. Table 4 reports a value and a plus-or-minus spread but does not identify the spread's statistical meaning. Aggregation: Not reported Therapeutics Data Commons: Machine Learning Datasets and Tasks for Drug Discovery and Development · Table 4, block(Toxicity), row(CNN), column(TDC.DILI)
Method: CNN	Task: TDC ADMET benchmark group TDC-HALF-LIFE: Excretion: TDC.Half Life Dataset subset: TDC.Half Life (TDC ADMET benchmark group split)	0.038 ± 0.138 spearman correlation · higher Uncertainty: type: reported_plus_minus_type_unresolved; value: 0.138 Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source CNN on TDC ADMET benchmark group TDC-HALF-LIFE: Excretion: TDC.Half Life Scaffold split, as the ADMET benchmark group defines it. Table 4 reports a value and a plus-or-minus spread but does not identify the spread's statistical meaning. Aggregation: Not reported Therapeutics Data Commons: Machine Learning Datasets and Tasks for Drug Discovery and Development · Table 4, block(Excretion), row(CNN), column(TDC.Half_Life)
Method: CNN	Task: TDC ADMET benchmark group TDC-HERG: Toxicity: TDC.hERG Dataset subset: TDC.hERG (TDC ADMET benchmark group split)	0.754 ± 0.037 auroc fraction · higher Uncertainty: type: reported_plus_minus_type_unresolved; value: 0.037 Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source CNN on TDC ADMET benchmark group TDC-HERG: Toxicity: TDC.hERG Scaffold split, as the ADMET benchmark group defines it. Table 4 reports a value and a plus-or-minus spread but does not identify the spread's statistical meaning. Aggregation: Not reported Therapeutics Data Commons: Machine Learning Datasets and Tasks for Drug Discovery and Development · Table 4, block(Toxicity), row(CNN), column(TDC.hERG)
Method: CNN	Task: TDC ADMET benchmark group TDC-HIA: Absorption: TDC.HIA Dataset subset: TDC.HIA (TDC ADMET benchmark group split)	0.869 ± 0.026 auroc fraction · higher Uncertainty: type: reported_plus_minus_type_unresolved; value: 0.026 Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source CNN on TDC ADMET benchmark group TDC-HIA: Absorption: TDC.HIA Scaffold split, as the ADMET benchmark group defines it. Table 4 reports a value and a plus-or-minus spread but does not identify the spread's statistical meaning. Aggregation: Not reported Therapeutics Data Commons: Machine Learning Datasets and Tasks for Drug Discovery and Development · Table 4, block(Absorption), row(CNN), column(TDC.HIA)
Method: CNN	Task: TDC ADMET benchmark group TDC-LD50: Toxicity: TDC.LD50 Dataset subset: TDC.LD50 (TDC ADMET benchmark group split)	0.675 ± 0.011 mae error · lower Uncertainty: type: reported_plus_minus_type_unresolved; value: 0.011 Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source CNN on TDC ADMET benchmark group TDC-LD50: Toxicity: TDC.LD50 Scaffold split, as the ADMET benchmark group defines it. Table 4 reports a value and a plus-or-minus spread but does not identify the spread's statistical meaning. Aggregation: Not reported Therapeutics Data Commons: Machine Learning Datasets and Tasks for Drug Discovery and Development · Table 4, block(Toxicity), row(CNN), column(TDC.LD50)
Method: CNN	Task: TDC ADMET benchmark group TDC-LIPO: Absorption: TDC.Lipo Dataset subset: TDC.Lipo (TDC ADMET benchmark group split)	0.743 ± 0.020 mae error · lower Uncertainty: type: reported_plus_minus_type_unresolved; value: 0.020 Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source CNN on TDC ADMET benchmark group TDC-LIPO: Absorption: TDC.Lipo Scaffold split, as the ADMET benchmark group defines it. Table 4 reports a value and a plus-or-minus spread but does not identify the spread's statistical meaning. Aggregation: Not reported Therapeutics Data Commons: Machine Learning Datasets and Tasks for Drug Discovery and Development · Table 4, block(Absorption), row(CNN), column(TDC.Lipo)
Method: CNN	Task: TDC ADMET benchmark group TDC-PGP: Absorption: TDC.Pgp Dataset subset: TDC.Pgp (TDC ADMET benchmark group split)	0.908 ± 0.012 auroc fraction · higher Uncertainty: type: reported_plus_minus_type_unresolved; value: 0.012 Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source CNN on TDC ADMET benchmark group TDC-PGP: Absorption: TDC.Pgp Scaffold split, as the ADMET benchmark group defines it. Table 4 reports a value and a plus-or-minus spread but does not identify the spread's statistical meaning. Aggregation: Not reported Therapeutics Data Commons: Machine Learning Datasets and Tasks for Drug Discovery and Development · Table 4, block(Absorption), row(CNN), column(TDC.Pgp)
Method: CNN	Task: TDC ADMET benchmark group TDC-PPBR: Distribution: TDC.PPBR Dataset subset: TDC.PPBR (TDC ADMET benchmark group split)	11.106 ± 0.358 mae error · lower Uncertainty: type: reported_plus_minus_type_unresolved; value: 0.358 Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source CNN on TDC ADMET benchmark group TDC-PPBR: Distribution: TDC.PPBR Scaffold split, as the ADMET benchmark group defines it. Table 4 reports a value and a plus-or-minus spread but does not identify the spread's statistical meaning. Aggregation: Not reported Therapeutics Data Commons: Machine Learning Datasets and Tasks for Drug Discovery and Development · Table 4, block(Distribution), row(CNN), column(TDC.PPBR)
Method: CNN	Task: TDC ADMET benchmark group TDC-VD: Distribution: TDC.VDss Dataset subset: TDC.VDss (TDC ADMET benchmark group split)	0.226 ± 0.114 spearman correlation · higher Uncertainty: type: reported_plus_minus_type_unresolved; value: 0.114 Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source CNN on TDC ADMET benchmark group TDC-VD: Distribution: TDC.VDss Scaffold split, as the ADMET benchmark group defines it. Table 4 reports a value and a plus-or-minus spread but does not identify the spread's statistical meaning. Aggregation: Not reported Therapeutics Data Commons: Machine Learning Datasets and Tasks for Drug Discovery and Development · Table 4, block(Distribution), row(CNN), column(TDC.VD)

Source checking is not independent reproduction. Release 2026-09-29-06401fd5b220.

Use this model

How it works, versions and access

Strengths, limitations and unresolved questions

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.

0 evidence rows matching the loaded filters