analiza aiqbench
MolLedger introduces a graph neural network architecture that achieves exact interpretability by decomposing ADME predictions into per-atom scores without sacrificing performance. By anchoring these attributions to chemical properties through an auxiliary loss, the model provides more faithful explanations for molecular property changes compared to standard interpretability methods. This advancement addresses a critical bottleneck in drug discovery where understanding model reasoning is as important as the prediction itself.
MolLedger: An Additive Graph Neural Network with Chemically Grounded ADME Attributions
arXiv:2608.30636v1 Announce Type: new Abstract: Optimizing absorption, distribution, metabolism, and excretion (ADME) is an important part of small molecule drug discovery. Many machine learning models have been built to predict ADME properties to facilitate this optimization process, but explaining model predictions is challenging. We propose a new graph neural network architecture with built-in meaningful per-atom attributions. Our model MolLedger outputs predictions that are the sum of per-atom scores. MolLedger's additive framework obtains exact interpretability at no cost to performance because the global context vector gives the additive head enough context to produce good per-atom scores. Furthermore, MolLedger produces attributions that are more faithful to chemical properties than other interpretability methods because the auxiliary loss in MolLedger anchors the atom scores to chemical properties. Our case studies comparing interpretations from multiple methods on molecular pairs reveal that MolLedger is much better at producing sensible explanations for predicted property changes.
Abstract:Optimizing absorption, distribution, metabolism, and excretion (ADME) is an important part of small molecule drug discovery. Many machine learning models have been built to predict ADME properties to facilitate this optimization process, but explaining model predictions is challenging. We propose a new graph neural network architecture with built-in meaningful per-atom attributions. Our model MolLedger outputs predictions that are the sum of per-atom scores. MolLedger's additive framework obtains exact interpretability at no cost to performance because the global context vector gives the additive head enough context to produce good per-atom scores. Furthermore, MolLedger produces attributions that are more faithful to chemical properties than other interpretability methods because the auxiliary loss in MolLedger anchors the atom scores to chemical properties. Our case studies comparing interpretations from multiple methods on molecular pairs reveal that MolLedger is much better at producing sensible explanations for predicted property changes.
Machine Learning (cs.LG)
arXiv:2608.30636 [cs.LG]
(or arXiv:2608.30636v1 [cs.LG] for this version)
https://doi.org/10.48550/arXiv.2608.30636 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 237 cuvinte
Subiecte conexe
News
Quantum Computing Report · 25 iul. 2026
Electronic Density versus Geometry for Machine-Learned Molecular Absorption Spectra
arXiv · 5 oct. 2026
PIRAMID: Progress and Plans
LessWrong · 27 iul. 2026
POI Recommendation with LLM-Augmented Multi-Graph Learning and Contrastive Alignment
arXiv · 19 aug. 2026