AI Paper Atlas

Research map / Probabilistic and scientific modeling / Molecules and proteins

QH9: A Quantum Hamiltonian Prediction Benchmark for QM9 Molecules

Haiyang Yu, Meng Liu, Youzhi Luo, Alex Strasser, Xiaofeng Qian, Xiaoning Qian, Shuiwang Ji

NeurIPS 2023 · Probabilistic and scientific modeling › Molecules and proteins · approach: Machine-learned interatomic potentials · 3 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
train on QH9-stable-ood (OOD) instead of stable-id (ID)QHNet · Intersection test setHamiltonian MAE all: 57.53 -> 72.78Hurt
train on QH9-stable-ood instead of stable-idQHNet · Intersection test seteigenenergy MAE: 442.78 -> 630.49Hurt
train on QH9-stable-ood instead of stable-idQHNet · Intersection test setorbital coefficient cosine similarity: 95.26 -> 94.01Hurt
OOD training on 29-atom moleculesQHNet · Intersection test set, 29 atomsHamiltonian MAE all: 47.00 -> 96.15Hurt
molecule-wise instead of geometry-wise split (100k)QHNet · QH9-dynamic-100kHamiltonian MAE all: 96.85 -> 173.92Hurt
molecule-wise instead of geometry-wise split (300k)QHNet · QH9-dynamic-300kHamiltonian MAE all: 100.19 -> 119.66Hurt
more trajectories: 300k-mol instead of 100k-molQHNet · QH9-dynamic-molHamiltonian MAE all: 173.92 -> 119.66Improved
more trajectories: 300k-mol instead of 100k-molQHNet · QH9-dynamic-moleigenenergy MAE: 9719.58 -> 2178.15Improved
more trajectories: 300k-geo instead of 100k-geoQHNet · QH9-dynamic-geoHamiltonian MAE all: 96.85 -> 100.19No clear effect
reduce training set from 104,664 to 10,000 (ablation)QHNet · QH9-stable-ideigenenergy MAE: 798.51 -> 2471.46Hurt
reduce training set from 104,664 to 10,000 (ablation)QHNet · QH9-stable-idHamiltonian MAE all: 76.31 -> 78.60No clear effect
reduce training set from 104,664 to 10,000 (ablation)QHNet · QH9-stable-idorbital coefficient cosine similarity: 95.85 -> 96.18No clear effect

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy