AI Paper Atlas

Research map / Probabilistic and scientific modeling / Molecules and proteins

Crystallization Learning with the Delaunay Triangulation

Jiaqi Gu, Guosheng Yin

ICML 2021 · Probabilistic and scientific modeling › Molecules and proteins · approach: Sparse graphical and tensor models · 0 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
Crystallization learning (Delaunay-based local linear fit) instead of k-NN, local linear, kernel regression and GPnonparametric regression of conditional expectation · Synthetic Scenario 1 and 2, d=5 and d=10log(MSE) ratio vs baseline: baseline log(MSE_method/MSE_CL) mostly positive, e.g. 0.23 to 0.57 at d=5, n=200Improved
Crystallization learning instead of existing methods in high dimensions (d=20, 50)nonparametric regression of conditional expectation · Synthetic Scenario 1 and 2, d=20 and d=50log(MSE)No clear effect
Increase sample size for crystallization learningcrystallization learning · Synthetic Scenario 1 and 2log(MSE): e.g. -1.11 at n=200 to -2.21 at n=2000 (d=5)Improved
Crystallization learning instead of k-NN, local linear, kernel regression and GP on UCI real dataprediction of response on CASP, Concrete, Parkinson's data · UCI real data bootstrap testsmean predictive squared error (MPSE)Improved
Data-driven LOO-CV selection of L versus fixed L (ablation)crystallization learning · Synthetic Scenario 1, d=5log(MSE)No 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