Improved decision making with similarity based machine learning: Applications in chemistry

Abstract

Despite the fundamental progress in autonomous molecular and materials discovery, data scarcity throughout chemical compound space still severely hampers the use of modern ready-made machine learning models as they rely heavily on the paradigm, 'the bigger the data the better'. Presenting similarity based machine learning (SML), we show an approach to select data and train a model on-the-fly for specific queries, enabling decision making in data scarce scenarios in chemistry. By solely relying on query and training data proximity to choose training points, only a fraction of data is necessary to converge to competitive performance. After introducing SML for the harmonic oscillator and the Rosenbrock function, we describe applications to scarce data scenarios in chemistry which include quantum mechanics based molecular design and organic synthesis planning. Finally, we derive a relationship between the intrinsic dimensionality and volume of feature space, governing the overall model accuracy.

References in corpus (28)

• On representing chemical environments
• Fast and Accurate Modeling of Molecular Atomization Energies with Machine Learning
• Machine Learning of Accurate Energy-Conserving Molecular Force Fields
• Less is more: sampling chemical space with active learning
• Machine Learning Unifies the Modelling of Materials and Molecules
• Active Learning of Uniformly Accurate Inter-atomic Potentials for Materials Simulation
• Machine learning for electronically excited states of molecules
• Estimating the intrinsic dimension of datasets by a minimal neighborhood information
• FCHL revisited: faster and more accurate quantum machine learning
• Towards self-driving laboratories: The central role of density functional theory in the AI age
• Combining SchNet and SHARC: The SchNarc machine learning approach for excited-state dynamics
• Intrinsic dimension of data representations in deep neural networks
• Equivariant Diffusion for Molecule Generation in 3D
• Machine learning based energy-free structure predictions of molecules (closed and open-shell), transition states, and solids
• Bayesian optimization with known experimental and design constraints for chemistry applications
• EquiformerV2: Improved Equivariant Transformer for Scaling to Higher-Degree Representations
• Torsional Diffusion for Molecular Conformer Generation
• High-throughput property-driven generative design of functional organic molecules
• Emergent autonomous scientific research capabilities of large language models
• Machine Learning of Free Energies in Chemical Compound Space Using Ensemble Representations: Reaching Experimental Uncertainty for Solvation
• Machine learning the computational cost of quantum chemistry
• Exploring the robust extrapolation of high-dimensional machine learning potentials
• Geometric Latent Diffusion Models for 3D Molecule Generation
• Impact of the characteristics of quantum chemical databases on machine learning predictions of tautomerization energies
• Reducing Training Data Needs with Minimal Multilevel Machine Learning (M3L)
• The Shape of Learning Curves: a Review
• Intrinsic dimension estimation for discrete metrics
• The Statistical Complexity of Interactive Decision Making

Cited by in corpus (2)

• Engineering Artificial Intelligence: Framework, Challenges, and Future Direction
• Integer linear programming for unsupervised training set selection in molecular machine learning