Skip to main content
U.S. Department of Energy
Office of Scientific and Technical Information

Off-the-shelf deep learning is not enough, and requires parsimony, Bayesianity, and causality

Journal Article · · npj Computational Materials
Abstract

Deep neural networks (‘deep learning’) have emerged as a technology of choice to tackle problems in speech recognition, computer vision, finance, etc. However, adoption of deep learning in physical domains brings substantial challenges stemming from the correlative nature of deep learning methods compared to the causal, hypothesis driven nature of modern science. We argue that the broad adoption of Bayesian methods incorporating prior knowledge, development of solutions with incorporated physical constraints and parsimonious structural descriptors and generative models, and ultimately adoption of causal models, offers a path forward for fundamental and applied research.

Research Organization:
Oak Ridge National Laboratory (ORNL), Oak Ridge, TN (United States)
Sponsoring Organization:
USDOE; USDOE Office of Science (SC), Basic Energy Sciences (BES)
Grant/Contract Number:
AC05-00OR22725
OSTI ID:
1763073
Alternate ID(s):
OSTI ID: 1819625
OSTI ID: 1787453
Journal Information:
npj Computational Materials, Journal Name: npj Computational Materials Journal Issue: 1 Vol. 7; ISSN 2057-3960
Publisher:
Nature Publishing GroupCopyright Statement
Country of Publication:
United Kingdom
Language:
English

References (39)

U-Net: Convolutional Networks for Biomedical Image Segmentation
  • Ronneberger, Olaf; Fischer, Philipp; Brox, Thomas
  • Medical Image Computing and Computer-Assisted Intervention – MICCAI 2015: 18th International Conference, Munich, Germany, October 5-9, 2015, Proceedings, Part III https://doi.org/10.1007/978-3-319-24574-4_28
book November 2015
Collective properties of neural networks: A statistical physics approach journal February 1984
Materials informatics journal January 2018
Potential energy surfaces for macromolecules. A neural network technique journal May 1992
Materials informatics: From the atomic-level to the continuum journal April 2019
Scoping the polymer genome: A roadmap for rational polymer dielectrics design and beyond journal September 2018
Causality book January 2009
Phases and Interfaces from Real Space Atomically Resolved Data: Physics-Based Deep Data Image Analysis journal August 2016
Automatic Chemical Design Using a Data-Driven Continuous Representation of Molecules journal January 2018
Designing Algorithms To Aid Discovery by Chemical Robots journal July 2018
Knowledge Extraction from Atomically Resolved Images journal October 2017
Deep Learning of Atomically Resolved Scanning Transmission Electron Microscopy Images: Chemical Identification and Tracking Local Transformations journal October 2016
Deep learning journal May 2015
Design of efficient molecular organic light-emitting diodes by a high-throughput virtual screening and experimental approach journal August 2016
Autonomy in materials research: a case study in carbon nanotube growth journal October 2016
Mutual information, neural networks and the renormalization group journal March 2018
How to explore chemical space using algorithms and automation journal January 2019
Accelerating the discovery of materials for clean energy in the era of smart automation journal April 2018
Occurrence of the potent mutagens 2- nitrobenzanthrone and 3-nitrobenzanthrone in fine airborne particles journal January 2019
A Kriging-Based Approach to Autonomous Experimentation with Applications to X-Ray Scattering journal August 2019
Re-epithelialization and immune cell behaviour in an ex vivo human skin model journal January 2020
Automated microfluidic platform for systematic studies of colloidal perovskite nanocrystals: towards continuous nano-manufacturing journal January 2017
Facile synthesis of anhydrous microparticles using plug-and-play microfluidic reactors journal January 2020
Building ferroelectric from the bottom up: The machine learning analysis of the atomic-scale ferroelectric distortions journal July 2019
Spatially resolved mapping of disorder type and distribution in random systems using artificial neural network recognition journal July 2011
Disorder Identification in Hysteresis Data: Recognition Analysis of the Random-Bond–Random-Field Ising Model journal October 2009
Crystal Graph Convolutional Neural Networks for an Accurate and Interpretable Prediction of Material Properties journal April 2018
Discovering Physical Concepts with Neural Networks journal January 2020
Using neural networks to construct models of the molecular beam epitaxy process journal January 2000
Strike (With) a Pose: Neural Networks Are Easily Fooled by Strange Poses of Familiar Objects conference June 2019
Conditional Random Fields Meet Deep Neural Networks for Semantic Segmentation: Combining Probabilistic Graphical Models with Deep Learning for Structured Prediction journal January 2018
One Pixel Attack for Fooling Deep Neural Networks journal October 2019
Theory-Guided Data Science: A New Paradigm for Scientific Discovery from Data journal October 2017
Deep reinforcement learning for de novo drug design journal July 2018
A general reinforcement learning algorithm that masters chess, shogi, and Go through self-play journal December 2018
Inverse molecular design using machine learning: Generative models for matter engineering journal July 2018
Organic synthesis in a modular robotic system driven by a chemical programming language journal November 2018
Principles of neurodynamics. Perceptrons and the theory of brain mechanisms report March 1961
An Introduction to Causal Inference posted_content February 2020

Similar Records

Related Subjects