DOE PAGES title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Unsupervised physics-informed disentanglement of multimodal data

Journal Article · · Foundations of Data Science
 [1];  [2];  [3];  [4];  [1];  [5];  [1];  [1];  [1];  [6]
  1. Sandia National Laboratories (SNL-NM), Albuquerque, NM (United States)
  2. Univ. of Pennsylvania, Philadelphia, PA (United States)
  3. Sandia National Laboratories (SNL-NM), Albuquerque, NM (United States); Arizona State Univ., Tempe, AZ (United States)
  4. Arizona State Univ., Tempe, AZ (United States)
  5. Northwestern Univ., Evanston, IL (United States)
  6. Center for Integrated Nanotechnologies

Here, we introduce physics-informed multimodal autoencoders (PIMA) - a variational inference framework for discovering shared information in multimodal datasets. Individual modalities are embedded into a shared latent space and fused through a product-of-experts formulation, enabling a Gaussian mixture prior to identify shared features. Sampling from clusters allows cross-modal generative modeling, with a mixture-of-experts decoder that imposes inductive biases from prior scientific knowledge and thereby imparts structured disentanglement of the latent space. This approach enables cross-modal inference and the discovery of features in high-dimensional heterogeneous datasets. Consequently, this approach provides a means to discover fingerprints in multimodal scientific datasets and to avoid traditional bottlenecks related to high-fidelity measurement and characterization of scientific datasets.

Research Organization:
Sandia National Laboratories (SNL-NM), Albuquerque, NM (United States)
Sponsoring Organization:
USDOE Office of Science (SC), Advanced Scientific Computing Research (ASCR); USDOE Laboratory Directed Research and Development (LDRD) Program; National Science Foundation (NSF)
Grant/Contract Number:
NA0003525
OSTI ID:
2441141
Report Number(s):
SAND--2024-12122J
Journal Information:
Foundations of Data Science, Journal Name: Foundations of Data Science Journal Issue: 1 Vol. 7; ISSN 2639-8001
Publisher:
AIMSCopyright Statement
Country of Publication:
United States
Language:
English

References (15)

The Hungarian method for the assignment problem journal March 1955
Applications of fluorescence spectroscopy in protein conformational changes and intermolecular contacts journal January 2023
Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations journal February 2019
DeepM&Mnet for hypersonics: Predicting the coupled flow and finite-rate chemistry behind a normal shock using neural-network approximation of operators journal December 2021
Enforcing exact physics in scientific machine learning: A data-driven exterior calculus on graphs journal May 2022
Long-time integration of parametric evolution equations with physics-informed DeepONets journal February 2023
Fingerprints of the quantum space-time in time-dependent quantum mechanics: An emergent geometric phase journal February 2022
The Evolution of High-Throughput Experimentation in Pharmaceutical Development and Perspectives on the Future journal May 2019
Autonomy in materials research: a case study in carbon nanotube growth journal October 2016
Author Correction: SciPy 1.0: fundamental algorithms for scientific computing in Python journal February 2020
Micromachined tunable metamaterials: a review journal September 2012
A Commentary on the Unsupervised Learning of Disentangled Representations journal April 2020
seaborn: statistical data visualization journal April 2021
Variational Deep Embedding: An Unsupervised and Generative Approach to Clustering conference August 2017
Solid State NMR Spectroscopy a Valuable Technique for Structural Insights of Advanced Thin Film Materials: A Review journal June 2021