Outsmarting neural networks: an alternative paradigm for machine learning
Conference
·
OSTI ID:393293
We address three problems in machine learning, namely: (i) function learning, (ii) regression estimation, and (iii) sensor fusion, in the Probably and Approximately Correct (PAC) framework. We show that, under certain conditions, one can reduce the three problems above to the regression estimation. The latter is usually tackled with artificial neural networks (ANNs) that satisfy the PAC criteria, but have high computational complexity. We propose several computationally efficient PAC alternatives to ANNs to solve the regression estimation. Thereby we also provide efficient PAC solutions to the function learning and sensor fusion problems. The approach is based on cross-fertilizing concepts and methods from statistical estimation, nonlinear algorithms, and the theory of computational complexity, and is designed as part of a new, coherent paradigm for machine learning.
- Research Organization:
- Oak Ridge National Lab., TN (United States)
- Sponsoring Organization:
- USDOE, Washington, DC (United States)
- DOE Contract Number:
- AC05-96OR22464
- OSTI ID:
- 393293
- Report Number(s):
- CONF-9610138--1; ON: DE96014603
- Country of Publication:
- United States
- Language:
- English
Similar Records
QUBO formulations for training machine learning models
PAC learning using Nadaraya-Watson estimator based on orthonormal systems
Journal Article
·
Mon May 10 20:00:00 EDT 2021
· Scientific Reports
·
OSTI ID:1833737
PAC learning using Nadaraya-Watson estimator based on orthonormal systems
Conference
·
Fri Aug 01 00:00:00 EDT 1997
·
OSTI ID:527900