Edoardo-Tronci/ITA

Incremental Training Algorithm

Incremental Training Algorithm (ITA)

ITA python code (see [1]). The implementation uses python+pytorch

Machine Learning applications

ITA has been tested on the following Machine Learning applications:

• Adult
• Ailerons
• Appliances Energy Prediction
• Arcene
• BlogFeedback
• Boston House Prices
• Breast Cancer Wisconsin (Diagnostic)
• CIFAR 10
• Gisette
• Iris
• MNIST Handwritten Digit
• Mv
• QSAR Oral Toxicity

Authors

A. De Santis¹, G. Liuzzi¹, S. Lucidi¹, E.M. Tronci¹

¹ Department of Computer, Control and Management Engineering, "Sapienza" University of Rome

• desantis@diag.uniroma1.it,
• liuzzi@diag.uniroma1.it,
• lucidi@diag.uniroma1.it,
• tronci@diag.uniroma1.it

Copyright 2022

References

[1] De Santis, A., Liuzzi, G., Lucidi, S., & Tronci, E. M. (2021). Avoiding local minima in multilayer network optimization by incremental training. arXiv preprint arXiv:2106.06477.

[2] Fukumizu, K., & Amari, S. I. (2000). Local minima and plateaus in hierarchical structures of multilayer perceptrons. Neural networks, 13(3), 317-327.