Why Retraining Can Be Harder Than Training
- In a rapidly changing world, humans are required to quickly adapt to a new environment. Neural networks show why this is easier said than done. Our article uses a perceptron to demonstrate why unlearning and relearning can be costlier than learning from scratch.
Author: | Christian KochORCiD, Markus Stadi |
---|---|
URL / DOI: | https://towardsdatascience.com/why-retraining-can-be-harder-than-training-489b3bc6ae02 |
Parent Title (English): | Towards Data Science |
Subtitle (English): | A neural network perspective on learning, unlearning and relearning |
Publisher: | Medium |
Document Type: | Report |
Language: | English |
Date of first Publication: | 2024/01/23 |
Release Date: | 2024/02/05 |
GND Keyword: | Künstliche Intelligenz; maschinelles Lernen; neuronales Netz |
Pagenumber: | 10 |
First Page: | ohne Zählung |
institutes: | Fakultät Betriebswirtschaft |
Dewey Decimal Classification: | 0 Informatik, Informationswissenschaft, allgemeine Werke |
Licence (German): | Creative Commons - CC BY - Namensnennung 4.0 International |