Dataset Reduction Methods for Data and Energy Efficient Deep Learning

Krol, Daan (2023) Dataset Reduction Methods for Data and Energy Efficient Deep Learning. Master's Thesis / Essay, Artificial Intelligence.

Preview	Text Master_Thesis_Daan_Krol_Final.pdf Download (16MB) | Preview
	Text toestemming.pdf Restricted to Registered users only Download (100kB)

Abstract

Artificial intelligence models that achieve high performance in tasks such as natural language processing and computer vision often require large and complex datasets. However, the use of these datasets can be costly in terms of computational expense, training time, and energy consumption. The carbon footprint of these models can also be significant. Data-efficient machine learning, which aims to reduce the amount of data required to achieve good performance, is therefore an important area of research. Dataset reduction methods aim to select or synthesize a smaller subset of the data that still allows good performance to be achieved. These methods can be applied non-adaptively by pruning the dataset in advance or adaptively by selecting a subset of the data every X epochs while training.

Item Type:	Thesis (Master's Thesis / Essay)
Supervisor name:	Mohades Kasaei, S.H.
Degree programme:	Artificial Intelligence
Thesis type:	Master's Thesis / Essay
Language:	English
Date Deposited:	16 Jan 2023 12:22
Last Modified:	01 Jun 2023 11:20
URI:	https://fse.studenttheses.ub.rug.nl/id/eprint/29137

Actions (login required)

View Item