University of Twente Student Theses
Reducing the Electricity Use of Large Language Models
Todirascu, Emil (2024) Reducing the Electricity Use of Large Language Models.
|
PDF
395kB |
| Abstract: | The use of artificial intelligence (AI) has dramatically increased over the past few years. With the recent surge of Large Language Models (LLMs) and text-to-image generation models, the general public has begun to see the possibilities of artificial intelligence and use them in their personal and professional lives. A large contributor to this is the advancements in both hardware and software technologies. However, these new technologies require careful consideration regarding their energy consumption. As models become more computationally intensive, their training drastically increases their energy consumption. Balancing the benefits LLMs can bring to our lives and their energy requirements is essential to ensure that this technological progress does not come at the expense of environmental impact. Therefore, researchers should aim to use efficient techniques that lower the electricity use of such artificial intelligence models. This research aims to create a model of the electricity consumption of training LLMs and explore techniques machine learning researchers should use to reduce the electricity use of training LLMs. |
| Item Type: | Essay (Bachelor) |
| Faculty: | EEMCS: Electrical Engineering, Mathematics and Computer Science |
| Subject: | 50 technical science in general, 54 computer science |
| Programme: | Business & IT BSc (56066) |
| Link to this item: | https://purl.utwente.nl/essays/101249 |
| Export this item as: | BibTeX EndNote HTML Citation Reference Manager |
Show download statistics for this publication
Show download statistics for this publicationRepository Staff Only: item control page