Sparse Training Theory for Scalable and Efficient Agents

Decebal Constantin Mocanu, Elena Mocanu, Tiago Pinto, Selima Curci, Phuong H. Nguyen, Madeleine Gibescu, Damien Ernst, Zita Vale

Research output: Chapter in Book/Report/Conference proceedingChapterAcademicpeer-review

3 Downloads (Pure)

Abstract

A fundamental task for artificial intelligence is learning. Deep Neural Networks have proven to cope perfectly with all learning paradigms, i.e. supervised, unsupervised, and reinforcement learning. Nevertheless, traditional deep learning approaches make use of cloud computing facilities and do not scale well to autonomous agents with low computational resources. Even in the cloud, they suffer from computational and memory limitations, and they cannot be used to model adequately large physical worlds for agents which assume networks with billions of neurons. These issues are addressed in the last few years by the emerging topic of sparse training, which trains sparse networks from scratch. This paper discusses sparse training state-of-the-art, its challenges and limitations while introducing a couple of new theoretical research directions which has the potential of alleviating sparse training limitations to push deep learning scalability well beyond its current boundaries. Nevertheless, the theoretical advancements impact in complex multi-agents settings is discussed from a real-world perspective, using the smart grid case study.
Original languageEnglish
Title of host publication20th International Conference on Autonomous Agents and Multiagent Systems, AAMAS 2021
Publication statusAccepted/In press - 25 Jan 2021
Event20th International Conference on Autonomous Agents and Multiagent Systems, AAMAS 2021 - Virtual Event, United Kingdom
Duration: 3 May 20217 May 2021
Conference number: 20
https://aamas2021.soton.ac.uk/

Conference

Conference20th International Conference on Autonomous Agents and Multiagent Systems, AAMAS 2021
Abbreviated titleAAMAS
CountryUnited Kingdom
CityVirtual Event
Period3/05/217/05/21
Internet address

Fingerprint Dive into the research topics of 'Sparse Training Theory for Scalable and Efficient Agents'. Together they form a unique fingerprint.

Cite this