Sparse Training Theory for Scalable and Efficient Agents
This addresses the problem of computational and memory constraints in deep learning for scalable agents, particularly in resource-limited settings, but it is incremental as it builds on existing sparse training methods.
The paper tackles the scalability and efficiency limitations of deep learning for autonomous agents with low computational resources by exploring sparse training, which trains sparse networks from scratch, and introduces new theoretical research directions to push deep learning scalability beyond current boundaries, using a smart grid case study for real-world impact.
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.