BottleFit: Learning Compressed Representations in Deep Neural Networks for Effective and Efficient Split Computing.
Yoshitomo MatsubaraDavide CallegaroSameer SinghMarco LevoratoFrancesco RestucciaPublished in: CoRR (2022)
Keyphrases
- neural network
- learning process
- reinforcement learning
- cost effective
- learning algorithm
- prior knowledge
- learning systems
- online learning
- cascade correlation
- recurrent networks
- multiple representations
- learning capabilities
- incremental learning
- learning problems
- learning tasks
- computationally efficient
- supervised learning
- mobile devices
- artificial neural networks
- pattern recognition
- machine learning