Excerpts
- Simple machine learning algorithms
- Сhoice of representation
- How does deep learning solve the central problem in representation learning?
- 2 main ways of measuring the depth of a deep learning model
- Relationships between deep learning, representation learning, machine learning, and artificial intelligence
- Historical trends in deep learning
- The deep learning’s history
- Simple linear models: predecessors of modern deep learning
- Limitations of linear models
- Why has the neuroscience’s role in deep learning been diminished?
- What has neuroscience given to deep learning?
- What is the «computational neuroscience»?
- What is «connectionism»?
- What is «distributed representation»?
- Deep neural networks in the mid-1990s - mid-2000s
- The third wave of neural networks research
- Increasing dataset sizes
- Increasing model sizes
- The number of connections per neuron in animals and artificial neural networks
- The number of neurons in animals and artificial neural networks
- Increasing accuracy, complexity and real-world impact