Pushing boundaries of machine learning
Finally, various alternative approaches emerged in recent years when it comes to model architectures, training, or even what a neural network should be. As I’ve written above, we’re still early in the process of understanding machine learning and its applications. That’s why each year brings breakthroughs, and it will stay so for the foreseeable future.
In this section, I’ve gathered concepts that push boundaries and are actively tested. Some of them might need years to fully blossom, while others would never make it out of research labs. Here are the top 4 trends in machine learning to watch in the upcoming years:
- Graph Neural Networks
- Bayesian Deep Learning
- Active Learning
- Federated Learning
Graph Neural Networks84 are deep learning methods that operate on graphs. A graph is a data structure that models a set of objects (nodes) and their relationships (edges). Standard neural networks like CNNs...