Deep Learning Essentials

Co-learning

Co-learning refers to transferring knowledge between modalities and their representations. Some challenges are as follows:

• Partially-observable views
• Pivot identification
• Collaborative overfitting
• Imperfect predictions
• Potential divergence

In the following section, we will cover a few application cases for some recent benchmarks of multimodality learning. We will be focusing on a high-level introduction of these different fields to give you a better overview and a starting point for how each application can be implemented. With them as an anchor point, you should be able to start your own exploration of deep learning on multimodalities.