Learning: Machine vs Deep
- Machine Learning is a subfield of Artificial Intelligence where we try to make ‘accurate’ predictions based on the training data presented to a model or classifier to solve a problem statement.
- We try to fit the training data using various algorithms preceded by feature extraction to formulate a model which generalises well on the test data (unseen) without being explicitly programmed.
- The algorithms to be used are task-specific and depends on the type of dataset in hand.
- Preferable to use for less complex problems where training dataset size is moderate and infrastructure is limited.
- Deep Learning is a subfield of Machine Learning that solves the same problem as Machine Learning but it approaches the problem by using ‘deep’ Artificial Neural Networks (ANN).
- ANN’s are inspired from the way a human brain works using a network of neurons to solve a problem.
- The neural network learns the features itself from the data presented using hidden layer architecture and the accuracy improves as more and more data is fed to the network.
- Preferable to use for complex problems like Image classification, Natural Language Processing where we have huge sets of data and high end infrastructure to support large computations.
Author – Abhishek Saxena