IBM defines machine learning as a branch of artificial intelligence and computer science that focuses on the use of data and algorithms to imitate the way that humans learn, gradually improving its accuracy.
However, we routinely see examples of machines only learning the most superficial aspects of the subject and imitating those.
Social media, for all its shortcomings, is adept at identifying images that instantly expose the underlying truth of a situation, and this tweet shows us how far machine learning still has to go.
This picture of jigsaw puzzle pieces pushed haphazardly into piles of the same color to create a recognizable imitation of the intended result will surely come to mind the next time a Tesla plows into a stationary police car on the shoulder of the road because its machine learning still can't assemble the puzzle pieces of a car parked on the shoulder of the highway where there aren't normally stopped cars.
UC Berkeley identifies three parts to machine learning algorithms:
- A Decision Process: In general, machine learning algorithms are used to make a prediction or classification. Based on some input data, which can be labeled or unlabeled, your algorithm will produce an estimate of a pattern in the data.
- An Error Function: An error function serves to evaluate the prediction of the model. If there are known examples, an error function can make a comparison to assess the accuracy of the model.
- A Model Optimization Process: If the model can fit better to the data points in the training set, then weights are adjusted to reduce the discrepancy between the known example and the model estimate. The algorithm will repeat this evaluate and optimize the process, updating weights autonomously until a threshold of accuracy has been met.
At Nvidia, they see four different methods of machine learning, which the company identifies as:
- Supervised learning: The dataset being used has been pre-labeled and classified by users to allow the algorithm to see how accurate its performance is.
- Unsupervised learning: The raw dataset being used is unlabeled and an algorithm identifies patterns and relationships within the data without help from users.
- Semisupervised learning: The dataset contains structured and unstructured data, which guide the algorithm on its way to making independent conclusions. The combination of the two data types in one training dataset allows machine learning algorithms to learn to label unlabeled data.
- Reinforcement learning: The dataset uses a “rewards/punishments” system, offering feedback to the algorithm to learn from its own experiences by trial and error.