Last Updated on September 11, 2022
Zero–shot learning and one–shot learning are both methods of machine learning where an algorithm can learn from a small amount of data. In zero–shot learning, the algorithm can learn from data that does not contain labels, while in one–shot learning, the algorithm can learn from data that only contains one label per data point.
Zero–shot learning is a relatively new method of machine learning that has shown promise in some applications. One example is image recognition, where the algorithm can learn to recognize objects from images that do not contain labels. Another example is text classification, where the algorithm can learn to classify documents without any labels.
One–shot learning has been around for longer and is more commonly used. One example is facial recognition, where the algorithm can learn to recognize faces from a single image. Another example is hand–written digit recognition, where the algorithm can learn to recognize digits from a single image.
Both zero–shot learning and one–shot learning are powerful methods of machine learning that can be used to learn from small amounts of data.
