The most popular and comprehensive Open Source ECM platform
Let us know more
Artificial Intelligence: Learning from Limited Data Sets
Machine Learning is celebrated for its ability to be able to do things like reading x-rays or creating images based on a text description. But Machine Learning is also notorious for needing days and days to train on massive amounts of data. But that needn’t be the case.
For cases where there are limited amounts of data, a technique called N-shot learning is often used. N refers to the number of data samples used. The technique often is classified as Zero-shot, one-shot, few-shot, or N-shot (where the data set is large).
Facial recognition is an example of a one-shot problem. Instead of a dataset of many images of a person, the training data consists of only a single image of a person and the problem is to determine whether or not a person matches the known image. When there is only a single image, rather than a classification problem, the problem is a difference-evaluation problem. In this example, the accuracy can be poor if one of the images shows the person with an accessory like a hat or glasses and that does not appear in the other image.
Bethann Noble, Head of Marketing at Continual, told TechTarget that “the ability to learn with limited labeled data opens new product possibilities and allows enterprises to use large pools of otherwise unusable data to be innovative. Enterprises typically sit on large pools of data, but most of this data does not have labels and cannot be used to build a model.”
Jelani Harper, an editorial consultant at KMWorld, said that “one-stop learning and other such approaches make machine learning much more accessible for enterprise purposes, while perhaps expanding this technology’s scope for solving business problems.”
