Machine learning or semantic technology is the basis for most commercial artificial-intelligence systems, and it is about to do things a little differently. This interesting information came from MIT News in their article, “More-flexible machine learning.”
Currently using an object-recognition algorithm to classify a particular image, MIT researchers are presenting a new way that enables semantically related concepts to reinforce each other. The object-recognition algorithm would learn to weigh the co-occurrence of classifications more heavily that before. Because the possible categories feels almost limitless, the conventional way of dealing with it imposes limits.
Traditional algorithms treats all other categories equally unfavorably. Because there are actually semantic similarities between those categories, they have developed a way of making use of that semantic similarity to essentially borrow data from close categories.
Semantic technology and new artificial intelligence language continue to evolve and be used in a variety of applications. It has never been more important to have someone with the expertise and knowledge handling your content, developing your taxonomies, and making your information findable.
Melody K. Smith
Sponsored by Data Harmony, a unit of Access Innovations, the world leader in indexing and making content findable.