Saffron 10, a major upgrade to Saffron Technology’s cognitive computing platform, debuts today, with new machine learning algorithms that support anticipating outcomes based on patterns found in large amounts of data. (You can read more about the platform and Saffron’s Associative MemoryBase here as well as in this article at our Dataversity sister site.)
One of the new machine learning algorithms is Saffron Universal Cognitive Distance, which the company says is the first non-linear, non-parametric similarity computation for making sense of massive amounts of data without requiring businesses to pre-model the data. The other is Saffron Mutual Information, which the vendor says addresses “sparse data” challenges by making it possible to perform classification with high accuracy on high-dimensional data (with tens of thousands of features).
CEO Gayle Sheppard explains that the focus is on looking for the signals that really matter from the noise of Big Data, as companies “merge outside-in intelligence with inside-out intelligence, increasing the amount of data now available for decision making.” With the new algorithms, the associations the Saffron platform makes among people, places and things data — counting an entity’s frequency and the contexts in which one thing is associated with something else — extends to discovering patterns to anticipate what happens next.