IBM is expanding its roster of Watson online services that can be deployed by developers to enhance their apps with analysis and data processing capabilities.
Launched in October 2013, the Watson Developer Cloud has so far attracted more than 5000 users, who have incorporated the services into more than 6000 apps. At present, all the Watson Developer Could services are available in beta form, so they are free to use.
Five new services announced Thursday are all designed to be embedded within mobile or desktop apps through APIs (application programming interfaces), and provide advanced functionality that otherwise would have been be too time-consuming and laborious for developers to create from scratch.
The new services can convert text to speech, convert speech to text, recognize visualize images, search a set of data by concept, and help users make decisions. They join eight services that IBM introduced a few months ago as part of its Watson Developer Cloud.
The Speech to Text service provides a way for mobile applications to receive input from the voices of their users, in much the same way Siri works on Apple devices. IBM claims the service has very low latency to ensure quick responsiveness. It uses machine learning techniques to recognize the quirks of grammar and language structures.
IBM also offers a service that converts text to speech. The same voice used by the Watson supercomputer that won the 2011 Jeopardy match can now be planted on developers' apps, offering to voice material in either English or Spanish.
The Visual Recognition service can analyze a picture and video and return a list of words summarizing the visual material. The service can recognize objects, settings and events. Such data could be used to help classify images, potentially saving a lot of manual labor, and use them as part of a larger workflow.
IBM's new new Concept Insights search service is novel in that it allows users to search for concepts rather than specific keywords. For each search, it can return a list of explicit results -- where the material contains the exact keywords the user entered -- as well as implicit results, where the documents returned match more general concepts underlying the search terms.
The fifth new service, called Tradeoff Analytics, dynamically weighs multiple options in order to decide which one is the best choice for the user. It uses a statistical analysis called Pareto filtering to weigh different choices in relation to each other.
Tradeoff Analytics could be used, for instance, to help choose treatment options for medical patients, or help a car buyer decide which vehicle to purchase.
IBM is not the only enterprise technology company to offer API-based functionality as a stand-alone service. Hewlett-Packard also offers a set of embeddable APIs for services derived from its Autonomy Idol analysis software. Idol OnDemand includes services such as image detection, face detection, entity extraction and barcode reading, all of which can be planted with applications.
IBM has not announced pricing for the Watson services, or given a timetable for general availability.