The Use Of Cognitive Computing Technologies At Work
Based on recent research & experiences, in the following we focus on the use of the IBM Watson platform for it is giving a significant contribution to the recent growth of cognitive computing applications, being straight accessible without the need of investments in both costly hardware and software. Moreover, owing to the IBM Academic Initiative, universities can freely experiment Watson through the services and Application Programming Interfaces (API), made available within the IBM Bluemix platform. In Addition, Watson is a fully featured cognitive computing system for the research and development of cognitive systems and services, which can easily interact with other applications and systems as well. It is a valuable tool, which includes the set of APIs within a varied catalog of building blocks available, which allow building advanced cognitive applications. For example, among the others, we cite the following boilerplates that can help enhance, scale, and accelerate human expertise: (i) Concept Expansion, which performs text analysis and can learn similar terms as well as words or phrases, based on context. Such an apparatus empowers clients to quickly make a vocabulary and an arrangement of related terms from informational indexes of content pieces or accumulations of records. Then, the output can be used to provide further understanding of data and improve text analytics pipelines; (ii) Concept Insights, which looks for associations of concepts inside sets of documents provided by users with a pre-existing graph of concepts based on the renowned free encyclopedia Wikipedia. Accordingly, the service identifies links of two types: a. explicit links in the case a document directly mentions a concept, b. implicit links, which connect the input documents to relevant concepts that are not directly mentioned. This service can also search for documents that are relevant to a concept or collection of concepts by exploring both the explicit and the implicit links; (iii) Dialog, which allows developers designing the interaction mechanisms of an application with an end-user, based on a natural conversational interface.
In practice, this service enables computer applications to use natural language and this capability can be profitably exploited in a variety of situations such as, e. g. , automatically respond to user questions, walk users through processes or applications, or even hand-hold users through difficult tasks;
- Natural Language Classifier, which applies cognitive computing techniques to analyze sentences or phrases in a given corpus and return the best matching classes;
- Relationship Extraction, which parses sentences into their various components, looking for relationships between the components;
- Speech-to-text and Text-to-speech, the former converts the human voice into the written word and the latter processes text and natural language to generate synthesized audio output complete with appropriate cadence and intonation even if on a limited set of languages;
- Tone Analyzer, which leverages cognitive linguistic analysis to identify tones that people show in their languages;
- Tradeoff Analytics, which helps people make better choices while considering multiple, often conflicting, goals that matter when making that choice.
The straight availability of such a quantity of services with advanced functionalities, opens new opportunities and enables both designers and programmers to realize solutions. As an example, let us consider the most famous demonstration of the IBM Watson’s ability, that is its successful competition in the Jeopardy! Game. The victory was possible, owing to the capacity to answer open questions of unlimited domain, crawling both structured and unstructured data. In most cases multiple answers emerge and the cognitive computing system provides a confidence estimation for each. The key feature is mastering the language because the game offer clues through open questions. These clues are to be analyzed because full of subtle meanings, irony and other complexities of natural language.
From a more technological point of view, Watson owes its smartness to the DeepQA Project, a massively parallel probabilistic evidence-based architecture, which can also be adapted to different business applications and additional exploratory challenge problems including medicine, enterprise search, and gaming.