Cognitive computing

Cognitive Computing describes systems that learn at scale, reason with purpose, and interact with humans naturally rather than being explicitly programmed. An official definition of cognitive computing is yet to be agreed upon; however, it is generally seen as a branch of artificial intelligence developed to think and work alongside humans. It is a mixture of both computer science and cognitive science and works to build computerized models that simulate human thought in complex situations where the answers may be ambiguous and uncertain. Cognitive computing is able to weigh complex, conflicting, and changing information contextually and offer the best-fitting solution—not necessarily the most algorithmically correct one.

Closely linked to artificial intelligence, cognitive computing works from many of the same building blocks. However, cognitive computing takes into account human elements. AI systems take in data and the algorithm over a long period of time learns variables and predicts outcomes. Cognitive computing typically describes AI systems that specifically aim to simulate human thought. This involves real-time analysis of the environment, context, intent, and many other variables to inform a person's ability to solve problems. Cognitive models for mimicking human thought include a number of AI technologies including machine learning, deep learning, neural networks, NLP, and sentiment analysis.

The term is associated with IBM's cognitive computer system, Watson. IBM has been developing cognitive computing technology for decades, combining more than a dozen disciplines of advanced computer science.

The Cognitive Computing Consortium, a forum bringing the community together to advance cognitive computing work, defined four specific characteristics common to systems:

• Adaptive—flexible enough to learn as information changes and goals evolve.
• Interactive—users must be able to interact with cognitive machines and these systems must also operate with existing hardware.
• Iterative and Stateful—able to recognize when a problem is not solvable and find answers necessary to solve it. Performed through remembering and comparing past situations.
• Contextual—able to understand context and use data in both structured and unstructured forms.

Cognitive computing use cases include the following:

• Healthcare decisions
• Investment strategies
• Enabling visually impaired users to navigate physical spaces