Cognitive computing systems help to remove human biases and augment decisions, as they rely on a broad information base, says technology and software multinational IBM Cognitive Solutions GM Jay Bellissimo.
Cognitive systems understand content in context, reason with purpose, learn at scale and interact naturally with humans. They comprise sophisticated algorithms that compare, correlate and draw on massive pools of structured and unstructured data – everything from numbers to images, video and social media content.
IBM’s Watson cognitive computing system uses more than 100 complex algorithms that regulate how Watson determines the relevance of data and the outcomes that are most consistent with the query. These algorithms are designed based on IBM’s rich technical and research expertise, and refined in close collaboration with experts in various disciplines and industries.
“The logic of the experts, the data they use in their processes, their method of interrogating data and the precedents they rely on are used as the basis for algorithm design. “We are trying to ‘teach’ Watson how these industry experts and virtuosi think, which means designing the algorithms to use similar methods, but at a much more massive scale, including using data that may be relevant but not necessarily specific to a discipline.”
This ability to query massive amounts of data specific to a topic, and make connections between data that may not be directly related, but are relevant to a topic, is a key benefit of Watson. It removes the problem of bias as a result of small data samples and, because the algorithm outputs are probabilistic, generating hypotheses rather than hard and fast answers, it minimises human bias.
The IBM Watson Health team has a dedicated oncology unit, and Watson is used in several oncology centres worldwide. Experts at these centres use Watson to help them to identify potential new precision treatment options, or map diagnoses against possible treatments and success rates, based on published scientific and medical data. This helps doctors to make evidence-based, individualised decisions for their patients with the help of insights from a massive pool of specific and relevant data that underpin a diagnosis, decision and treatment.
Watson is also accelerating scientific research, similar to informatics methods, by helping researchers leverage massive volumes of relevant scientific and medical research studies and identify new patterns in the data. This could lead to breakthroughs in discovery and potential new treatment options for patients.
Cognitive systems are generally provided in a cloud format because they require powerful processing resources. This enables large businesses to implement these systems in their enterprise cloud systems and democratises their use because they can be accessed remotely.
“Watson does not replace human intelligence; it augments it. Cognitive systems are still in the early phase of development. “We believe that they will contribute massively to augmenting human capability in future. Right now, cognitive systems like Watson are being used in complex disciplines and industries to augment the expertise of people through the interrogation of massive datasets,” adds Bellissimo.
He notes that IBM is providing businesses with cognitive solutions to improve their core and peripheral functions, and that the firm foresees cognitive systems improving many public and commercial services.