IBM has launched a collaborative research initiative including Carnegie Mellon University’s Eric Nyberg that will advance the development and deployment of cognitive computing systems. Like IBM’s famed Watson, these systems can learn, reason and help human experts make complex decisions involving extraordinary volumes of data.
MIT, New York University and Rensselaer Polytechnic Institute also are part of the initiative, which will lay the groundwork for the Cognitive Systems Institute, which IBM envisions will comprise universities, research institutes and IBM clients.
Nyberg and other members of the initiative will study enabling technologies and methods for building a new class of systems that better enable people to interact with Big Data. Nyberg, a professor in the Language Technologies Institute, will lead development of architectures that would allow these systems to support intelligent, natural interaction with all kinds of information in support of complex human tasks.
"The cost-effective creation of cognitive systems for complex analytic tasks will require fundamental advances in the rapid construction, optimization, and constant adaptation of large ensembles of analytic components,” Nyberg said. “Personalized information agents will rapidly adapt and optimize their task performance based on direct interaction with the end user.”
Nyberg is a leading researcher in the field of question-answering systems. He pioneered the Open Advancement of Question Answering (OAQA), an architecture and methodology for accelerating collaborative research in automatic question answering. Notably, OAQA supported the Watson system, which bested human champions on the TV game show “Jeopardy!”
"IBM has demonstrated with Watson that cognitive computing is real and delivering value today," said Zachary Lemnios, vice president of strategy for IBM Research. "It is already starting to transform the ways clients navigate Big Data and is creating new insights in healthcare, how research can be conducted and how companies can support their customers. But much additional research is needed to identify the systems, architectures and process technologies to support a new computing model that enables systems and people to work together across any domain of expertise."