Recent advances in AI, and specifically cognitive computing had received a lot of attention and interest from researchers and practitioners. One notable area, which particularly has generated a lot of interest, is cognitive assistants, specifically with the recent proliferation of intelligent apps in personal assistant space including Siri, Google Now, Cortana, and others.
While there is a significant progress made in the development of personal assistants, there are still many open challenges and the need for innovation to enable the development of cognitive assistants, specifically in enterprise and government context. The following slides were presented in a discussion group, in which I tried to review the opportunities, gaps/challenges and share lesson learned from Watson Jeopardy challenge experience on how to build a coalition and partnership between industry, academia and government to create an Open Collaboration Platform to tackle a such big challenge, see here:
In reviewing these, and comparing human intelligence in terms of cognitive abilities with machine intelligence, one fundamental question is whether the same level of human cognitive abilities (discussed in slide 10 above) is needed by a cognitive agent? And, if not, how do we characterize the cognitive skills needed by a cogs, and in particular personal cogs, work-focused cogs and specialized/expert cogs? And, in general, how the division of the cognitive skills of human and machine would look like in order to realize the partnership (augmenting human intelligence)?
And, a related discussion point, is building on Jeopardy! DeepQA challenge experience (mentioned in slide 22), in forming and supporting an open collaboration model between academia, industry and government on cognitive assistance. How to enable and grow such an open collaboration platform where visions, data, knowledge/expertise, and interoperable artifacts (algorithms and APIs) can be shared to support advancing cognitive assistance vision?