Michael Karasick from the Almaden lab came up next to discuss Big Data and its role in the IBM research agenda. Almaden’s mission is “science and data to extend human capability.”
IBM invests heavily in 12 research labs around the world and focuses on pure research as well as R&D for products. Once a year IBM does a survey of interesting topics for their research agenda, culling a list > 100 to a small number of really interesting topics. As part of this they identified four mega trends/technology drivers:
- Growing scale/lower barrier of entry
- Increasing complexity yet easier to consume
- Fast pace
- Contextual overload
The 2013 research agenda identified two sets of topics:
- Rapidly evolving infrastructure
- Mobile first
- Scalable services ecosystems with cloud
- Software defined environments
- Future of Big Data and Analytics
- Multimedia and visual analytics (driving big data analytics)
- Contextual Enterprise (fusing dta and processes)
- Personalized Education
Michael focused in on two of these – contextual enterprise and multimedia analytics.
The contextual enterprise is all about building context from data dynamically at scale, discovering new value and combining structured, unstructured, static and streaming data. Combining in particular what IBM calls systems of record and systems of engagement. To make this work, IBM sees a pipeline that gathers data from lots of different sources, connects this so knowledge and learning can happen, reasoning on this to make decisions and take action. Finally you must adapt, improving over time.
Multimedia is important because so much of the growth in data is coming from voice, image and video. It takes 1000x the compute power to extract meaning from a movie,for instance, than from text so this has huge implications. Consider scenarios like using video for appraisals, healthcare imagery, using cameras in transportation and safety scenarios, using imagery of geological data in exploration and more.
All this is part of what IBM sees as the growth of cognitive systems – building on tabulating and programming systems to developing systems that make decisions and use the kind of cognitive capabilities demonstrated by Watson. One neat example,combining this with visual analysis, is a “sieve” for reviewing breast cancer imagery by using computing to identify shape, boundaries etc and highlight the stuff that should be reviewed, separating it from the less useful material.
Great to see IBM continuing with its long standing investment in research.