Utilizing cognitive computing and AI in Knowledge Management
Cognitive systems and artificial intelligence are changing the way organizations search for and discover important information and how it is analyzed to create knowledge.
A new era in computing and information management is on the horizon.
KMWorld recently held a webinar with David Seuss, chief executive officer, Northern Light Group, and Amy Guarino, chief operating officer, Kyndi, Inc who discussed how cognitive computing and AI are setting the stage for next-gen KM.
According to Seuss, currently, 80% of U.S. CEOs believe that AI will change the way they do business in the next five years. However, there are challenges: defining an AI strategy, finding AI-literate workers, getting data AIready, and ensuring AI is trustworthy and responsible.
Delivering on the promise of machine learning requires lining up many disparate pieces across many domains, Seuss said.
Before diving into cognitive computing, there are several steps that should be taken:
- Pick a problem that can be solved using techniques and resources available today
- Pick a problem that is central to an important business process
- Pick a problem that cannot be solved any other way
- You have to provide all the parts associated with the solution, combining multiple machine learning techniques as well as providing for the non-machine learning components
The problem most businesses face now is that unstructured data creates workplace inefficiencies, Guarino explained.
The challenges with current approaches include training data, language/text, and explanation. With Big data issues of volume and velocity enterprises can’t find what they want. There is too much information for humans to read and we can’t see the hidden relationship through the forest of big data, Guarino said.
Kyndi can help, Guarino said. The platform can:
- Gather documents related to a particular area of interest
- Ingest, convert to text
- Create cognitive memory graph
- Generate proto-ontology
- Calibrate as needed
- Generate signals
An archived on-demand replay of this webinar is available here.