Intuitive Explanation of Trends of AI/ML for CDOs
isCDO Webinar Presentation May 27, 2021, 0800PST
James C S Meng, Senior Scientist, SuperComputer Center, University of California, San Diego, California, USA
A summary of historical trends is presented based on the AI/ML building blocks and major drivers and challenges for its elements. The AI/ML are not magic. AI/ML are tools for data analytics, it reliably free up human labor for more meaningful and productive tasks. Data managers must be knowledgeable to leverage AI/ML in enterprise data transformation. This presentation aims to explain AI/ML to CDOs from a holistic and evolutionary perspective.
We start with explaining what is AI/ML? and principles of AI/ML. AI/ML are software that reads and reasons making human-generated text, machine-readable, real-time at scale. It enables users to understand volumes of human-generated text by turning it into something that a machine can read & process. AI/ML are computer programs which can carry out a simulation of the human brain. AI/ML is about problem solving using computers for discovery & automation.
Key AI/ML building blocks are: Representation of commonsense knowledge; Translation into computer operations; Model human problem solving by using rules consistent of conditions (antecedent) and actions (consequence). By analyzing the relations, inference net is generated among conditions and actions. Logic-oriented constraint net allows dependency-directed, non-chronological backtracking to sort out the contradictions imbedded in common knowledge assumptions. In summary, building blocks of AI/ML are knowledge representation and automated reasoning.
Drivers for emergence of AI/ML and relevance to CDO’s: First, Traditional data processing is saturated and overwhelmed; data is being generated at record levels, data growth rate is exponential. It is certain that the pace at which the world creates data will never be this slow again. Second, enterprise need AI/ML to stay competitive in the ever-faster-changing world.
AI/ML performs exceptionally fast, accurately and at scale. Structured data analytics can describe and explain what’s happened, & unstructured data analytics can explain why it’s happening. Together CDOs get the whole picture. Without both, you’re half blind.