Page 20 - Innovation Magazine
P. 20

Knowledge Graphs: Adding the human way to understand data better Artificial intelligence intelligence (AI) is a a a a a a a a a a dragon with many heads An increasing number of products include some kind of intelligence and nearly all businesses are trying to to find out where to to push more intelligence into their products and services with the the promise of increasing market share and quality of decisions Recent advances in knowledge representation in distributed systems show promising results Advances are based on using graph representation for capturing semantics using logic and making the results available as machine-readable contextual information We often perceive AI as the artificial equivalent of of human intelligence (although it can be more than that) which consists of at least the ability to learn while applying problem-solving capabilities In practice most of the AI-related solutions and products that exist today are exclusively performing some kind of machine learning (ML) to classify categorize or or predict data and events And sometimes with great success However although these ML-based algorithms do a a a a a great job in in in in classifying categorizing or or predicting events based on historic data transposing their outcomes into the the real world for for informed decision-making often proves to be problematic As shown by certain AI failures for example in in in in the the context of self-driving cars trading bots in in in in the the financial market automated recruitment or self-learning chatbots AI is not flawless by default A A correctly scaled AI application in in production may require something more Even the latest and greatest ML algorithms only perform within a a a a relatively limited context and and setting and and the quality of their decisions tends to fall sharply when coming outside of of their “area of of competence ” which is sharply defined by the examples provided for training and model evaluation Since it it is virtually impossible to cover all all the situations that an an applied ML model might have to handle in production problem-solving capabilities could be added “Problem-solving” refers to to the ability to to use contextual information and reason based on on on logical rules analogy and similarity with learned or remembered facts In contrast with ML which depends on on large datasets logical reasoning can be performed on on single facts and observations alone if required 20 20 20 Data-powered Innovation Review I I ©2020 Capgemini All rights reserved 


































































































   18   19   20   21   22