Abstract

Learning goes beyond knowledge acquisition. It is about using and refining knowledge to solve problems and enhance abilities to deliver value. It has always been an exemplary manifestation of intelligence. Human being learns based on experience and interactions. The surprises and unexpected scenarios create new and interesting opportunities for learning. Machine learning tries to mimic human way of learning to exhibit human-like behaviour. Most of the machine learning models are developed around human learning philosophies. These models include bioinspired models and probabilistic models (Floreano and Mattiussi, Bio-inspired artificial intelligence: Theories, methods, and technologies. MIT Press, 2008). Machine learning models try to get the best from quantitative abilities and connectionist intelligence of humans. Human learning has four different aspects—behaviourism, humanism, reinforcement, and social learning. Canonical theory of dynamic decision-making captures ontological, cognitive as well as relational aspects for learning. Human learning at times results in creative activities. When we look at existing ML models from a creativity perspective, many standard models fall apart. Creativity is typically defined as a very humanish act. It is the central dimension of human achievements that always fascinated scientists (Cotterill, Prog Neurobiol 64:1–33, 2001). It is about producing something new, interesting, and useful. Creative intelligence is about combination and transformation. It does carry an element of surprise and differentiation, even in high entropy and uncertain states. At some point, while going with patterns, associations, and selective combinations, it introduces ‘out of pattern’ results. Interestingly, it breaks the rules at some very unobvious but logical junction point. Is it possible to learn this ability to surprise? Is it the thing what human learns? Or creativity is just an outcome of an accident or an offshoot of routine work? Then how can machines learn to produce these surprises? These questions just try to find their routes through theories of distance measurements, connectionist models, probabilistic associations, information gain, and outliers. This chapter tries to unfold different facets of human learning and machine learning with this unexplored element of surprising creativity. It further tries to formulate creative learning models to build abilities in machines to deliver ingenious solutions.