A Marriage of Neurons and Symbols – Opportunities and Obstacles

Research interest in neuro-symbolic artificial intelligence (NSAI) is on the rise. Currently, most researchers and artificial intelligence (AI) practitioners rely on deep learning (DL) algorithms as solutions to their tasks. The drawbacks and limitations have been widely documented, ranging from extremely resource-hungry processes to ambiguous and complex inner mechanics rendering a certain lack of interpretability. The modern generation acknowledges the importance of producing systems that we can fully understand. We also understand the significance of developing efficient processes that can achieve high-performance levels with fewer resources. For example, the craze of cryptocurrency (especially Bitcoin) mining saw a huge spike in computing electricity consumption worldwide. Forbes estimates that Bitcoin mining consumes roughly 127 terawatt-hours of electricity annually. To put this into perspective, that is greater...