Recently , potential-based approaches are attracting substantial focus within the computational intelligence field . Distinct from conventional deep learning architectures , these designs specify a chance set not directly , but via a sophisticated energy function . This enables for representing ext