This is a recurring event: View all events in the series “Data Bites”
Artificial Intelligence (AI) has found a great degree of success in recent decades, mostly due to the availability of vast amounts of data and processing power.
While AI models have shown incredible capabilities in processing data beyond what humans are capable, they still struggle to reliably show the same level of reasoning capabilities as humans.
Neural-Symbolic AI, i.e. the integration of artificial neural networks and symbolic reasoners, has led to the development of AI models that are capable of processing distributed data, while maintaining knowledge representation on the symbolic level.
Neural-Symbolic models have been shown to be more explainable, robust to noisy data, capable of generalizing beyond the training data, as well as learning from less data and using less processing power.
The key aspect of Neural-Symbolic AI that allows for these improvements in performance is the ability to not only learn from data, but also to reason about what has been learned.
About the Speaker:
Sofoklis Kyriakopoulos is currently a PhD student at City, University of London. His research focuses on the application of Neural-Symbolic AI models in continual learning scenarios.
He has a MSc of Artificial Intelligence, a MSc of Computer Engineering, and over 5 years of industry experience as a freelance computer engineer.
Attendance at City events is subject to our terms and conditions.