个人简介：Dr Varun Ojha is an Associate Professor of Artificial Intelligence at theSchool of Computing, Newcastle University. He is Co-Investigator on the UKRI-fundedNational Edge AI Hub. He is the lead of Safe AI at the Centre for AI Safety. He is theHead of the Intelligent Systems Research Group at Newcastle University. In the past, DrOjha served as an Assistant Professor of Computer Science at the University of Reading,UK, and as a Postdoctoral Fellow at the Swiss Federal Institute of Technology, Zurich,Switzerland. Before this, Dr Ojha was a Marie Curie Fellow (funded by the EuropeanCommission) at the Technical University of Ostrava, Czech Republic. He works inTrustworthy Artificial Intelligence, Sparsity, Explainability, and Security. He innovatesin Deep Learning, Neural Networks, Machine Learning, and Data Science. More on:ojhavk.github.io.

报告摘要: Artificial Intelligence (Al) algorithms have become an inevitable part ofour lives and are so pervasive that users often engage with them unconsciously.However, AI systems have advantages and drawbacks, with safety and sustainability assignificant concerns. AI models are vulnerable to malicious attacks, unintentionalchanges, omissions of context in training data, sensor ageing, and other unforeseencircumstances. This makes edge AI applications, such as smartphones and cars,susceptible to defects in training data and AI models. Our research in safeguarding AIfocuses on protecting data integrity and the quality of learning in edge computingenvironments and federated learning. Similarly, AI energy demand is outpacing globalenergy production, with catastrophic environmental impacts. Improving AI inferenceenergy efficiency across edge devices will support sustainability. Our research aims tomake AI resources more efficient through quantisation, pruning, distillation, and sparse model training .