Dr Reza Shokri

The Data Privacy and Trustworthy Machine Learning research laboratory, led by NUS Presidential Young Professor Dr Reza Shokri, pioneers research in an emerging field – trustworthy machine learning.

“Advanced AI has enabled many new technologies.  However, there are many risks in using machine learning on sensitive data and in critical decision-making processes. AI systems need to be trained on a large amount of data to learn how to make accurate decisions in a given task (for example, automatically translating documents). The question is: Can we trust AI algorithms to have access to our personal data (for example, when they are used in medical and financial applications, or when they are used in our smartphones)? What if they leak our sensitive information to untrusted entities? Can we trust AI systems to be robust against hackers and adversarial manipulations (for example, when they are used in a variety of applications in smart cities)?” explains Dr Shokri. We know that computer systems can be very vulnerable to attacks. AI systems are no exception. So, although we want to take advantage of the power of AI algorithms, we need to make sure they are safe and trustworthy.

The lab, which has received support from donors such as Intel, Meta (Facebook), Google, and VMWare, develops theoretical foundations of data privacy in AI, designs algorithms for enabling machine learning across organization, designs ethical and fair AI, and works on making AI technologies more transparent and accountable. The lab also develops tools to enable auditing privacy for personal data in AI algorithms: The tool, named ‘privacy meter’, is being used to quantitatively evaluate the privacy risks of AI algorithms, across a wide range of systems. The lab is also very active in educating students and researchers on ethical and responsible computing. This helps designing systems which are, by design, safe, ethical and privacy-preserving.

We are very grateful to donors for helping us to achieve our goals," expressed Prof Shokri. 

Google also supports PhD Fellowships to recognise outstanding graduate students at NUS doing exceptional and innovative research in areas relevant to computer science and related fields.