Machine learning can be applied in various ways in security including malware analysis, making security-wise predictions, and detecting previously unknown attacks. Moreover, it becomes widely accessible via Machine-Learning-as-a-Service (MLaaS), a service that provides machine learning tools as part of cloud computing services. Our research includes privacy-preserving machine learning, secure distributed machine learning, federated learning, deepfake detection, backdoor attacks, membership inference, watermarking for machine learning, etc.

Can data-driven services be hosted to untrusted cloud storage while achieving secure data management? Our study includes (i) flexible and expressive access control using ID/attribute-based access control, hidden access policy, and privacy-preserving identification, and (ii) secure query processing over encrypted databases using searchable encryption, and flexible data search over encrypted data, etc.

Functional encryption is a distinguished cutting-edge technique in cryptography. Our research includes quantum-secure encryption, predicate encryption, inner-product encryption, provable security analysis, along with designing customized protocols and algorithms in pairings-based cryptography.

The pattern with which a system or program reads and writes memory on secondary storage becomes a critical security hole. Our research includes access pattern hiding techniques such as ORAM, PIR, obfuscation and/or more sophisticated leakage-abusing attacks.

Fine-grained access control over public storage is of extreme interest in various applications for secure data sharing. Our research includes privacy-enhanced and efficient construction of ID-based encryption, variants of attribute-based encryption.

Cryptocurrency has emerged as the most successful digital asset supporting anonymity. But it faces lots of threat models such as double-spending, de-anonymization, etc. Our research includes security and privacy on cryptocurrency and blockchain, blockchain-based security protocols and application,  etc.

Many web applications require their own independent security mechanisms for authentication and session management, such as validating certificates and cookie management. Our research includes detecting web cookie vulnerabilities, certificate transparency, etc.

A trusted execution environment (TEE) is a secure area of the main processor. Our study includes a TEE-based secure protocol design.