Anrin Chakraborti | University of Illinois at Chicago

Research Interests

I build cryptographically secure systems and design protocols that address privacy concerns in emerging technologies. My research combines cryptography with system design, and from this vantage point, I work on a variety of topics which includes

Secure Cloud Computing: I research different aspects of cloud security, including secure storage technologies, databases, and virtualization. I design cryptographic mechanisms that protect confidentiality of cloud-hosted data against side-channels resulting from access pattern leakage, and build storage and database technologies that data protect integrity against malicious service providers.
Information Sharing in Collaborative Learning and Cyber-Physical Systems: I research privacy-preserving mechanisms for sharing information in collaborative learning and anomaly detection pipelines. These tools enable parties to jointly detect coordinated attacks and reduce false positives. To this end, on one hand, I develop new theory for private fuzzy matching of vectors (e.g., feature vectors corresponding to attack identifiers) based on their distances in a metric space. On the other hand, I build new tools to compare outputs of vision sensors, e.g., LIDAR points clouds, enabling better visibility and fault detection in connected autonomous vehicles (CAVs).
Deniable Communication and Storage: I research storage and communication systems that retain privacy of sensitive information – either stored at rest with full-disk encryption, or in transit with end-to-end encryption – even when the encryption keys are leaked due to coercion, backdoors, vulnerabilities, etc. To this end, I build block stores, filesystems and firmware for secure storage, and propose mechanisms to make end-to-end encrypted communication systems more resilient to backdoors and key compromise.

As part of my research, I often find myself trying to answer questions such as

how do we scale cryptographic tools to handle new data types and computational tasks in clouds and big data?
how can we optimize cryptographic protocols to enable reliable and private communication between low-resource parties, e.g., cyber-physical systems?
how do we make private end-to-end communication (and storage) truly private in the face of intrusive privacy laws, censorship, coercion and backdoors?