Overview

The DACES (Data structures and Algorithms for Communications Engineering Safety) is a research group devoted to the study of data structures, algorithms, and information theory for ensuring the trust, reliability, and security of communications infrastructure. We aim to design and analyze algorithms for trustworthy systems with provable guarantees on communication cost, privacy levels, memory usage, and computational and statistical efficiency. To this end, we use a variety of mathematical and computational tools to advance the state-of-the-art. Whenever possible, we liaise with experts on socio-politics and economics to further understand the consequences of our technical work.

Topics

Here are some subject areas we devote our attention to:

• Creating and Auditing Trustworthy Statistics and Machine Learning:
  • Trust in Data Models and Curation Mechanisms
  • Attack Models for Data Privacy and Security
  • Solution Models for Data Privacy and Security
  • Attack Models for Adversarial Robustness
  • Solution Models for Robustness
• Representation, Learning, and Verification of Structured Data and Outcomes:
  • Graph Embeddings and Relational Graph Learning
  • Scientific Data Management
  • Spatiotemporal and Statistical Databases
  • Interpretable Machine Learning
  • Verification of Computational Materials Science
  • Safety of Quantum Information