The AAAI-22 Workshop on

Engineering Dependable and Secure

Machine Learning Systems

March 1, 2022

held at the thirty-sixth AAAI Conference on Artificial Intelligence (AAAI-22)

Nowadays, machine learning solutions are widely deployed. Like other systems, ML systems must meet quality requirements. However, ML systems may be non-deterministic; they may re-use high-quality implementations of ML algorithms; and, the semantics of models they produce may be incomprehensible. Consequently, standard notions of software quality and reliability such as deterministic functional correctness, black box testing, code coverage, and traditional software debugging become practically irrelevant for ML systems. This calls for novel methods and new methodologies and tools to address quality and reliability challenges of ML systems.

In addition, broad deployment of ML software in networked systems inevitably exposes the ML software to attacks. While classical security vulnerabilities are relevant, ML techniques have additional weaknesses, some already known (e.g., sensitivity to training data manipulation), and some yet to be discovered. Hence, there is a need for research as well as practical solutions to ML security problems.

With these in mind, this workshop solicits original contributions addressing problems and solutions related to dependability, quality assurance and security of ML systems. The workshop combines several disciplines, including ML, software engineering (with emphasis on quality), security, and game theory. It further combines academia and industry in a quest for well-founded practical solutions. Topics of interest include, but are not limited to:

  • Vulnerability, sensitivity and attacks against ML

  • Adversarial ML and adversary-based learning models

  • Strategy-proof ML algorithms

  • Case studies of successful and unsuccessful applications of ML techniques

  • Correctness of data abstraction, data trust

  • Choice of ML techniques to meet security and quality

  • Size of the training data, implied guaranties

  • Application of classical statistics to ML systems quality

  • Sensitivity to data distribution diversity and distribution drift

  • The effect of labeling costs on solution quality (semi-supervised learning)

  • Reliable transfer learning

  • Software engineering aspects of ML systems and quality implications

  • Testing of the quality of ML systems over time

  • Debugging of ML systems

  • Quality implication of ML algorithms on large-scale software systems