Inspect, Understand, Overcome: A Survey of Practical Methods for AI Safety
It addresses safety concerns for machine learning experts and safety engineers in applications like mobile health and autonomous driving, but is incremental as it synthesizes existing research.
This survey tackles the problem of AI safety in deep neural networks by providing a structured overview of practical methods for detecting, quantifying, or mitigating model shortcomings, such as lack of generalization and interpretability, without presenting new experimental results.
The use of deep neural networks (DNNs) in safety-critical applications like mobile health and autonomous driving is challenging due to numerous model-inherent shortcomings. These shortcomings are diverse and range from a lack of generalization over insufficient interpretability to problems with malicious inputs. Cyber-physical systems employing DNNs are therefore likely to suffer from safety concerns. In recent years, a zoo of state-of-the-art techniques aiming to address these safety concerns has emerged. This work provides a structured and broad overview of them. We first identify categories of insufficiencies to then describe research activities aiming at their detection, quantification, or mitigation. Our paper addresses both machine learning experts and safety engineers: The former ones might profit from the broad range of machine learning topics covered and discussions on limitations of recent methods. The latter ones might gain insights into the specifics of modern ML methods. We moreover hope that our contribution fuels discussions on desiderata for ML systems and strategies on how to propel existing approaches accordingly.