This paper assesses the fault-detection capabilities of modern deep-learning models. It highlights that a naive deep-learning approach optimized for accuracy is unsuitable for learning fault-detection models from time-series data. Consequently, out-of-the-box deep-learning strategies may yield impressive accuracy results but are ill-equipped for real-world applications. The paper introduces a methodology for estimating fault-detection delays when no oracle information on fault occurrence time is available. Moreover, the paper presents a straightforward approach to implicitly achieve the objective of minimizing fault-detection delays. This approach involves using pseudo-multi-objective deep optimization with data windowing, which enables the utilization of standard deep-learning methods for fault detection and expanding their applicability. However, it does introduce an additional hyperparameter that needs careful tuning. The paper employs the Tennessee Eastman Process dataset as a case study to demonstrate its findings. The results effectively highlight the limitations of standard loss functions and emphasize the importance of incorporating fault-detection delays in evaluating and reporting performance. In our study, the pseudo-multi-objective optimization could reach a fault-detection accuracy of 95% in just a fifth of the time it takes the best naive approach to do so.

The field of Artificial General Intelligence needs a simple and understandable measure of progress to aid in passing the difficult Turing Test. The proposed Ladder Metric introduces a sequence of levels that intelligent agents must climb in order to reach the final, Turing indistinguishable, level. Just like the Turing test, the levels can only be defined in terms of general properties and behaviour rather than specific operational requirements. To track the progress and make comparisons one must climb and determine the achieved rung of the proposed ladder.