New publication at ASE 2020 conference
I am proud to announce that my latest paper, Hybrid Deep Neural Networks to Infer State Models of Black-Box Systems, is accepted for publication at ACM/IEEE conference ASE ‘20.
In that work, I created a deep neural network architecture to detect internal state of an autopilot system (developed by our industry partner, MicroPilot Inc), as a blackbox system. For more details, I encourage you to read the paper, available here. This is the first milestone of this project, later after submitting this paper I worked on replicating this paper on another case study as well as improving it on other aspects as well. The results of the next extensions can be read in my MSc thesis. Source code (incl. docs) and data are publicly available on Github as well. [1], [2].
Due to the current devastating COVID situation, I will be presenting it online. Presentation slides and probably a recording of the presentation will become available later, check publications page on this website out for more information.
Abstract:
Inferring behavior model of a running software system is quite useful for several automated software engineering tasks, such as program comprehension, anomaly detection, and testing. Most existing dynamic model inference techniques are white-box, i.e., they require source code to be instrumented to get run-time traces. However, in many systems, instrumenting the entire source code is not possible (e.g. when using black-box third-party libraries) or might be very costly. Unfortunately, most black-box techniques that detect states over time are either univariate, or make assumptions on the data distribution, or have limited power for learning over a long period of past behavior. To overcome the above issues, in this paper, we propose a hybrid deep neural network that accepts as input a set of time series, one per input/output signal of the system, and applies a set of convolutional and recurrent layers to learn the non-linear correlations between signals and the patterns, over time. We have applied our approach on a real UAV auto-pilot solution from our industry partner with half a million lines of C code. We ran 888 random recent system-level test cases and inferred states, over time. Our comparison with several traditional time series change point detection techniques showed that our approach improves their performance by up to 102%, in terms of finding state change points, measured by F1 score. We also showed that our state classification algorithm provides on average 90.45% F1 score, which improves traditional classification algorithms by up to 17%.
