�ݺ�ߣshows by User: MohammadJafarMashhad

�ݺ�ߣshows by User: MohammadJafarMashhad / http://www.slideshare.net/images/logo.gif �ݺ�ߣshows by User: MohammadJafarMashhad / Sun, 04 Oct 2020 03:00:02 GMT �ݺ�ߣShare feed for �ݺ�ߣshows by User: MohammadJafarMashhad Hybrid Deep Neural Networks to Infer State Models of Black-Box Systems /slideshow/hybrid-deep-neural-networks-to-infer-state-models-of-blackbox-systems/238729522 asepresentationslides-201004030002
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%. ]]>
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%. ]]> Sun, 04 Oct 2020 03:00:02 GMT /slideshow/hybrid-deep-neural-networks-to-infer-state-models-of-blackbox-systems/238729522 MohammadJafarMashhad@slideshare.net(MohammadJafarMashhad) Hybrid Deep Neural Networks to Infer State Models of Black-Box Systems MohammadJafarMashhad 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%. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/asepresentationslides-201004030002-thumbnail.jpg?width=120&height=120&fit=bounds" /><br> 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%.

Hybrid Deep Neural Networks to Infer State Models of Black-Box Systems from Mohammad Jafar Mashhadi

]]> 150 0 https://cdn.slidesharecdn.com/ss_thumbnails/asepresentationslides-201004030002-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 Black-box Behavioral Model Inference for Autopilot Software Systems /slideshow/blackbox-behavioral-model-inference-for-autopilot-software-systems/238433901 mscpresentation-200909212527
MSc thesis oral exam presentation.]]>
MSc thesis oral exam presentation.]]> Wed, 09 Sep 2020 21:25:27 GMT /slideshow/blackbox-behavioral-model-inference-for-autopilot-software-systems/238433901 MohammadJafarMashhad@slideshare.net(MohammadJafarMashhad) Black-box Behavioral Model Inference for Autopilot Software Systems MohammadJafarMashhad MSc thesis oral exam presentation. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/mscpresentation-200909212527-thumbnail.jpg?width=120&height=120&fit=bounds" /><br> MSc thesis oral exam presentation.

Black-box Behavioral Model Inference for Autopilot Software Systems from Mohammad Jafar Mashhadi

]]> 164 0 https://cdn.slidesharecdn.com/ss_thumbnails/mscpresentation-200909212527-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 An Empirical Study On Practicality Of Specification Mining Algorithms On A Real-world Application /slideshow/an-empirical-study-on-practicality-of-specification-mining-algorithms-on-a-realworld-application/238279907 icpc19presentation-200827003143
An Empirical Study on Practicality of Specification Mining Algorithms on a Real-world Application at the International Conference on Program Comprehension (ICPC) 2019]]>
An Empirical Study on Practicality of Specification Mining Algorithms on a Real-world Application at the International Conference on Program Comprehension (ICPC) 2019]]> Thu, 27 Aug 2020 00:31:43 GMT /slideshow/an-empirical-study-on-practicality-of-specification-mining-algorithms-on-a-realworld-application/238279907 MohammadJafarMashhad@slideshare.net(MohammadJafarMashhad) An Empirical Study On Practicality Of Specification Mining Algorithms On A Real-world Application MohammadJafarMashhad An Empirical Study on Practicality of Specification Mining Algorithms on a Real-world Application at the International Conference on Program Comprehension (ICPC) 2019 <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/icpc19presentation-200827003143-thumbnail.jpg?width=120&height=120&fit=bounds" /><br> An Empirical Study on Practicality of Specification Mining Algorithms on a Real-world Application at the International Conference on Program Comprehension (ICPC) 2019

An Empirical Study On Practicality Of Specification Mining Algorithms On A Real-world Application from Mohammad Jafar Mashhadi

]]> 182 0 https://cdn.slidesharecdn.com/ss_thumbnails/icpc19presentation-200827003143-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 Machine Learning Workshop /MohammadJafarMashhad/machine-learning-workshop-238279699 machinelearningworkshop-200827002808
I used to deliver a machine learning workshop for a group of engineering school graduate students]]>
I used to deliver a machine learning workshop for a group of engineering school graduate students]]> Thu, 27 Aug 2020 00:28:07 GMT /MohammadJafarMashhad/machine-learning-workshop-238279699 MohammadJafarMashhad@slideshare.net(MohammadJafarMashhad) Machine Learning Workshop MohammadJafarMashhad I used to deliver a machine learning workshop for a group of engineering school graduate students <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/machinelearningworkshop-200827002808-thumbnail.jpg?width=120&height=120&fit=bounds" /><br> I used to deliver a machine learning workshop for a group of engineering school graduate students

Machine Learning Workshop from Mohammad Jafar Mashhadi

]]> 38 0 https://cdn.slidesharecdn.com/ss_thumbnails/machinelearningworkshop-200827002808-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 https://cdn.slidesharecdn.com/profile-photo-MohammadJafarMashhad-48x48.jpg?cb=1599354117 mjafar.me https://cdn.slidesharecdn.com/ss_thumbnails/asepresentationslides-201004030002-thumbnail.jpg?width=320&height=320&fit=bounds slideshow/hybrid-deep-neural-networks-to-infer-state-models-of-blackbox-systems/238729522 Hybrid Deep Neural Net... https://cdn.slidesharecdn.com/ss_thumbnails/mscpresentation-200909212527-thumbnail.jpg?width=320&height=320&fit=bounds slideshow/blackbox-behavioral-model-inference-for-autopilot-software-systems/238433901 Black-box Behavioral M... https://cdn.slidesharecdn.com/ss_thumbnails/icpc19presentation-200827003143-thumbnail.jpg?width=320&height=320&fit=bounds slideshow/an-empirical-study-on-practicality-of-specification-mining-algorithms-on-a-realworld-application/238279907 An Empirical Study On ...