DENAS: Automated Rule Generation by Knowledge Extraction from Neural Networks
Deep neural networks (DNNs) have been widely applied in the software development process to automatically learn patterns from massive data. However, many applications still make decisions based on rules that are manually crafted and verified by domain experts due to safety or security concerns.
In this paper, we aim to close the gap between DNNs and rule-based systems by automating the rule generation process via extracting knowledge from well-trained DNNs. Existing techniques with similar purposes either rely on specific DNNs input instances or use inherently unstable random sampling of the input space. Therefore, these approaches either limit the exploration area to a local decision-space of the DNNs or fail to converge to a consistent set of rules. The resulting rules thus lack representativeness and stability.
In this paper, we address the two aforementioned shortcomings by discovering a global property of the DNNs and use it to remodel the DNNs decision-boundary. We name this property as the activation probability, and show that this property is stable.
With this insight, we propose an approach named DENAS including a novel rule-generation algorithm. Our proposed algorithm approximates the non-linear decision boundary of DNNs by iteratively superimposing a linearized optimization function.
We evaluate the representativeness, stability, and accuracy of DENAS against five state-of-the-art techniques (LEMNA, Gradient, IG, DeepTaylor, and DTExtract) on three software engineering and security applications: Binary analysis, PDF malware detection, and Android malware detection. Our results show that DENAS can generate more representative rules consistently in a more stable manner over other approaches. We further offer case studies that demonstrate the applications of DENAS such as debugging faults in the DNNs and generating signatures that can detect zero-day malware.
Wed 11 Nov Times are displayed in time zone: (UTC) Coordinated Universal Time change
|17:30 - 17:32|
|AMS: Generating AutoML Search Spaces from Weak Specifications|
Jose CambroneroMassachusetts Institute of Technology, USA, Jürgen CitoTU Wien and MIT, Martin RinardMassachusetts Institute of Technology, USADOI
|17:33 - 17:34|
|Continuous Experimentation on Artificial Intelligence Software: A Research Agenda|
Visions and ReflectionsDOI
|17:35 - 17:36|
|DENAS: Automated Rule Generation by Knowledge Extraction from Neural Networks|
SiminChen University of Texas at Dallas, USA, Soroush BateniUniversity of Texas at Dallas, USA, Sampath GrandhiUniversity of Texas at Dallas, USA, Xiaodi LiUniversity of Texas at Dallas, USA, Cong LiuUniversity of Texas at Dallas, USA, Wei YangUniversity of Texas at Dallas, USADOI
|17:37 - 17:38|
|On Decomposing a Deep Neural Network into ModulesACM SIGSOFT Distinguished Paper Award|
Research PapersDOI Media Attached
|17:39 - 17:40|
|Synthesizing Correct Code for Machine Learning Programs|
Student Research Competition
Joshua GisiNorth Dakota State University, USADOI
|17:41 - 18:00|
|Conversations on ML Model Building|