Data Sanity Check for Deep Learning Systems via Learnt Assertions

Data Sanity Check for Deep Learning Systems via Learnt Assertions Haochuan Lu †, Huanlin Xu †, Nana Liu †, Y angfan Zhou †, Xin Wang † School of Computer Science, Fudan University, Shanghai, China † Shanghai Key Laboratory of Intelligent Information Processing, Shanghai, China Abstract --Deep learning (DL) techniques have demonstrated satisfactory performance in many tasks, even in safety-critical applications. Reliability is hence a critical consideration to DLbased systems. However, the statistical nature of DL makes it quite vulnerable to invalid inputs, i.e., those cases that are not considered in the training phase of a DL model. This paper proposes to perform data sanity check to identify invalid inputs, so as to enhance the reliability of DLbased systems. T o this end, we design and implement a tool to detect behavior deviation of a DL model when processing an input case, and considers it the symptom of invalid input cases. Via a light, automatic instrumentation to the target DL model, this tool extracts the data flow footprints and conducts an assertion-based validation mechanism. The assertions are built automatically, which are specifically-tailored for DL model data flow analysis. Our experiments conducted with real-world scenarios demonstrate that such an assertion-based data sanity check mechanism is effective in identifying invalid input cases. Moreover, SaneDL is lightweight, easy-to-construct, and non-intrusive to the target DL model. I NTRODUCTION In recent years, deep learning (DL) techniques have shown great effectiveness in various aspects. A huge amount of DLbased applications and systems have been proposed in favor of peoples daily life and industrial production [1]-[3], even in safety-critical applications. Image recognition module for auto-driving vehicles [1], for instance, determines what operation should be taken according to the real-time images captured by cameras. In such safety-critical scenarios, any unreliable system misbehavior may cause severe incidents. Reliability is hence of great significance for practical DLbased systems. It is widely-accepted that every software system has its valid input domain [4]-[7]. Inputs staying in such a domain, namely, valid inputs, are expected to receive proper execution results. Unfortunately, in real circumstances, there is no guarantee the inputs are always valid. Anomalous, unexpected inputs may arrive and result in unpredictable misbehavior, which in turn degrades reliability.