Quality Management for AIoT
Quality Management (QM) is responsible for overseeing all activities and tasks needed to maintain a desired level of quality. QM in Software Development traditionally has four main components: quality planning, quality assurance and quality control. In many agile organizations, QM is becoming closely integrated with the DevOps organization. Quality Assurance (QA) is responsible for setting up the oganization and its processes to ensure the desired level of quality. In an agile organization, this means that QA needs to be closely aligned with DevOps. Quality Control (QC) is responsible for the output, usually by implementing a test strategy along the various stages of the DevOps cycle. Quality Planning is responsible for setting up the quality and test plans. In a DevOps organization, this will be a continuous process. For safety and security relevant systems, Verfication & Validation (V&V) usually plays an important role as well. V&V is closely related to Quality Control; it aims to link back test results to requirements.
QM for AIoT-enabled systems must take into consideration all the specific challenges of AIoT-development, including QM for combined hardware / software development, QM for highly distributed systems (including edge components in the field), as well as any homologation requirements of the specific industry. In case of critical systems, Independent Verfication & Validation (IV&V) via an independent 3rd party can be required.
Quality Assurance and AIoT DevOps
Verification & Validation
Verification and Validation for AIoT
V&V for AI
From the perspective of the integrated system, verification and validation of the AI-related services is usually focusing on functional testing, considering the AI-based services a black box ("Black Box Testing"), which is tested in the context of the other services which make up the complete AIoT system. However, it will usually be very difficult to ensure a high level of quality if this the only test approach. Consequently, V&V for the AI services in an AIoT system also requires a "white box" approach, which specifically focuses on the AI-based functionality.
In a recent paper published by the European Union Aviation Safety Agency (EASA), the authors propose a W-shaped model for the "Design Assurance for Neural Networks". This model (see below) assigns data management, learning process management and model training to the left side of the W. The middle part is the learning process verification. On the right side, model implementation, inference model verfication and data verification can be found.
The W-model proposed by EASA could be an interesting approach for addressing some of the common challenges found in verification and validation of AI models, including poor matching quality, data bias, over- and underfitting, model decay and adverserial attacks.
In AI, important techniques for Training Data Validation include the validation of data quality, completeness and representativeness. Techniques for Model Testing & Optimization include K-fold Cross-Validation, Holdout Method, Hyper Parameter Tuning, Statistical Validation Methods and AI Model Simulations. Field tests and production tests can also provide critical feedback to the V&V team of the AI services.
Integrated V&V for AIoT
On the service-level, AI services can usually be tested using the methods outlined in the previous section. After the initial tests performed by the AI service team, it is important that the AI services will be integrated into the overall AIoT product for real-world integration tests. This means that the AI services are integrated with the remaining IoT services, to build the full AIoT system. This is shown in the figure below. The fully integrated system can then be used for User Acceptance Tests, load and scalability tests, and so on.
Authors and Contributors
|