The development of Spin-Transfer Torque Magnetic RAMs (STT-MRAMs) mass production requires high-quality test solutions. Accurate and appropriate fault modeling is crucial for the realization of such solutions. This paper targets fault modeling and test generation for all interconnect and contact defects in STT-MRAMs and shows that using the defect injection and circuit simulation for fault modeling without incorporating the impact of magnetic coupling will result in an incomplete set of fault models; hence, not obtaining accurate fault models. Magnetic coupling introduced by the stray field is an inherent property of STT-MRAMs and may foster the occurrence of additional memory faults. Not considering the magnetic coupling clearly will give rise to test escapes. The paper introduces a compact model for STT-MRAM that incorporates the intra- and inter-cell stray field, uses this model to derive the full set of fault models for interconnect and contact defects, and finally proposes an efficient test solution.

The graduation project is about personalizing the multimedia guide (abbreviated as MMT) of the Van Gogh Museum. The personalization goal is to facilitate future visitors to engage with the collection and maximize the meaning of visiting the museum based on visitors’ diverse concerns. A model is brought up during the project and can describe four popular concerns of the adult visitors at this museum, which are about “understanding Van Gogh”, “learning art techniques”, “enjoying an aesthetic experience”, and “finding out the landmarks”. The concern model contains two parts for each concern: the first part collects the typical minds-on behaviors and terms for each category, and can be used to identify each visitor’s concerns; while the second part contains the facilitation design principles and examples for each category, and can be used to inspire the new content designs for each concern type. It’s also found out that, the core of the personalized product/service system is to collect enough data from the visitors, identify their concern(s) with the help of a robust backend algorithm, and then present selected content to the visitors. A MMT can be called as a personalized MMT as long as it can collect data from visitors, and select one content out of many according to their data input, but with a premise that the content should be relevant to the identified concern. After several design iterations, 17 design principles were generated for guiding the data input and content output design. An embodiment design was made as a personalized MMT example. The final design requires visitors’ interested terms in their daily life and collects them through a concern quiz which will be presented to the visitors at the beginning of their visit; their input will be calculated by the backend Bayesian network, and then ten personal highlights and some personal activities will be selected for the visitors immediately; visitors can see their personal highlights on a digital map and use it for their actual visit; they will also receive invitations of some personal activities from their conversational MMT assistant during the break. All the interaction happens only on the MMT screen. The embodiment design was tested by eight participants from TUDelft who role-played four different charterers in the Van Gogh Museum and visited the permanent exhibition with a Wizard-of-Oz prototype. The evaluation result is positive, which means the final concept can achieve the personalization goal in the desired way.