With the advent of intelligent Internet-of-Things (IoT), processing Deep-Learning (DL) algorithms in resource-constrained edge devices is challenging. To tackle this challenge, we must rely on novel DL inference accelerators that consume little power/energy, i.e., within the limited energy budgets of edge devices, while providing high-quality and low-latency inference results, thus realizing ubiquitous intelligence.

In this context, this project investigates approximate computing as a promising path towards increased energy efficiency. We aim to design novel approximate low-power architectures that will enable DL-based inference in edge devices without compromising accuracy or dependability.

Towards this goal, several challenges must be addressed. First, an important challenge in approximate computing is to analyse and model the effects of approximation on the target applications. To address this challenge, we aim to design novel analysis methods and develop the tools that can evaluate and justify the proposed approximations within the target application. Second, we envision using such tools in combination with approximate hardware models and statistical techniques, thus exploring the design space of possible solutions. Finally, through design-space exploration targeting efficiency and performance metrics, we expect to determine, design, implement, and validate the promising quality-efficiency designs in FPGA/ASIC, thus assessing the energy efficiency gains.

Overall, we expect the project will make conceptual contributions to the approximate computing, low-power architectures, and edge AI accelerators fields, while also providing methods and tools for application and system architects to evaluate and realize these designs.

Are you interested in these topics? Please apply! We will welcome you to our CAES group, where we ensure an inclusive and collaborative learning atmosphere. You will be encouraged to co-supervise research assignments and labs (MSc./BSc.), and collaborate with other groups within and outside of University of Twente.

• You have a MSc degree in Computer Science, Software/Computer/Electrical engineering or associated field.
• You have sufficient knowledge of computer architecture and digital hardware design.
• You have experience and/or strong interest in algorithms, design space exploration, and/or statistical (error resilience) analysis
• You are interested to contribute to education tasks – such as TA or BSc/MSc thesis co-supervisor.
• You are passionate about non-conventional computing techniques and open to take up related courses/training.
• You are an enthusiastic and highly motivated researcher.
• You have a creative mindset and very good analytical and communication skills.
• You are a team-player, and want to work in an interdisciplinary, internationally oriented, and collaborative environment.
• You are proficient in English.

• As a PhD candidate at UT, you will be appointed to a full-time position for four years, with a qualifier in the first year, within a very stimulating and exciting scientific environment;
• The University offers a dynamic ecosystem with enthusiastic colleagues;
• Your salary and associated conditions are in accordance with the collective labour agreement for Dutch universities (CAO-NU);
• You will receive a gross monthly salary ranging from € 2.770,- (first year) to € 3.539,- (fourth year);
• There are excellent benefits including a holiday allowance of 8% of the gross annual salary, an end-of-year bonus of 8.3%, and a solid pension scheme;
• The flexibility to work (1 day/week) from home;
• A minimum of 232 leave hours in case of full-time employment based on a formal workweek of 38 hours. A full-time employment in practice means 40 hours a week, therefore resulting in 96 extra leave hours on an annual basis.
• Free access to sports facilities on campus
• A family-friendly institution that offers parental leave (both paid and unpaid);
• You will have a training programme as part of the Twente Graduate School where you and your supervisors will determine a plan for a suitable education and supervision;
• We encourage a high degree of responsibility and independence, while collaborating with close colleagues, researchers and other staff.

The PhD student will join CAES, a group working on the most efficient and effective computer architectures of the future, from large-scale data-center servers to low-power/small-scale embedded systems. The group's research sits at the border between EE and CS, aiming to bridge any gaps between these two sides of computing systems.