Research

The SEEDER Group carries out research in the broad research area of Electronics for Computing Systems. Within this scope of research, the the group is focused on two key research thrusts: (i) Computing Paradigms and (ii) Advanced Memories.

1. Computing Paradigms for Intelligent Electronic Systems

The demand for intelligent electronics that perform their functions based on data gathered locally and/or globally has exploded in recent times. These devices will enable the Cognitive Internet of Things and revolutionize urban transportation (i.e. autonomous vehicles), manufacturing (Industry 4.0), healthcare (i.e. wearable and/or implantable electronics), etc.

The SEEDER Group is very interested in designing intelligent electronic systems that take advantage of emerging device technologies. Our members design hardware architectures and circuits based on emerging device technologies and address design issues at every level of the hardware stack. Research topics along this research thrust include:

• Novel device concepts
  • Artificial synapses
  • Artificial neurons
  • Mapping of neuromorphic algorithms (e.g., spiking neural networks, supervised/unsupervised learning, etc.) to device operation
  • Ising bits
• Hardware accelerator microarchitectures
  • Spiking neural network emulators
  • Ising machines

2. Advanced Memories

Memory is an ubiquitous component of modern electronics. As the requirements for electronic systems evolve, so too do the demands on the memory subsystem.

We are developing concepts for memory subsystems that opens new possibilities for future electronic systems. The SEEDER Group explores memory design at every level of the design hierarchy, from device technology all the way up to the memory architecture. Research topics along this research thrust include:

• Device concepts for ultra-low power, high density non-volatile memories
• Circuits and hardware architectures for functionalized memories (e.g. in-memory compute engines)

Methodologies and Approaches

The development of next-generation electronic systems can no longer follow the tradition paradigm of “making circuits better by improving device behavior” and will require a broad understanding of the design challenges across the abstraction layers in the design hierarchy. The SEEDER group explores the design of electronic systems across various layers of design abstraction of electronics hardware–from material physics all the way up to the hardware architecture. By studying the interactions between various layers of design abstraction and understanding how/where/when/why various physical phenomena works (and does not work), we enable physical design of novel electronic circuits and systems that fully utilize new device technologies. The following is a summary listing (not comprehensive) of the research activities that the SEEDER group is actively involved in.

1. Device Concepts, Simulations and Compact Modeling
   • Spintronic/Ferroelectric/Multiferroic Materials and Devices
   • Field-Effect Transistors (Negative-Capacitance FET, based on 2D materials such as MoS₂, CNT/graphene, etc.)
   • TCAD Modeling of Devices for Circuits/Systems
   • Novel Approaches to High Quality Compact Model Generation
2. Device/Circuit Co-Design
   • Novel Mixed-Signal Circuits enabled by Emerging Device Technologies
   • Non-Volatile Memory (bitcells and memory arrays)
   • Flexible/stretchable Electronics
   • Hardware Security Primitives
   • SRAM Design
3. Hardware Architectures
   • Neuromorphic Computing
   • Bio-inspired Approaches to Boolean and Non-Boolean Computing
   • In-Memory Processing/Compute
   • Embedded NVM Caching Schemes
   • Approximate / Stochastic Computing