Tania Roy
Duke University, NC·Engineering·Early-career
Today's AI models rely on huge, power-hungry data centers, in part because ordinary computers store and process data in separate places. Brain-inspired, or neuromorphic, hardware instead builds artificial neurons and synapses that compute where data are stored or sensed.
- Fields
Engineering
Materials Science
Computer Science
- Topics
neuromorphic computing
artificial synapses
artificial neurons
- Grants
- 2 known grants, through Jul 2027
What they work on
- Develop in-pixel computing circuits using optoelectronic synaptic devices made from 2D materials, sensing several wavelengths, that maximize each pixel's fill factor and form the first layer of a convolutional neural network.
- Integrate in-sensor pixel arrays with crossbar arrays of memristive synapses and nanomaterial peripheral circuits to classify and segment images, using a hybrid analog/digital scheme to offset device imperfections.
- Build the full in-sensor computing system from recyclable materials and study the recyclability of every circuit to reduce electronic waste.
- Use volatile resistive switching in vertically standing MoS2 sheets between graphene electrodes to make integrate-and-fire artificial neurons, and study the stochastic nature of their firing.
- Develop low-power artificial synapses from multi-level, non-volatile switching in horizontal MoS2 sheets, investigate why vertical and horizontal sheets switch differently, and integrate neurons and synapses on one chip, including on flexible substrates.
More about the research
The grants pursue this idea using two-dimensional (2D) materials, crystals only a few atoms thick. One project builds energy-efficient neurons and synapses from stacked 2D layers, and the other builds image sensors that process pictures inside the sensor itself, made from recyclable materials.
Why it matters
Energy-efficient AI hardware could let devices like laptops run image-processing models without cloud servers, cutting power use and the climate impact of AI. Applications named include self-driving cars, industrial monitoring, athlete motion analysis, health-monitoring wearables for elderly people living alone, and lightweight circuits for spacecraft. Recyclable electronic materials could also reduce electronic waste.
Known grants
From NIH and NSF research grants. Labs often have other funding we can't see.
Closest labs
Davood Shahrjerdi
New York University, NY·Materials Science
Determine how polymer residues such as polyvinyl alcohol and paraffin interact with 2D materials like graphene and molybdenum disulfide, viewing the problem in terms of interface energy.
Sumit Jha
University of Florida, FL·Computer Science
Develop formal verification methods that prove neural networks run correctly when accelerated on analog in-memory computing hardware, which can be more error-prone than standard CMOS chips.
Rashmi Jha
University of Cincinnati, OH·Engineering
Model, design and synthesize specific synthetic DNA sequences, varying base pairs, sequence, length and crystallinity, for integration with organometallic halide perovskite semiconductors.
The school
Duke University: 61 labs on PI Pages, $1.6B a year on research, 417 research doctorates a year.
Reach out
A start for your email. Fill in the brackets.