Prizes 2025
27 January, 2025
2025 Misha Mahowald Prizes
The 2025 Misha Mahowald Prize is awarded to the OPUS team from UC Santa Barbara led by Kerem Y. Çamsari for their work on Stochastic Neuromorphic Computing with Probabilistic Bits.
The 2025 Mahowald Early Career Award is awarded equally to:
Mark M. Iskarous for the work on Invariant neuromorphic representations of tactile stimuli improves robustness of a real-time texture classification
and
Jens Egholm Pedersen for the work on Neuromorphic Intermediate Representation
The 2025 Misha Mahowald Prize
The MMP Jury awarded the prize to the team of Navid A. Aadit, Shuvro Chowdhury, Shaila Niazi, Kemal Selçuk, Nihal S. Singh, and Kerem Y. Çamsarı at UC Santa Barbara for their work on Stochastic Neuromorphic Computing with Probabilistic Bits
Jury Citation: The Orchestrating Physics for Unconventional Systems (OPUS) Lab implemented an architecture with special-purpose digital hardware based on three neuromorphic principles – massive parallelism, asynchronous dynamics, and sparsity – and demonstrated applications to problems in combinatorial optimization, energy-based machine learning, and quantum simulations.
The 2025 Mahowald Early Career Awards
The MECA Jury equally awarded the MECA to two projects:
Mark M. Iskarous at Neuroengineering and Biomedical Instrumentation Lab, Johns Hopkins University for the work on Invariant neuromorphic representations of tactile stimuli improves robustness of a real-time texture classification
Jury Citation: This application impressed the jury because it combined neural modelling with hardware and experimental testing in a sensory domain that so far has been less explored in neuromorphic engineering. The work creates a force and speed invariant representation of texture that improves accuracy and computational efficiency of texture classification, which is tested in real world experiments.
Jens Egholm Pedersen at Neurocomputing Systems Lab, KTH Royal Institute of Technology for the work on Neuromorphic Intermediate Representation
Jury Citation: The Neuromorphic Intermediate Representation is an important effort to lower the barrier of entry to using neuromorphic computing hardware by providing a standardised set of computational primitives. This effort will allow many more researchers to use and try out neuromorphic hardware, which in turn we expect to lead to advances in neuromorphic computation and neuromorphic hardware.