Are you interested in working on cutting-edge research projects that combine machine learning and neuromorphic computing? Do you have a strong interest in Bayesian neural networks (BNNs) and theirapplications? If yes, then you might be the perfect candidate for a student research assistant position at our lab!

We are looking for a motivated and talented student who can help us implement BNNs on neuromorphic in-memory computing platforms. Neuromorphic computing is a non-von Neumann computational paradigm that mimics the structure and function of biological brains using neurons and synapses. In-memory computing is a technique that performs certain computational tasks in place in memory, thereby reducing the energy cost of data movement. By combining these two approaches, we aim to achieve high performance and efficiency for BNNs, which are probabilistic versions of traditional neural networks that can capture uncertainty and robustness.

Requirements:
     •   A bachelor’s or master’s degree in computer science, electrical engineering, or related fields
     •   A good understanding of BNNs and conventional neural networks or willingness to learn
     •   Experience with PyTorch or willingness to learn
     •   Motivated to get familiarized with neuromorphic computing concepts and architectures.
     •   Good programming skills in Python

Contact:
     •  Prof. Dr. Mehdi Tahoori, Chair of Dependable Nano Computing
     •  Soyed Tuhin Ahmed, Chair of Dependable Nano Computing

References:
     1. S. T. Ahmed, K. Danouchi, C. Münch, G. Prenat, L. Anghel and M. B. Tahoori, "SpinDrop: Dropout-Based Bayesian Binary Neural Networks With Spintronic Implementation," in IEEE Journal on Emerging and Selected Topics in Circuits and Systems, vol. 13, no. 1, pp. 150-164, March 2023, doi:         10.1109/JETCAS.2023.3242146.
     2. Soyed Tuhin Ahmed, Kamal Danouchi, Christopher Münch, Guillaume Prenat, Lorena Anghel, and Mehdi B. Tahoori. 2022. Binary Bayesian Neural Networks for Efficient Uncertainty Estimation Leveraging Inherent Stochasticity of Spintronic Devices. In IEEE/ACM NANOARCH
     3. Gawlikowski, Jakob, et al. "A survey of uncertainty in deep neural networks." arXiv preprint arXiv:2107.03342 (2021).
     4. Mobiny, Aryan, et al. "Dropconnect is effective in modeling uncertainty of bayesian deep networks." Scientific reports 11.1 (2021): 1-14.