Profile
Professor in electronics engineering at Digital University Kerala, and as the Chief Scientist and CTO at Graphene Aurora, He lead the research and development of cutting-edge graphene-based devices and neural chips for various applications, such as biomedical, IoT, and AI. With over 20+ years of experience in the field of electrical and electronics engineering, He had developed multiple patents, publications, and products that have advanced the state-of-the-art in brain-inspired computing.
Areas of current work: AI chips (Memristors, Crossbars, Neuromorphic circuits, ANN, SNN, CNN, LSTM, HTM, LLMs), RF circuits, antenna designs (UWB, Phased Array Antennas, Parabolic Antennas, Horn Antennas, Log-Periodic Antennas, Patch (Microstrip) Antennas, Yagi-Uda Antennas, Spiral Antennas, Helical Antennas, and Monopole Antenna), various 2D materials (Graphene), Memory circuits (ReRAMs, STTRAM, PCM), drones, image processing, AI/ML, learning circuits, security circuits, impedance matching, crosstalk, and signal reflections, jitter and timing analysis, EMI and EMC
Topic
Crossbar based Mixed-Signal Neural Architectures under Variability and Parasitics
Abstract
In this lecture, the variability impact, compensation techniques, and variability-aware neural architectures are discussed. Variability and paracitics pose significant challenges in ensuring accurate multiply and accumulate (MAC) computations emulated with a crossbar. In most neural networks, the MAC forms the core computing module to implement a neuron and, consequently, neural networks. The analog MAC blocks are particularly susceptible to variability and parasitics, which can result in inaccuracies in the computation process. Therefore, developing techniques to mitigate these challenges is crucial for the successful implementation of variability averse mixed-signal neural architectures.