M Ihsan Al Hafiz

Lindstedtsvägen 30

Stockholm, Sweden

Due to the rise of Artificial Intelligence, energy consumption is expected to rise. A brain as a source of inspiration actually requires ~20 W to work. Is it because of the current AI algorithm? Or a computation machine that is not compatible? Or both?

My research is to explore the neuromorphic computing, or how we do computation more biologically close to the brain, in a custom computation machine that is more compatible with it. Currently, I am working with two biologically plausible algorithms, Bayesian Confidence Propagation Neural Networks (BCPNNs) and Spiking Neural Networks (SNNs). BCPNN draws inspiration from the columnar structure of the brain and uses Bayesian statistics for its learning mechanism. While SNN represents the computation more like the brain, as spikes signal. Due to the specific algorithm of neuromorphic computing, general-purpose machines like the Central Processing Unit (CPU) or the Graphic Processing Unit (GPU) will not be able to optimize low-energy implementation. Therefore, a specialized hardware accelerator is needed as an alternative to a general-purpose computing machine for neuromorphic computing. The target is to replace the current AI workload with more efficient, low-energy-consuming neuromorphic computing to support a more sustainable future.

My research interests include hardware accelerators, neuromorphic computing, High-Performance Computing, FPGAs, and reconfigurable computing. More information about me can be found at ihsanalhafiz.github.io.

news

Apr 30, 2026	Just got acceptance paper in Euro-Par conference 2026 for the paper NeuroRing: Scaling Spiking Neural Networks via Multi-FPGA Bidirectional Ring Topologies and Stream-Dataflow Architectures
Apr 01, 2024	Started my PhD in Hardware Accelerator for Neuromorphic Computing at KTH Royal Institute of Technology under supervision of Artur Podobas.

latest posts

May 07, 2026	play with always_comb and always_ff in systemverilog
Dec 08, 2023	Precision in PCB Layout: The Ultimate Checklist for Flawless PCB Routing Review

selected publications

ARC 2025

A Reconfigurable Stream-Based FPGA Accelerator for Bayesian Confidence Propagation Neural Networks

Muhammad Ihsan Al Hafiz, Naresh Ravichandran, Anders Lansner, and 2 more authors

In Applied Reconfigurable Computing. Architectures, Tools, and Applications, 2025

Abs

Brain-like algorithms are attractive and emerging alternatives to classical deep learning methods for use in various machine learning applications. Brain-like systems can feature local learning rules, both unsupervised/semi-supervised learning and different types of plasticity (structural/synaptic), allowing them to potentially be faster and more energy-efficient than traditional machine learning alternatives. Among the more salient brain-like algorithms are Bayesian Confidence Propagation Neural Networks (BCPNNs). BCPNN is an important tool for both machine learning and computational neuroscience research, and recent work shows that BCPNN can reach state-of-the-art performance in tasks such as learning and memory recall compared to other models. Unfortunately, BCPNN is primarily executed on slow general-purpose processors (CPUs) or power-hungry graphics processing units (GPUs), reducing the applicability of using BCPNN in Edge systems, among others. In this work, we design a reconfigurable stream-based accelerator for BCPNN using Field-Programmable Gate Arrays (FPGA) using Xilinx Vitis High-Level Synthesis (HLS) flow. Furthermore, we model our accelerator’s performance using first principles, and we empirically show that our proposed accelerator (full-featured kernel non-structural plasticity) is between 1.3x - 5.3x faster than an Nvidia A100 GPU while at the same time consuming between 2.62x - 3.19x less power and 5.8x - 16.5x less energy without any degradation in performance.
MCSoC 2025

Embedded FPGA Acceleration of Brain-Like Neural Networks: Online Learning to Scalable Inference

Muhammad Ihsan Al Hafiz, Naresh Ravichandran, Anders Lansner, and 2 more authors

In 2025 IEEE 18th International Symposium on Embedded Multicore/Many-core Systems-on-Chip (MCSoC), Dec 2025

ISSN: 2771-3075

Abs DOI

Edge AI increasingly requires models that learn and adapt on-device under a tight energy budget. Mainstream deep learning models, while powerful, are often overparameterized, energy-hungry and dependent on cloud connectivity. Brain-Like Neural Networks (BLNNs), such as the Bayesian Confidence Propagation Neural Network (BCPNN), propose a neuromorphic alternative by mimicking cortical architecture and biologicallyconstrained learning. They offer sparse architectures with local learning rules and unsupervised/semi-supervised learning, making them well-suited for low-power edge intelligence. However, existing BCPNN implementations rely on GPUs or datacenter FPGAs. This work presents the first embedded FPGA accelerator for BCPNN on a Zynq UltraScale+ SoC (ZCU104) using High-Level Synthesis. We implement both online learning and inference-only kernels with configurable precision (FP32, FP16, and mixed FP16/FXP16). Evaluated on MNIST, Pneumonia, and Breast Cancer datasets, our accelerator delivers up to 17.55% lower latency and 94.1% energy savings over ARM baselines. Our work brings practical, brain-like online learning and scalable inference to edge devices.