Fabrizio Ottati

The C-Transformer architecture dynamically routes language model workloads between ANN and SNN cores based on spike sparsity, reducing LLM energy by up to 72%.

IBM's NorthPole chip intertwines memory and compute, achieving 25x higher energy and 5x higher space efficiency than conventional architectures.

A deep dive into IBM's NorthPole, a brain-inspired AI accelerator. Understand its architecture, 10 core axioms, and how it achieves groundbreaking energy efficiency for neural inference.

Learn how neuromorphic mechanisms like SoftHebb learning and short-term synaptic plasticity can surpass conventional AI in speed and adversarial robustness.

SoftHebb learning and short-term plasticity mechanisms improve state-of-the-art AI performance on dynamic tasks without relying on non-local backpropagation.

Innatera's Talamo SDK streamlines deploying spiking neural network architectures to mixed-signal neuromorphic hardware for low-power edge audio processing.

A bio-inspired visual attention model leverages event cameras and SpiNNaker neuromorphic hardware to give robotic agents low-latency, depth-aware object focus.

Dynamically scheduled dataflow circuits overcome static high-level synthesis limits to enable out-of-order memory and speculative execution from C/C++ code.

Discover how the brain could implement gradient descent, addressing weight transport and multiplexing through feedback alignment and node perturbation.

Examining the biological plausibility of gradient descent reveals how node perturbation and feedback alignment successfully bypass the weight transport problem.

The Microbrain architecture combines asynchronous processing and Forward Propagation Through Time (FPTT) to train 6.2-million-neuron SNNs.

The Exodus backend accelerates BPTT training of PyTorch-based spiking neural networks in Sinabs before deploying them to the event-driven Speck hardware chip.

See a complete workflow for training spiking neural networks in PyTorch with Sinabs and deploying them directly to the event-driven Speck neuromorphic chip.

Explore the innovative TrueNorth neuromorphic chip, its event-driven architecture, low power operation, massive parallelism, real-time capabilities, and scalable design.

The EONS framework applies evolutionary algorithms to co-design spiking neural network topologies and parameters for diverse neuromorphic hardware.

Learn how to overcome the discrepancies between synchronous GPU training and asynchronous deployment when building SNNs for SynSense's Speck chip.

snnTorch applies surrogate gradients to overcome the non-differentiability of spikes, enabling SNN training and quantization using standard PyTorch pipelines.

Discover methods to efficiently encode and store event-based data from high-resolution event cameras, striking a balance between file size and fast retrieval for spiking neural network training.

Stay up-to-date with cutting-edge digital hardware designs for neuromorphic applications. Explore recent research on power-efficient event-driven spiking neural networks and state-of-the-art processors like TrueNorth and Loihi.

Learn how to model Leaky Integrate and Fire (LIF) neurons in digital hardware. Understand spike communication, synapse integration, and more for hardware implementation.