Neuromorphic computing at scale
Neuromorphic computing is a brain-inspired approach to hardware and algorithm design
that efficiently realizes artificial neural networks. Neuromorphic designers apply the …
that efficiently realizes artificial neural networks. Neuromorphic designers apply the …
SpiNNaker2: A large-scale neuromorphic system for event-based and asynchronous machine learning
HA Gonzalez, J Huang, F Kelber, KK Nazeer… - ar** enhances exact gradient learning with eventprop in spiking neural networks
Event-based machine learning promises more energy-efficient AI on future neuromorphic
hardware. Here, we investigate how the recently discovered Eventprop algorithm for …
hardware. Here, we investigate how the recently discovered Eventprop algorithm for …
Language Modeling on a SpiNNaker2 Neuromorphic Chip
As large language models continue to scale in size rapidly, so too does the computational
power required to run them. Event-based networks on neuromorphic devices offer a …
power required to run them. Event-based networks on neuromorphic devices offer a …
Advancing State of the Art in Language Modeling
Generalization is arguably the most important goal of statistical language modeling
research. Publicly available benchmarks and papers published with an open-source code …
research. Publicly available benchmarks and papers published with an open-source code …
Beyond weights: deep learning in spiking neural networks with pure synaptic-delay training
E Grappolini, A Subramoney - … of the 2023 International Conference on …, 2023 - dl.acm.org
Biological evidence suggests that adaptation of synaptic delays on short to medium
timescales plays an important role in learning in the brain. Inspired by biology, we explore …
timescales plays an important role in learning in the brain. Inspired by biology, we explore …
Activity sparsity complements weight sparsity for efficient RNN inference
Artificial neural networks open up unprecedented machine learning capabilities at the cost
of ever growing computational requirements. Sparsifying the parameters, often achieved …
of ever growing computational requirements. Sparsifying the parameters, often achieved …
Spiking Music: Audio Compression with Event Based Auto-encoders
M Lisboa, G Bellec - arxiv preprint arxiv:2402.01571, 2024 - arxiv.org
Neurons in the brain communicate information via punctual events called spikes. The timing
of spikes is thought to carry rich information, but it is not clear how to leverage this in digital …
of spikes is thought to carry rich information, but it is not clear how to leverage this in digital …
Residual Echo State Networks: Residual recurrent neural networks with stable dynamics and fast learning
A Ceni, C Gallicchio - Neurocomputing, 2024 - Elsevier
Residual connections have been established as a staple for modern deep learning
architectures. Most of their applications are cast towards feedforward computing. In this …
architectures. Most of their applications are cast towards feedforward computing. In this …
Harnessing Manycore Processors with Distributed Memory for Accelerated Training of Sparse and Recurrent Models
J Finkbeiner, T Gmeinder, M Pupilli, A Titterton… - Proceedings of the …, 2024 - ojs.aaai.org
Current AI training infrastructure is dominated by single instruction multiple data (SIMD) and
systolic array architectures, such as Graphics Processing Units (GPUs) and Tensor …
systolic array architectures, such as Graphics Processing Units (GPUs) and Tensor …