Low-Power Deep Learning Inference using the SpiNNaker Neuromorphic Platform

Vineyard, Craig Michael; Dellana, Ryan; Aimone, James B.; Rothganger, Fredrick; Severa, William

doi:10.1145/3320288.3320300

Cited by 6 publications

(2 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In [47], Whetstone is deployed on SpiNNaker [48], with slight drop in accuracy due to issues with input/output encoding. Here, we optimize the network using Whetstone, but we do not map the resulting networks to a neuromorphic hardware implementation, such as SpiNNaker [48] or Loihi [49].…”

Section: Discussionmentioning

confidence: 99%

“…Here, we optimize the network using Whetstone, but we do not map the resulting networks to a neuromorphic hardware implementation, such as SpiNNaker [48] or Loihi [49]. As observed in [47], several other hyperparameters such as input/output encoding, different network topologies and training parameters will have an effect on this mapping performance. In the future, we plan to include how the network performs on real neuromorphic hardware as part of our training objectives in the hyperparameter and network architecture optimization process.…”

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

Hyperparameter Optimization in Binary Communication Networks for Neuromorphic Deployment

Parsa

Schuman

Date

et al. 2020

2020 International Joint Conference on Neural Networks (IJCNN)

Self Cite

View full text Add to dashboard Cite

Training neural networks for neuromorphic deployment is non-trivial. There have been a variety of approaches proposed to adapt back-propagation or back-propagation-like algorithms appropriate for training. Considering that these networks often have very different performance characteristics than traditional neural networks, it is often unclear how to set either the network topology or the hyperparameters to achieve optimal performance. In this work, we introduce a Bayesian approach for optimizing the hyperparameters of an algorithm for training binary communication networks that can be deployed to neuromorphic hardware. We show that by optimizing the hyperparameters on this algorithm for each dataset, we can achieve improvements in accuracy over the previous state-of-theart for this algorithm on each dataset (by up to 15 percent). This jump in performance continues to emphasize the potential when converting traditional neural networks to binary communication applicable to neuromorphic hardware.

show abstract