Skip to content Zenke Lab Computational Neuroscience at the FMI Selected talks from the lab Research Funding Spiking Heidelberg Digits and Spiking Speech Commands Auryn Spiking Network Simulator LaTeX rebuttal/response to reviewers template Great free text books Preprint: The remarkable robustness of surrogate gradient learning for instilling complex function in spiking neural networks June 30, 2020 fzenke We just put up a new preprint https://www.biorxiv.org/content/10.1101/2020.06.29.176925v1 in which we
Understand Generative Adversarial Networks (GANs) and how banks use them to generate synthetic data for AI training without compromising privacy
Spiking neural networks (SNNs) are believed to be highly computationally and energy efficient5 for specific neurochip hardware real-time solutions. However
Could psychedelic drugs one day play a part in the treatment of mental health conditions? The idea is getting less and less far-fetched, after scientists successfully used drugs including MDMA and LSD to repair neurons in animal tests and cultured cells.
I tidied up the code for SwitchNet neural network and pointed out a linear mapping mathematical justification. https://sites.google.com/view/algorithmshortcuts/switchnet-newer I’m not a paid computer scientist. I don’t
I have rarely been as enthusiastic about a new research direction. We call them GFlowNets, for Generative Flow Networks
Mysterianism is the idea that human cognition, mind, cannot be understood. Taking this concept and applying it to known mechanism — such that claims are made that we do not know how engineered systems, such as artificial neural networks (ANNs), work, or that they constitute black boxes that we can only open with difficulty — is inappropriate at best and malicious at worst. We do know the mechanistic structure of such models because we designed and built them. We also do know their functional role (what
Hi, I’m working on the train_step of week 4 assignment 2. I’m in the for loop, directionally i’m: getting the vgg_model_outputs for the current image computing the style cost using the style layers and current image …
## Utilizing Visual Forms of Japanese Characters for Neural Review Classification Yota Toyama , Makoto Miwa , Yutaka Sasaki We propose a novel method that exploits visual information of ideograms and logograms in analyzing Japanese review documents. Our method first converts font images of Japanese characters into character embeddings using convolutional neural networks. It then constructs document embeddings from the character embeddings based on Hierarchical Attention Networks, which represent the docum
A simple regularization technique called cutout, involving random masking of input regions, significantly improves the performance and robustness of convolutional neural