[D] Is it possible for us to make fixed-size multilayer perceptrons (MLP's) provably converge?

• tchlux

  5 - -

Have you actually tested this before or are you just considering this problem theoretically? We can use some very simple code to show that this is an impractical statement. Here is a simple ReLU MLP trained with gradient descent implemented in numpy, and I've used this to generate the results I'm talking about here. Here's a GIF showing that it is capable of correctly minimizing the error of an approximation.

• VDCNN

  1 171 10.0 Python

  Implementation of Very Deep Convolutional Neural Network for Text Classification

Code for https://arxiv.org/abs/1502.01852 found: https://github.com/zonetrooper32/VDCNN

• InfluxDB

  www.influxdata.com featured

  Power Real-Time Data Analytics at Scale. Get real-time insights from all types of time series data with InfluxDB. Ingest, query, and analyze billions of data points in real-time with unbounded cardinality.

  

• hnmt

  2 1 10.0 Python

  Helsinki Neural Machine Translation system

Code for https://arxiv.org/abs/1607.06450 found: https://github.com/Waino/hnmt

Suggest a related project