GitPlanet

Cheap and reliable Node.js hosting starts at $3/month, and $1/month static HTML hosting

Created with love in Canada, visit hostnodejs.com today

Feel like to post an Ad? Learn Details
All Projects → jostmey → Rwa

jostmey / Rwa

Licence: bsd-3-clause
Machine Learning on Sequential Data Using a Recurrent Weighted Average

Programming Languages

python
139335 projects - #7 most used programming language

Labels

time-seriesresearchrecurrent-neural-networks

Projects that are alternatives of or similar to Rwa

Time Series Machine Learning
Machine learning models for time series analysis
Stars: ✭ 261 (-55.99%)
Mutual labels:  time-series, recurrent-neural-networks
Tensorflow Cnn Time Series
Feeding images of time series to Conv Nets! (Tensorflow + Keras)
Stars: ✭ 49 (-91.74%)
Mutual labels:  time-series, recurrent-neural-networks
Deep Learning Time Series
List of papers, code and experiments using deep learning for time series forecasting
Stars: ✭ 796 (+34.23%)
Mutual labels:  time-series, recurrent-neural-networks
AC-VRNN
PyTorch code for CVIU paper "AC-VRNN: Attentive Conditional-VRNN for Multi-Future Trajectory Prediction"
Stars: ✭ 21 (-96.46%)
Mutual labels:  time-series, recurrent-neural-networks
LSTM-Time-Series-Analysis
Using LSTM network for time series forecasting
Stars: ✭ 41 (-93.09%)
Mutual labels:  time-series, recurrent-neural-networks
unicornn
Official code for UnICORNN (ICML 2021)
Stars: ✭ 21 (-96.46%)
Mutual labels:  time-series, recurrent-neural-networks
Lstm anomaly thesis
Anomaly detection for temporal data using LSTMs
Stars: ✭ 178 (-69.98%)
Mutual labels:  time-series, recurrent-neural-networks
dts
A Keras library for multi-step time-series forecasting.
Stars: ✭ 130 (-78.08%)
Mutual labels:  time-series, recurrent-neural-networks
Tensorflow Lstm Regression
Sequence prediction using recurrent neural networks(LSTM) with TensorFlow
Stars: ✭ 433 (-26.98%)
Mutual labels:  time-series, recurrent-neural-networks
Multilabel Timeseries Classification With Lstm
Tensorflow implementation of paper: Learning to Diagnose with LSTM Recurrent Neural Networks.
Stars: ✭ 519 (-12.48%)
Mutual labels:  time-series
Promxy
An aggregating proxy to enable HA prometheus
Stars: ✭ 562 (-5.23%)
Mutual labels:  time-series
Holodeck
High Fidelity Simulator for Reinforcement Learning and Robotics Research.
Stars: ✭ 513 (-13.49%)
Mutual labels:  research
Karta
Karta - source code assisted fast binary matching plugin for IDA
Stars: ✭ 519 (-12.48%)
Mutual labels:  research
Ios
Most usable tools for iOS penetration testing
Stars: ✭ 563 (-5.06%)
Mutual labels:  research
Qlib
Qlib is an AI-oriented quantitative investment platform, which aims to realize the potential, empower the research, and create the value of AI technologies in quantitative investment. With Qlib, you can easily try your ideas to create better Quant investment strategies. An increasing number of SOTA Quant research works/papers are released in Qlib.
Stars: ✭ 7,582 (+1178.58%)
Mutual labels:  research
Warez
All your base are belong to us!
Stars: ✭ 584 (-1.52%)
Mutual labels:  research
Awesome Time Series Database
🕖 A curated list of awesome time series databases, benchmarks and papers
Stars: ✭ 501 (-15.51%)
Mutual labels:  time-series
Pgm Index
🏅State-of-the-art learned data structure that enables fast lookup, predecessor, range searches and updates in arrays of billions of items using orders of magnitude less space than traditional indexes
Stars: ✭ 499 (-15.85%)
Mutual labels:  research
Dnc Tensorflow
A TensorFlow implementation of DeepMind's Differential Neural Computers (DNC)
Stars: ✭ 587 (-1.01%)
Mutual labels:  research
Quantmod
Quantitative Financial Modelling Framework
Stars: ✭ 578 (-2.53%)
Mutual labels:  time-series
View All Similar Projects ➔

Description

This repository holds the code to a new kind of RNN model for processing sequential data. The model computes a recurrent weighted average (RWA) over every previous processing step. With this approach, the model can form direct connections anywhere along a sequence. This stands in contrast to traditional RNN architectures that only use the previous processing step. A detailed description of the RWA model has been published in a manuscript at https://arxiv.org/pdf/1703.01253.pdf.

alt text

Because the RWA can be computed as a running average, it does not need to be completely recomputed with each processing step. The numerator and denominator can be saved from the previous step. Consequently, the model scales like that of other RNN models such as the LSTM model.

In each folder, the RWA model is evaluated on a different task. The performance of the RWA model is compared against a LSTM model. The RWA is found to train considerably faster on most tasks by at least a factor of five. As the sequences become longer, the RWA model scales even better. See the manuscript listed above for the details about each result.

Note: The RWA model has failed to yield competitive results on Natural Language Problems.

Download

  • Download: zip
  • Git: git clone https://github.com/jostmey/rwa

Requirements

The code is written in Python3. The scripts have been upgraded to run using version 1.0 of TensorFlow.

Alternative Implementations

Acknowledgements

Thanks Alex Nichol for correcting the equations for numerical stability.

Corrections (Changelog)

  • March 17th, 2017: Corrected equations used to rescale the numerator and denominator terms, which is used to avoid overflow and underflow conditions. Results for the RWA model were recomputed.
  • March 26th, 2017: Corrected a bug specific to the code for loading the permuted MNIST task. Results for permuted MNIST task were recomputed.
  • April 3rd, 2017: Corrected bug in the LSTM model. This bug affected all the results except for the copy problem. Results for the LSTM model were recomputed. No significant changes in performance were observed.
Note that the project description data, including the texts, logos, images, and/or trademarks, for each open source project belongs to its rightful owner. If you wish to add or remove any projects, please contact us at [email protected].