edouardelasalles / Stnn

Licence: bsd-2-clause
Code for the paper "Spatio-Temporal Neural Networks for Space-Time Series Modeling and Relations Discovery"

Projects that are alternatives of or similar to Stnn

End To End Sequence Labeling Via Bi Directional Lstm Cnns Crf Tutorial
Tutorial for End-to-end Sequence Labeling via Bi-directional LSTM-CNNs-CRF
Stars: ✭ 87 (-3.33%)
Mutual labels:  jupyter-notebook
Blog Resources
This repo will contain the resources available in my blog for learning
Stars: ✭ 89 (-1.11%)
Mutual labels:  jupyter-notebook
Fast tabnet
TabNet for fastai
Stars: ✭ 90 (+0%)
Mutual labels:  jupyter-notebook
Speech Emotion Recognition
Detecting emotions using MFCC features of human speech using Deep Learning
Stars: ✭ 89 (-1.11%)
Mutual labels:  jupyter-notebook
Data science fundamentals
Stars: ✭ 89 (-1.11%)
Mutual labels:  jupyter-notebook
Checklist
Beyond Accuracy: Behavioral Testing of NLP models with CheckList
Stars: ✭ 1,290 (+1333.33%)
Mutual labels:  jupyter-notebook
Spotify Recsys Challenge
A complete set of Recommender Systems techniques used in the Spotify Recsys Challenge 2018 developed by a team of MSc students in Politecnico di Milano.
Stars: ✭ 89 (-1.11%)
Mutual labels:  jupyter-notebook
Adaptivemaskedproxies
Adaptive Masked Proxies for Few Shot Semantic Segmentation
Stars: ✭ 90 (+0%)
Mutual labels:  jupyter-notebook
Deeper Traffic Lights
[repo not maintained] Check out https://diffgram.com if you want to build a visual intelligence
Stars: ✭ 89 (-1.11%)
Mutual labels:  jupyter-notebook
Deeplearning2020
course materials for introduction to deep learning 2020
Stars: ✭ 90 (+0%)
Mutual labels:  jupyter-notebook
Textclustering
Stars: ✭ 89 (-1.11%)
Mutual labels:  jupyter-notebook
Crl
Implementation of the paper "Cascade Residual Learning: A Two-stage Convolutional Neural Network for Stereo Matching"
Stars: ✭ 89 (-1.11%)
Mutual labels:  jupyter-notebook
Pysheaf
Python Cellular Sheaf Library
Stars: ✭ 89 (-1.11%)
Mutual labels:  jupyter-notebook
Programming Concepts For Data Science
Basics of programming: algorithms, data structures, object oriented programming
Stars: ✭ 89 (-1.11%)
Mutual labels:  jupyter-notebook
Computervision tutorials in russian
Stars: ✭ 90 (+0%)
Mutual labels:  jupyter-notebook
Xpedite
A non-sampling profiler purpose built to measure and optimize performance of ultra low latency/real time systems
Stars: ✭ 89 (-1.11%)
Mutual labels:  jupyter-notebook
Psketch
Modular multitask reinforcement learning with policy sketches
Stars: ✭ 89 (-1.11%)
Mutual labels:  jupyter-notebook
Deep Dream In Pytorch
Pytorch implementation of the DeepDream computer vision algorithm
Stars: ✭ 90 (+0%)
Mutual labels:  jupyter-notebook
Logohunter
Deep learning tool to find brand logos in everyday pictures
Stars: ✭ 90 (+0%)
Mutual labels:  jupyter-notebook
Grokking Deep Learning With Julia
Stars: ✭ 90 (+0%)
Mutual labels:  jupyter-notebook

Spatio-Temporal Neural Networks for Space-Time Series Forecasting and Relation Discovery

ICDM 2017 - IEEE International Conference on Data Mining series (ICDM)

Conference Paper

Journal Extension

Commands for reproducing synthetic experiments:

Heat Diffusion

STNN

python train_stnn.py --dataset heat --outputdir output_heat --manualSeed 2021 --xp stnn

STNN-R(efine)

python train_stnn.py --dataset heat --outputdir output_heat --manualSeed 5718 --xp stnn_r --mode refine --patience 800 --l1_rel 1e-8

STNN-D(iscovery)

python train_stnn.py --dataset heat --outputdir output_heat --manualSeed 9690 --xp stnn_d --mode discover --patience 1000 --l1_rel 3e-6

Modulated Heat Diffusion

STNN

python train_stnn.py --dataset heat_m --outputdir output_heat_m --manualSeed 679 --xp stnn

STNN-R(efine)

python train_stnn.py --dataset heat_m --outputdir output_heat_m --manualSeed 3488 --xp stnn_r --mode refine --l1_rel 1e-5

STNN-D(iscovery)

python train_stnn_.py --dataset heat_m --outputdir output_m --xp test --manualSeed 7664 --mode discover --patience 500 --l1_rel 3e-6

Data format

The file heat.csv contains the raw temperature data. The 200 rows correspond to the 200 timestep, and the 41 columns are the 41 space points. The file heat_relations.csv contains the spatial relation between the 41 space points. It is a 41 by 41 adjacency matrix A, where A(i, j) = 1 means that series i is a direct neighbor of series j in space, and is 0 otherwise.

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].