Introduction
This repository contains projects for time-series forecasting and image classification. Codes are written in Python.
-
📗 First Topic: Data Driven Prediction of Dynamical Systems
I have exploited ConvLSTM neural networks to predict solution of Inviscid Burgers equation. Click the link below for more details.
-
📙 Second Topic: Convolutional Neural Networks with TensorFlow
This is a tutorial project to show how to use TensorFlow to load data and build simple models for prediction. Click the link below for more details.
-
📘 Third Topic: Image Classification with Very Deep Convolutional Neural Networks
This is a project on Kaggle for detecting Cassava Leaf disease. I have implemented a pretrained Efficientnet with TensorFlow. EfficientNet is a very deep convolutional neural network architecture which uses a scaling method to uniformly scale the dimensions. Click the link below for more details.
Topics
| Topics | |
|---|---|
| Data Driven Prediction of Dynamical Systems | |
| Convolutional Neural Networks with TensorFlow | |
| Image Classification with Very Deep Convolutional Neural Networks |
References
- M. Tan and Q. V. Le, "EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks", arXiv:1905.11946, 2019
- U. Achatz, S. I. Dolaptchiev, I. Timofeyev, "Subgrid-scale closure for the inviscid Burgers-Hopf equation", Communications in Mathematical Sciences, 2013
- A. Chattopadhyay, P. Hassanzadeh, and D. Subramanian, "Data-driven prediction of a multi-scale Lorenz 96 chaotic system using deep learning methods: Reservoir computing, ANN, and RNN-LSTM", Nonlinear Processes in Geophysics, 2020