Hyperparameter Tuning with Neural Network Intelligence

Overview

Discover the power of Microsoft's Neural Network Intelligence (NNI) in our comprehensive 2-hour guided project, designed to equip participants with fundamental knowledge and practical skills in hyperparameter tuning using the NNI toolkit. The open-source AutoML toolkit by Microsoft facilitates the automation of various machine learning processes such as feature engineering, hyperparameter tuning, neural architecture search, and model compression.

This session focuses on hyperparameter tuning to optimize neural network performance.

Participants will gain hands-on experience with the NNI toolkit, exploring its capabilities in conducting a hyperparameter tuning experiment. Utilizing the MNIST dataset, known for its hand-written digit images, learners will train a basic neural network.

The project aims to fine-tune various parameters including batch size, learning rate, activation function choice for the hidden layer, number of units in the hidden layer, and the dropout rate to enhance model accuracy.

To fully benefit from this guided project, a competent understanding of the Python programming language is essential. Familiarity with neural networks and frameworks like TensorFlow and Keras will also be beneficial.

This course is optimized for learners in the North America region, with efforts underway to extend a similar learning experience to other regions.

Provided by Coursera, this project falls under multiple categories, offering a rich learning experience in Python Programming, Neural Networks, TensorFlow, Keras, Hyperparameter Optimization, and Hyperparameter Tuning.

Taught by

Amit Yadav

Subjects