COURSE OBJECTIVES:
In this course you will learn the fundamentals of Deep Learning through a series of hands on exercises guided by the instructor. Students will learn about the foundational underpinnings of machine learning and deep learning as well as how to put that knowledge into action with practical exercises and homework projects targeted to the day’s lesson. Students will also be challenged to build a larger capstone project throughout the class, which they will present on the final day of class.
The course takes unique project focused approach to teach you deep learning by building deep learning models. The instructor will walk you through a series of curated projects, and explain the key concepts as they arise. Students will learn the theory and how these models work under the hood while writing code, and building neural networks.
In the first half of each class students will review the pre-reading and flesh out the new concepts and theory in a session led by the instructor. In the second half of each class students will immediately put those concepts into practice with a challenging (but achievable) exercise — with support from the instructor and fellow students. Using Tensorflow 2 students will build and train several neural network, using a variety of different architectures to perform:
Classification
Object Localization
Scene Segmentation
Students who take this course will be able to:
Identify and frame problems that can be solved by deep learning
Choose the right techniques to the problems
Understand key deep learning concepts and how deep learning models work
Identify and fix problems with messy datasets
Build deep neural nets for classification and regression using the Keras framework
Build convolutional neural networks for image classification, object localization and segmentation using the Keras
COURSE SCHEDULE:
- Session 1: April 28, 10am PST
- Session 2: April 30, 10am PST
- Session 3: May 5th, 10am PST
- Session 4: May 7th, 10am PST
- Session 5: May 12th, 10am- PST
- Session 6: May 14th, 10am PST
- Session 7: May 19th, 10am PST
- Session 8: May 21st, 10am PST
COURSE INCLUDE:
- 16 hours/ 8 sessions
- 10 lectures / 8 hands-on code labs
- Live Sessions, Real time interaction
- Capstone project, Github portfolio
- Watch sessions replay anywhere any time
- Slack supports to projects and homeworks
COURSE CONTENT:
Check the content tab for full course outlines.
WHO SHOULD LEARN:
Developers, data scientists, students who want to get started on building deep learning projects or applications.
PREREQUISITE:
Familiarity with Python, or willingness to learn it quickly
Basic familiarity with statistics and probability theory is a plus, but not required
Comfort with the command line a plus
FREE TRIAL
We believe the course will be helpful for you like majority of other attendees. Even if you can find it not in your expectation level you will get full refund upon request before the first session ends (April 28th, 2020 12pm PT).
SESSION REPLAY
If you miss the live session or want to learn again, you can watch recorded sessions any time, along with interactive learning tools, slides, course notes
Certificate of completing course is available upon request
Job referral service is available, contact us for details
Scholarship is available, contact us for applications
Module 1: Machine Learning & Neural Network Fundamentals
- Deep learning applications, key concepts and terminology
- Neural network architecture and theory
- Capstone project description will be provided
- Exercise 1: Comparing simple neural networks on the MNIST dataset
Module 2: Deep Neural Networks in Detail
- Identify tasks that are well and ill suited to deep learning
- Feature engineering, Optimizers and training parameters
- Exploring common activation functions
- Exploring common loss functions
- Exploring common optimizers
- Underfitting, overfitting, and regularization tactics
- Exercise 2: Achieve 97% validation accuracy on MNIST
Module 3: Data Cleaning and Preprocessing
- Define and identify common problems with training data
- Describe and apply tactics for improving data quality and improve network performance
- Evaluate the appropriateness of applying said tactics.
- Exercise 3: Clean up a messy dataset
- Exercise 4: Apply data augmentation
Module 4: Convolutional neural networks for classification and segmentation
- Identify tasks for which CNNs are well suited
- Compare and contrast classification with segmentation
- Build CNNs from scratch using Keras and evaluate their performance
- Exercise 5: Achieve 95% accuracy on Fashion MNIST with CNN
Module 5: Convolutional Neural Networks and Transfer Learning
- Transfer learning and identify examples of situations where it might help
- Import well known CNN architectures and leverage transfer learning using Keras
- Exercise 6: Apply Transfer Learning
Module 6: Object Localization and Image Segmentation
- Define object localization and image segmentation and what they’re used for
- Exercise 7: Perform Object Localization
- Exercise 8: Perform Image Segmentation
Module 7: Generative Adversarial Networks (GAN)
- Define GANs and what they’re used for
- Implement a GAN
- Exercise 9: Build a C-GAN
Module 8: Capstone Presentations
- Students present capstone projects
Tyler Bettilyon
Tyler is an educator, technologist, programmer, and all around curious human. He holds a bachelor’s degree in computer science and completed his MBA by counter example in San Francisco’s startup scene. Looking out from inside the Bay Area Bubble he realized that the world is not prepared for the future that technology is bringing. He is now focused on technology education, outreach, and policy
