Udacity is excited to reintroduce an update to a program in our School of Artificial Intelligence: the Deep Learning Nanodegree program. Originally launched a few years ago, the new refresh maintains all the important parts of the old content while streamlining the lessons and adding a few more modern concepts to the courses. This modern take on our highly successful Nanodegree program has everything you need to be a skilled deep learning professional. Keep reading to learn more.
What is Deep Learning?
For those who don’t know, deep learning is a highly advanced version of machine learning. Essentially, it is many layers of a neural network that has the ability to learn new concepts without human intervention or training — a concept known as unsupervised learning. Deep learning is commonly used in cutting-edge technologies, like natural language processing and facial recognition.
Deep Learning Nanodegree Program Details
The Deep Learning Nanodegree program will provide students with a solid introduction to the world of artificial intelligence (AI). In particular, this program focuses on helping learners master the fundamentals, then gain a deeper understanding of the field by delving into cutting-edge topics such as Neural Networks, Convolutional Neural Networks, Recurrent Neural Networks, and Generative Adversarial Networks.
To get the most out of this program, it’s important to have experience working with or a thorough understanding of:
- Derivatives
- Linear Algebra
- Vectors
- Matrix Multiplication
- Python Data Types
- Numpy and Pandas
- Basic python
- Generators
- Jupyter notebooks
Additionally, students will need a computer with the ability to access NLTK, SKLearn, BeautifulSoup, and Numpy.
In as little as four months (at 10 hours a week), students who enroll in the Deep Learning Nanodegree program will learn how to create neural networks (NN), build convolutional neural networks (CNN), and implement recurrent neural networks (RNN), as well as RNN variants, all using PyTorch.
The Deep Learning Nanodegree Projects
Project 1: Developing a Handwritten Digits Classifier with PyTorch
Using PyTorch, students will create a handwritten digit recognition system by developing a neural network. The NN will be trained via a training loop that uses preprocessed data, loaded by the students into PyTorch. Once the model is trained, students can apply additional advanced training techniques to increase the accuracy of their handwritten digit recognition system.
Project 2: Landmark Classification and Tagging for Social Media
Students will design a program to automatically classify landmarks around the world in user photos. The landmark classification project will be designed by students from end to end, including performing data preprocessing, designing and training CNNs, comparing the accuracy of different CNNs, and deploying an app based on the best CNN that was trained.
Project 3: LSTM Seq2Seq Chatbot
Students will build an AI chatbot using LSTMs, Seq2Seq, and word embeddings for increased accuracy. First, they will write a neural network architecture using PyTorch, then they will train it with a dataset of conversational dialogue, and finally, they will tune the network hyperparameters to increase the accuracy.
Project 4: Face Generation
Students will build and train a custom GAN architecture, including a generator and discriminator, on the CelebA dataset. Next, students will try out various loss functions they learned about, including the Binary Cross Entropy and the Wasserstein loss. Finally, students will try out training stabilization methods, such as label smoothing.
Learning from Top Professionals in Artificial Intelligence
To develop this program’s world-class curriculum, we collaborated with professionals from top-rated tech companies. Each of these collaborators contributed guidance and feedback to focus the program on the most in-demand skills. Each of the instructors has extensive AI, deep learning, and teaching experience.
Instructors
- Erick Galinkin, Principal AI Researcher at Rapid7
- Giacomo Vianello, Principal Data Scientist
- Nathan Klarer, Head of ML & COO at Datyra
- Thomas Hossler, Sr. Machine Learning Engineer
Enroll in the Refreshed Deep Learning Nanodegree Today
If you’re a software engineer, data scientist, or any kind of tech professional who is interested in learning more about machine learning/deep learning, this is the Nanodegree program for you. This program is a quick, yet comprehensive, way to build foundational hands-on skills to be used in real-life projects for cutting-edge technology.
It’s no secret that artificial intelligence is the next big thing. Many companies are discovering ways that it can be employed to increase efficiency, expand their business, and do a plethora of exciting new things. In order to stay ahead of the curve, businesses are willing to pay big bucks to well-trained machine learning engineers. According to Glassdoor, the average salary of a machine learning engineer in the US is almost $125,000 a year.
With Udacity’s combination of hands-on project-centric learning and mentorship, there’s no better way to meet the demand than by registering today for the Deep Learning Nanodegree program. Enroll now to learn more!
