My Journey with Udacity’s Deep Learning Nanodegree Program
By — P Nithin Jayakar
Udacity?
“e-learning” — possibility of losing interest is very high in the journey.
BUT!!!
Udacity has a special way to maintain enthusiasm for every next video, next project, next Nanodegree Program. Trust me it will make you addictive towards learning.
“Udacity” — A tech DRUG!!!
“In general every one of us is a teacher.” — OK….Is it possible to teach a machine? Fortunately YES!!! This is where I started. I was fed up of manually deleting duplicate & blurry images from directories. Since I am a passionate CSE guy , I tried to find a solution to this problem by using technology, & at last I found a way to build a model using Machine Learning. I started learning Machine Learning especially fell in love with a section in Machine Learning known as Deep Learning.
“Deep Learning” — The spinal cord for Artificial intelligence.
I believe it will open the doors of intelligence one by one for a machine in particular conditions. I enrolled in Deep Learning Nanodgree with a fear(will I make it?)but now I have the courage to complete my nanodgree. Thanks to Udacity! It formed a one way bridge between fear and courage. Every quiz at videos made me more intuitive about the concepts. I would sometimes jump into other sections of the course just for brief insight. During the process I got a lot of exposure in the form of :
- Slack channel — the best part is office hours where I can interact with industrial experts directly.
2. Project review — they have excellent supporting review system where I learnt in depth about some difficult concepts like parameter tuning.
3. Forums — a place to learn and clarify doubts with a great community.
3. Mentor support — a dedicated mentor checks your progress every week.I am lucky to have my mentor @NG Fang.
4. AWS credits — gpu support is the best part for experimenting on projects.
DLND has 6 sections with 5 projects, |Introduction|Neural Networks|Constitutional Networks|Recurrent Networks|Generative adversarial Networks|Deep reinforcement learning|
In section-1 a good knowledge about deep learning along with some basics was given.
In section-2 I learned how to build a neural network, how neural network works, how to tune the parameters much more. After completing the section I had submitted a project. Working on this project review gave me a complete practical overview on the neural network.
Going through section-3 I learned about CNN and image processing to build a neural network, which analyzes dog image and predicts it’s breed. I found this project the most amazing and fun project of all.
Moving on to section-4 I built a RNN in tensor flow to process text and generate a new TV script.
At present I am going through section-5 where I will soon build a pair of multi layer neural networks and make them compete with each other in order to generate faces.
At every project I had a milestone to achieve, but still I came through. All thanks to the amazing support and help that my mentors and community gave me! The best part of Udacity is that it flexibly fir with my time & weekly goals and bracket due dates helped me to stay on track.
“Delay is nothing decision is everything”
Enroll into Deep Learning Nanodegree here
About the Author | P Nithin Jayakar
Al Optimist. •Epigrammatist. •Sketch Artist. •Palestrato.