This year I have promised myself to attend as many tech events as I can. So, a couple of days ago, I attended a workshop which is a part of NVIDIA deep learning Institute. The name of the workshop was Fundamentals of deep learning for computer vision.
These kinds of events do not happen very often in where I live. Therefore, I did not know what to expect from such an event. Through this blog post, I will share my experiences related to this workshop.
These are in-person workshops given by NVIDIA DLI Certified instructors. During the event, you learn how to implement and deploy an end-to-end project by getting hands-on training for 8 hours. These workshops are offered at conferences, customer sites and universities. The one that I attended was given by Alptekin Temizel who is an Associate Professor at the Graduate School of Informatics in my university. The workshop was also free to academics and students, if they sign up to the event with their “.edu” mails.
The workshop that I attended included 5 different tasks and 1 assignment in order to get a certificate. At the beginning of each task, the instructor introduced us to the subject, and gave us step-by-step guides on what is expected from us in each task. There were also two other assistants together with the instructor to help the ones who have questions through their tasks.
Since I had already been familiar with the fundamentals of deep learning and computer vision, I did not learn much from the theoretical side of the event. Yet, the slides were really clear for the ones who had little experience with computer vision and deep learning. They were really intriguing.
We started with the a little history and the definition of the deep learning as well as the driving forces that brought us to where we are today for this special topic. Let me give some insights from the beginning of the workshop.
Artificial Intelligence is actually quite an old topic. It started to emerge in 1950s, and with the help of the personal computers it was in our lives. Who doesn’t remember those Tic-Tac-Toe games where the user played opposed to the computer? After that, we started using mails but because of the presence of the malicious emails, we wanted to classify the emails as spam. So, we met with machine learning. The neural networks that we talk about quite often these days also existed in the early 2000’s but at some point we felt kind of stuck. What saved us was the emergence of deep learning. Three things led to the AI boom in early 2010’s:
and BOOM! So, here we are.
Yet, this doesn’t mean that all the problems that we have in our hands should be solved by using deep learning. Neural networks only reside in the Connectionists tribe of machine learning. ML has different tribes and one has to know not only about the Connectionists but also about the others in order to find a good solution with a lower cost.
After this part, we talked about various technical terms such as underfitting / overfitting, learning rate, and so on.
From the practical side, before this event, I had never worked with Caffe framework, so it was totally a plus to learn about a new framework. The tools that are presented to us really made it easier to go through the tasks.
Also, there were a lot of people in the same room who were curious about the same subject, and this led to great conversations and friendships. I also met with someone who will be attending to the same conference with me next week, so we will meet there again.
I had a great day in this event. It was a great opportunity to both connect with new people and learn new things. I really want to participate if they held another workshop similar to this concept.
References:
NVIDIA deep learning, AI, & HPC Classes & Workshops. (n.d.). Retrieved from https://www.nvidia.com/en-us/deep-learning-ai/education/
The Difference Between AI, machine learning, and deep learning? NVIDIA Blog. (2018, September 06). Retrieved from https://blogs.nvidia.com/blog/2016/07/29/whats-difference-artificial-intelligence-machine-learning-deep-learning-ai/