Deep Learning on Azure Hands-on Training!!
Hi! I’m HIRO and member of MSP Japan.
The other day, I organized Deep Learning hands on training with HAIT(Artificial Intelligence Development Group at U.Tokyo).
I would like to report this event.
(This blog was translated by Google translation of what was written in Japanese.)
What is HAIT
HAIT is a student organization composed of members mainly composed of Tokyo University, Waseda and Keio, established in March 2016.
We are working with the goal of “Generating 10,000 AI / data analytical technicians from the same generation of students” and after having entered the foundation of machine learning within just three months, I am aiming for improving skills by conducting AI entrusted development / data analysis and internship in companies.
Although such HAIT is still in its second year of establishment, 110 new students have entered this year, IT university students’ groups are unexpectedly excited. For details, please refer to the link below.
Event situation
Despite the first day of Golden Week for this event, we came to 82 people! This seminar room accommodates 90 people, but almost as packed as you can see!
The venue is surrounded by enthusiasm of a passionate student about AI.
Deep Learning session
Session on Deep Learning in the morning.
At this study meeting, we met with Yoshida who is studying CG and artificial intelligence in the laboratory of the university and two of Kitagawa who has deep knowledge about natural language processing.
First of all, I am 2nd year at Urban University of Tokyo Studies and a session called “Network of us” by MSP Yoshida Menhiro.
We will talk to participants deeply about the story of “drawing AI” using image generation technology and the concept and mind set necessary for engaging in artificial intelligence in the future.
“It is better to listen to the needs of many companies and organizations in the phase of applying research to real society instead of suddenly following the most advanced research on artificial intelligence and try creating a service” It seems that some students have moved to action immediately after the event.
Next is the session “Deep Learning in Natural Language Processing” by Kitagawa who graduated MSP the other day.
Kitagawa belonged to MSP until last year and was mainly active in a project such as Imagine Cup. I graduated from college in March of this year, and now I work as an engineer at Yahoo Japan Corporation.
This session is about natural language processing that was being studied at university. Based on examples such as “School Girls High School AI” Rin “” provided by Microsoft and “Bot” Personality Radio “BOTIO” developed in case of Imagine Cup 2017, He explained the outline and possibilities in detail.
And here is the question corner after the session is over.
80 people participating in the event will attack two of the speakers as questions. From mathematical story to singularity story carefully answered a wide range of questions.
When the question time is over, the morning section is over and lunch break is entered.
Hands-on!!
In the afternoon it is different from the morning and it changes, not learning a story but learning while actually moving the hand. We set up a virtual machine in Microsoft Azure and we will do 80 people to implement a simple image recognition program using that virtual machine.
First of all, MSPs Michigami-kun is in charge of Azure’s environment construction, and one participant places one virtual machine.
And in less than an hour, about 80 virtual machines started up! ! From there, I will baton to Mr. Ishii of HAIT and move on to image recognition.
In this hands-on handwriting recognition will be carried out using the data set called MNIST which is also called ‘Hello World’ of Deep Learning.
More than half of the participants were in the first year at university, but everyone seems to be able to implement handwriting recognition without difficulty.
A social gathering
Lastly, we had a social gathering for deepening friendship among participants.
Here we introduce the situation.
At the end
HAIT who held a study group jointly this time was second year in establishment and history is shallow, but I received the impression that each member is very motivated, active and vigorous student organization.
I hope that we can move forward while competing hard so that MSP does not defeat HAIT.
We look forward to your continued support in the future.
