Data Science Dojo is sponsoring Kaggle Days Tokyo, and I want to continue the conversation about what Kaggle Days are and cover the agenda for the Tokyo event.
A Brief Overview
Kaggle is an online learning platform for data science and machine learning. The educator uses competitions to help its users (called Kaggelers) practice and grow their data science skillset with publicly available datasets.
Kaggle Days are events that take place around the world. They started out as a partnership between LogicAI and Kaggle as a way to bring Kaggelers together for an offline event. Competitions, seminars, workshops, and networking opportunities are available for Kaggelers to participate in. These events take place as one-off local events (Meetups) as well as multiday global events (conferences).
Kaggle Days Tokyo – Agenda
The global event in Tokyo is taking place this December 11-12. Registration closed within a matter of days of opening, which shows the amount of popularity these events have among its participants. The agenda is jam-packed with exciting talks and tutorials from Kaggle Grandmasters and data science professionals, and I’d like to highlight a few.
Raja Iqbal – Tutorial on Model Validation and Parameter Tuning
Raja Iqbal is the CEO, Chief Data Scientist, and Lead Instructor at Data Science Dojo. He has an MS from Stanford and PhD from Tulane University. He spent more than 6 years at Microsoft Bing and Bing Ads working on various data science and machine learning research projects. Below is a description, given by Raja, of his workshop:
“Cross validation is a popular technique for model validation and parameter tuning. In this tutorial we will discuss other model validation and parameter techniques in scenarios where k-fold cross validation may not be the best choice. We will also discuss some parametric and non-parametric statistical tests for comparing models.”
Why should you attend?
Modern machine learning is about gathering the right data, feature engineering, validation, and parameter tuning. Not understanding the concepts or using the techniques correctly renders machine learning useless.
Time: 10:15 am – 11:45 am
Location: 27F Hanabi Room
Jin Zhan – My Jouney to Grandmaster: Success and Failure
Becoming a Kaggle Grandmaster (GM) is no small accomplishment. It takes years of practice to obtain this impressive title. Jin Zhan has multiple years of experience in data science and machine learning, as well as Hadoop. Currently, Zhan is a Data Scientist at Fast Retailing, where he focuses on demand forecasting, recommender systems, and customer comment analysis.
Why should you attend?
The original reason I chose this out of the bunch was because Jin is going to talk about his failures before becoming a grandmaster. Talking about our failures is often difficult, but we can learn more from them than our success. After (admittedly) combing through his LinkedIn profile, I found Zhan to be the perfect picture of success on Kaggle. His experience doesn’t come from one place and his education comes from multiple sources. Besides, Zhan’s a Grandmaster. What other reason to attend do you need?
Time: 4:35 pm – 4:45 pm
Location: 27F Matsuri Room
During a Kaggle competition, typically the only help or mentoring you receive is from your teammates or through Kaggle Kernels. At the competition in Tokyo (as well as the other global events) mentors will be available to help you along the way.
The mentors for the competition in Tokyo include:
- Ryuji Sakata – Kaggle Grandmaster and Data Scientist/Researcher at Panasonic Corporation
- Walter Reade – Data Scientist on Kaggle Competitions Team
- Dimitry Gordeev – Kaggle Grandmaster and Data Scientist at UNIQA
- Pawel Jankiewicz – Kaggle Grandmaster and Owner/Founder at LogicAI
- Jin Zhan – Kaggle Grandmaster and Data Scientist at Fast Retailing
You should feel compelled to pick their brains as much as you can. All of these people are successful and established data scientists with an extensive knowledge of Kaggle competitions. Get as much out of them as you can.
Time: 10:30 am – 6:30 pm