• Recommendation Systems (study sessions)

    Tokyo Metropolitan Central Library

    ๐Ÿ“Œ Session #2 focuses on following topics: โ— Aki Saarinen: Learnings from building a simple collaborative filtering recommender on the MovieLens dataset โ— Aakash Nand: Topic TBD โ— We are still open for a third topic, please let us know if youโ€™re interested: https://forms.gle/PHSinqG5f8nho3wk7 Going forward, we would like to expand focused discussion to directions aligned with our membersโ€™ interests. The study sessions are organized by Aki Saarinen. http://linkedin.com/in/akisaarinen ๐Ÿ“Œ For each study session we will have: โ— A couple short presentations on topics to spark a discussion. If youโ€™d like to present, please contact the organizers on MLT Slack (#recommendation_systems), or join the meetup and voice your interest. It can be for example learnings from a recommender related project youโ€™re working on, an interesting paper you read, or any other relevant topic youโ€™d like to share with our community. โ— Discussion where we hope participants proactively bring in their questions, experiences, and thoughts, so we can talk together in an interactive format. This is not a lecture series, but rather a forum for us all to learn more on these super interesting timely topics. You can join regardless of your current level, as long as you are interested in understanding how these systems work under-the-hood. ๐Ÿ“Œ RESOURCES โ— Book: Practical Recommender Systems by Kim Falk * A good introduction to recommender systems * Book info: https://www.manning.com/books/practical-recommender-systems * Example project discussed in the book: https://github.com/practical-recommender-systems/moviegeek โ— Introductory free online articles: * https://www.coursera.org/learn/recommender-systems-introduction * https://towardsdatascience.com/introduction-to-recommender-systems-6c66cf15ada * https://hackernoon.com/introduction-to-recommender-system-part-1-collaborative-filtering-singular-value-decomposition-44c9659c5e75 โ— Stanford Course: Mining Massive Datasets (Lecture videos available on Youtube): * Lecture 41 - Overview of recommender systems: https://youtu.be/1JRrCEgiyHM * Lecture 42 - Content-based Recommendations: https://youtu.be/2uxXPzm-7FY * Lecture 43 - Collaborative Filtering: https://youtu.be/h9gpufJFF-0 * Lecture 44 - Implementing Collaborative Filtering: https://youtu.be/6BTLobS7AU8 ๐Ÿ’™ โ€“โ€“ MLT PATRON โ€“โ€“ Always on the waiting list for MLT events? Become a MLT Patron and get early access to workshops, study groups and talks. https://www.patreon.com/MLTOKYO โ€“โ€“ FIND RESOURCES โ€“โ€“ Github: https://github.com/Machine-Learning-Tokyo Youtube: https://www.youtube.com/MLTOKYO Slack: https://goo.gl/WnbYUP โ€“โ€“ RECRUITING โ€“โ€“ MLT events are for community building and knowledge sharing. We politely ask that company representatives, recruiters and consultants looking to hire or sell their services do not come to this event. โ€“โ€“ CODE OF CONDUCT โ€“โ€“ MLT promotes an inclusive environment that values integrity, openness and respect. https://github.com/Machine-Learning-Tokyo/MLT_starterkit

    7
  • NLP paper reading session: "The Natural Language Decathlon"

    Tokyo Metropolitan Central Library

    ๐Ÿ“Œ In the third NLP paper reading session we will discuss the paper "The Natural Language Decathlon: Multitask Learning as Question Answering" (https://arxiv.org/pdf/1806.08730.pdf) The paper focuses on multi-task learning. It introduces the Natural Language Decathlon (decaNLP), a framework in which multitask learning is cast and made easy to understand as a question answering. We will discuss decaNLP and the multitask question answering network (MQAN) that jointly learns ten tasks: question answering, machine translation, summarization, natural language inference, sentiment analysis, semantic role labeling, relation extraction, goal-oriented dialogue, semantic parsing, and commonsense pronoun resolution. ๐Ÿ“Œโ€“โ€“ SCHEDULE โ€“โ€“ โ€“โ€“ Preparation: September 30 - October 5: Self reading, and preparing. โ— October 6: Paper reading session. โ— 12:50 (Optional) Meet for lunch. โ— 14:00 - 16:00 Interactive paper presentation Happy Buzaaba Reading Guidelines: https://discuss.mltokyo.ai/t/paper-reading-guidelines/241 โ€“โ€“ FIND RESOURCES โ€“โ€“ Github: https://github.com/Machine-Learning-Tokyo Youtube: https://www.youtube.com/MLTOKYO Slack: https://goo.gl/WnbYUP โ€“โ€“ RECRUITING โ€“โ€“ MLT events are for community building and knowledge sharing. We politely ask that company representatives, recruiters and consultants looking to hire or sell their services do not come to this event. โ€“โ€“ CODE OF CONDUCT โ€“โ€“ MLT promotes an inclusive environment that values integrity, openness and respect. https://github.com/Machine-Learning-Tokyo/MLT_starterkit

  • MLT ไธ€่ˆฌ็คพๅ›ฃๆณ•ไบบ NPO Launch

    We are thrilled to announce our MLT ไธ€่ˆฌ็คพๅ›ฃๆณ•ไบบ NPO Launch Event and partnership with Mistletoe Japan, Inc. and would like to invite the community to join us. We're honored to welcome Serial Entrepreneur and Investor Taizo Son for a Fireside Chat tonight at the event. ๐Ÿค–MLT is all about open source, open education and open science. The organization is volunteer-driven and consists of a core team of ML Engineers and Researchers, 40-50 active contributors and 3,500 members in Tokyo. MLT held more than 60 technical Deep Learning workshops, talks and study sessions with more than 4,000 participants in the past 1.5 years. https://mltokyo.ai/ ๐Ÿค–Mistletoe Japan, Inc. is a Collective Impact Community with the mission to re-create a sustainable human-centered future using technology. The community is made up of those who lead the forefront of the global startup movement including entrepreneurs, investors, researchers and visionaries, with the mutual goal to solve global social challenges that humanity will face in the near future. Main activities range from startup investment, research & development, joint ventures to ecosystem development. http://mistletoe.co/en/ Join us for a wonderful evening, celebrating AI, the community, the team and our new partnership. ๐Ÿ“Œ EVENT SCHEDULE 7:00 pm Doors open 7:30 pm Machine Learning Tokyo 7:50 pm Fireside Chat with Taizo Son, Mistletoe Japan, Inc. Light finger food and drinks will be provided. ๐Ÿ’™-- THANK YOU -- A huge THANK YOU goes out to Mistletoe of Tokyo for having us for this event. Mistletoe of Tokyo is a place to connect with people who are searching for a better world and thinking about the future. https://mistletoe.tokyo/about/

    7
  • Python tools for Deep Learning--Chainer, CuPy, and Optuna

    ๆ ชๅผไผš็คพPreferred Networks

    We're excited to explore the Deep Learning framework Chainer with a workshop by Crissman Loomis, Engineer / Business Development, Preferred Networks, Inc. Starting with an introduction to the Deep Learning framework Chainer, Crissman will present how to structure basic Deep Learning models. Also covered will be CuPy, a NumPy-like API for calculations on NVIDIA GPUS, and Optuna, a hyper parameter optimization library. Attendees will be able to do hands-on work using Google Colab and to understand the uses and benefits of all three open source projects. ๐Ÿ“Œโ€“โ€“ SCHEDULE โ€“โ€“ 6:30 pm Doors open 7:00 pm โ€“ 9:00 pm Workshop 9:00 pm โ€“ 9:30 pm Wrap up and Closing โ€“โ€“ SPEAKER BIO โ€“โ€“ Crissman has worked at Preferred Networks on the Chainer team for over two years, focusing on improving the documentation for Chainer and giving presentations on Chainer at Open Data Science Conferences, Euro SciPy, PyCon, GTC, and other venues. His ODSC West workshop on Chainer was selected as one of the top 10 workshops for learning Machine Learning. โ€“โ€“ ACCESS DETAILS โ€“โ€“ Meetup area is located on the 3nd floor of Otemachi building which is directly connected to subway station Otemachi. If you take the C7 exit, you will find the entrance to the Otemachi building, then enter the building and take the nearest stairways to 3rd floor and turn right. The meetup area is just some steps away from the stairways. ๐Ÿ’™ โ€“โ€“ THANK YOU โ€“โ€“ A big Thank You goes out to Preferred Networks for having us! PFN is a Tokyo-based AI Startup with Deep Learning at its core โ€“ from developing the deep learning framework Chainer and building large-scale compute clusters, to exploring a wide variety of domains (including robotics, life science, and others). https://preferred.jp/en/ โ€“โ€“ MLT PATRON โ€“โ€“ Always on the waiting list for MLT events? Become a MLT Patron and get early access to workshops, talks and study sessions. https://www.patreon.com/MLTOKYO โ€“โ€“ SUBSCRIBE โ€“โ€“ Subscribe to our monthly newsletter: https://mltokyo.ai/membership-join โ€“โ€“ FIND MLT RESOURCES โ€“โ€“ Github: https://github.com/Machine-Learning-Tokyo Youtube: https://www.youtube.com/MLTOKYO Slack: https://goo.gl/WnbYUP โ€“โ€“ RECRUITING โ€“โ€“ MLT events are for community building and knowledge sharing. We politely ask that company representatives, recruiters and consultants looking to hire or sell their services do not come to this event.

    9
  • Recommendation Systems (study group)

    Tokyo Metropolitan Central Library

    ๐Ÿ“Œ Session #1 focuses on following topics: โ— Discussion on participants' goals for the study group โ— Overview: what are the different types of recommender systems in common use, and what are they used for โ— Collecting user data, producing non-personalized recommendations, and handling the cold-start problem โ— Collaborative filtering: Theory, examples and how to evaluate a recommender system Going forward, we would like to expand focused discussion to directions aligned with our members' interests. The study sessions are organized by Aki Saarinen http://linkedin.com/in/akisaarinen ๐Ÿ“ŒFor each study session we will have: โ— A set of home study materials we hope participants take a look at before joining. We may base some of the discussion on books, but it's optional and we also link to free online resources. โ— A short summary presentation of the topics for the day at the beginning of the session (prepared by organizer). This will hopely get us all the a similar starting line. โ— A set of discussion areas where we hope participants proactively bring in their questions, experiences, and thoughts, so we can talk together in an interactive format. This is not a lecture series, but rather a forum for us all to learn more on these super interesting timely topics. You can join regardless of your current level, as long as you are interested in understanding how these systems work under-the-hood. ๐Ÿ“Œ RESOURCES โ— Book: Practical Recommender Systems by Kim Falk * Book info: https://www.manning.com/books/practical-recommender-systems * Example project discussed in the book: https://github.com/practical-recommender-systems/moviegeek โ— Introductory free online articles: * https://www.coursera.org/learn/recommender-systems-introduction * https://towardsdatascience.com/introduction-to-recommender-systems-6c66cf15ada * https://hackernoon.com/introduction-to-recommender-system-part-1-collaborative-filtering-singular-value-decomposition-44c9659c5e75 โ— Stanford Course: Mining Massive Datasets (Lecture videos available on Youtube): * Lecture 41 - Overview of recommender systems: https://youtu.be/1JRrCEgiyHM * Lecture 42 - Content-based Recommendations: https://youtu.be/2uxXPzm-7FY * Lecture 43 - Collaborative Filtering: https://youtu.be/h9gpufJFF-0 * Lecture 44 - Implementing Collaborative Filtering: https://youtu.be/6BTLobS7AU8 โ€“โ€“ MLT PATRON โ€“โ€“ Always on the waiting list for MLT events? Become a MLT Patron and get early access to workshops, study groups and talks. https://www.patreon.com/MLTOKYO โ€“โ€“ SUBSCRIBE โ€“โ€“ Subscribe to our monthly newsletter: https://mltokyo.ai/membership-join โ€“โ€“ FIND RESOURCES โ€“โ€“ Github: https://github.com/Machine-Learning-Tokyo Youtube: https://www.youtube.com/MLTOKYO Slack: https://goo.gl/WnbYUP โ€“โ€“ RECRUITING โ€“โ€“ MLT events are for community building and knowledge sharing. We politely ask that company representatives, recruiters and consultants looking to hire or sell their services do not come to this event. โ€“โ€“ CODE OF CONDUCT โ€“โ€“ MLT promotes an inclusive environment that values integrity, openness and respect. https://github.com/Machine-Learning-Tokyo/MLT_starterkit

    10
  • NLP & ML: Build a Chatbot from Scratch

    Code Chrysalis

    Ever wondered how chatbots work? Chris Gerpheide is leading the engineering team at Bespoke, the creators of Bebot, a chatbot for hospitality and tourism in Japan. Bespoke develops all of their technology in house; during the workshop, Chris will show you some simple machine learning and natural language processing techniques to create the guts of a basic chatbot from scratch. This is an interactive workshop, where each participant can create their own simple chatbot backend in Python using their own laptops. ๐Ÿ“ŒSCHEDULE โ— 7:00pm Doors open โ— 7:30pm Introduction and Presentation โ— 8:00pm Workshop โ— 9:30pm Wrap up โ€“โ€“ SPEAKER BIO โ€“โ€“ Chris Gerpheide leads the engineering team at Bespoke Inc. https://www.be-spoke.io Before joining Bespoke, she was engineering manager at Amazon Web Services. In her free time, she enjoys learning Japanese, hiking, and teaching children programming. ๐Ÿ’™ โ€“โ€“ THANK YOU โ€“โ€“ A big Thank You goes out to Code Chrysalis for having us! Code Chrysalis is a 12-week, full-time advanced coding bootcamp located in the heart of Tokyo ๐Ÿ—ผ. See why we are an industry leader in technical education in Japan ๐Ÿ—พ. They also provide a wide range of technical workshops and talks for the community. https://www.codechrysalis.io โ€“โ€“ MLT PATRON โ€“โ€“ Always on the waiting list for MLT events? Become a MLT Patron and get early access to workshops, talks and study sessions. https://www.patreon.com/MLTOKYO โ€“โ€“ SUBSCRIBE โ€“โ€“ Subscribe to our monthly newsletter: https://mltokyo.ai/membership-join โ€“โ€“ FIND MLT RESOURCES โ€“โ€“ Github: https://github.com/Machine-Learning-Tokyo Youtube: https://www.youtube.com/MLTOKYO Slack: https://goo.gl/WnbYUP โ€“โ€“ RECRUITING โ€“โ€“ MLT events are for community building and knowledge sharing. We politely ask that company representatives, recruiters and consultants looking to hire or sell their services do not come to this event. โ€“โ€“ CODE OF CONDUCT โ€“โ€“ MLT promotes an inclusive environment that values integrity, openness and respect. https://github.com/Machine-Learning-Tokyo/MLT_starterkit

    6
  • Deep Perception Hackathon: Deep Learning in VR/AR/MR

    MISTLETOE OF TOKYO

    Machine Learning Tokyo, the Digital Nature Group at the University of Tsukuba and ON-1 are organizing a 2-day hackathon to combine Deep Learning with interactive AR/VR/MR environments. The goal is to explore both fields and find novel ways of synthesis and interaction. We'll have teams of 3-5 people with different technical and creative backgrounds to work on diverse project ideas which will be presented at the end of the hackathon. The event will be held at Mistletoe of Tokyo. Play with different VR hardware like HTC Vive, Oculus RIFT and complementary hardware like the Jetson Nano. ๐Ÿ›‘โ€“โ€“ APPLICATION: OPEN UNTIL SEPTEMBER 7 โ€“โ€“ ๐Ÿ›‘ We have 40 open spots for engineers, artists and designers. The application form is open until September 7, accepted participants will be notified on September 8. Please apply here in this form: https://forms.gle/fFv3UBXFwfjC19R8A ๐Ÿ’ก โ€“โ€“ SCHEDULE โ€“โ€“ ๐Ÿค– DAY 1 | SATURDAY, SEPTEMBER 14 9:00 am - Welcome note 9:30 am - Team building and start hacking 12:30 am - Lunch break 5:30 pm - Invited talk and Q&A 8:00 pm - Closing Day 1 ๐Ÿค– DAY 2 | SUNDAY, SEPTEMBER 15 9:00 am - Start hacking 12:30 am - Lunch break 5:30 pm - Project Presentations 6:30 pm - Dinner & Drinks 7:00 pm - Announcing winning teams 8:00 pm - Closing Day 2 ๐Ÿ’™-- THANK YOU -- A huge THANK YOU goes out to Mistletoe of Tokyo for hosting this event. Mistletoe of Tokyo is a place to connect with people who are searching for a better world and thinking about the future. https://mistletoe.tokyo/about/ -- MLT PATRON -- Become a MLT Patron and help us to keep MLT meetups like this inclusive and for free. https://www.patreon.com/MLTOKYO Find MLT resources Github: https://github.com/Machine-Learning-Tokyo Youtube: https://www.youtube.com/MLTOKYO -- RECRUITING -- MLT events are for community building and knowledge sharing. We politely ask that company representatives, recruiters and consultants looking to hire or sell their services do not come to this event. Image source: https://www.stambol.com/2018/05/28/the-future-of-ar-vr-headset-design-is-hybrid/

    7
  • Generators, Manifolds, and Adversarial Mixup Resynthesis, Talk & Discussion

    Economics Research Annex(Kojima Hall)

    We're excited to welcome Alex Lamb for a talk on "Generators, Manifolds, and Adversarial Mixup Resynthesis" at the University of Tokyo. A motivating idea behind deep learning is that a model can learn to map from the high-dimensional space of observations unto a low-dimensional space of salient explanatory factors which vary across the data. Generative models with latent variables perhaps embody this idea most closely. In this talk we'll explore the relationship between deep models and manifolds, trying to make the relationship more explicit and rigorous. At the same time we'll discuss a recent paper, "Adversarial Mixup Resynthesis", that takes a new perspective on latent variables in generative models. ๐Ÿ“ŒSCHEDULE โ— 7:00pm Doors open โ— 7:30pm โ€“ 8:20pm "Generators, Manifolds, and Adversarial Mixup Resynthesis", Alex Lamb โ— 8:20pm โ€“ 8:45pm Q&A โ— 8:45pm โ€“ 9:00pm Closing โ€“โ€“ SPEAKER BIO โ€“โ€“ Alex Lamb grew up in Western Maryland and did his undergraduate at Johns Hopkins University. Afterwards he worked on new forecasting algorithms and systems at Amazon for a few years. Now he's a PhD student at the Montreal Institute for Learning Algorithms (MILA) in Yoshua Bengio's lab, and enjoys working on new algorithms for deep learning. On the side he also enjoys working on medieval Japanese document recognition. http://alexmlamb.github.io/ ๐Ÿ’™ โ€“โ€“ THANK YOU โ€“โ€“ A big Thank You goes out to the University of Tokyo and Michael Fabinger for hosting this event. http://www.cirje.e.u-tokyo.ac.jp/research/workshops/stateng/stateng.html โ€“โ€“ MLT PATRON โ€“โ€“ Become a MLT Patron and help us to keep MLT meetups like this inclusive and for free. https://www.patreon.com/MLTOKYO โ€“โ€“ SUBSCRIBE โ€“โ€“ Subscribe to our monthly newsletter: https://mltokyo.ai/membership-join โ€“โ€“ FIND MLT RESOURCES โ€“โ€“ Github: https://github.com/Machine-Learning-Tokyo Youtube: https://www.youtube.com/MLTOKYO Slack: https://goo.gl/WnbYUP โ€“โ€“ RECRUITING โ€“โ€“ MLT events are for community building and knowledge sharing. We politely ask that company representatives, recruiters and consultants looking to hire or sell their services do not come to this event. โ€“โ€“ CODE OF CONDUCT โ€“โ€“ MLT promotes an inclusive environment that values integrity, openness and respect. https://github.com/Machine-Learning-Tokyo/MLT_starterkit

    3
  • Decision (Neuro)Science in Society, Talk & Discussion

    Economics Research Annex(Kojima Hall)

    We're excited to collaborate with the RIKEN Center for Brain Science and welcome Decision Neuroscientist and RIKEN CBS Unit Leader Rei Akaishi for a talk & discussion at the University of Tokyo. Data and its analysis create value by assisting humans in the decision making process. The mergence of decision science and data science opens new perspectives and opportunities for research and industry. In this talk Rei Akaishi gives an introduction to decision neuroscience and behavioral economics and discusses the synthesis of decision science and data science with the audience. ๐Ÿ“ŒSCHEDULE โ— 7:00pm Doors open โ— 7:30pm โ€“ 8:20pm Decision (Neuro)Science in Society, Rei Akaishi, RIKEN Center for Brain Science โ— 8:20pm โ€“ 8:45pm Interactive discussion โ— 8:45pm โ€“ 9:00pm Closing โ€“โ€“ SPEAKER INFO โ€“โ€“ Rei Akaishi obtained his PhD from Graduate School of Medicine, the University of Tokyo. He did his post-doctoral jobs in University of Oxford, Tokyo Metropolitan Institute of Medical Science, University of Rochester, Center for Information and Neural Networks and conducted research projects on decision making of humans and animals. He is currently a Unit Leader in RIKEN Center for Brain Science. ๐Ÿ’™ โ€“โ€“ THANK YOU โ€“โ€“ A big Thank You goes out to the University of Tokyo and Michael Fabinger for hosting this event. http://www.cirje.e.u-tokyo.ac.jp/research/workshops/micro/micro.html โ€“โ€“ MLT PATRON โ€“โ€“ Become a MLT Patron and help us to keep MLT meetups like this inclusive and for free. https://www.patreon.com/MLTOKYO โ€“โ€“ SUBSCRIBE โ€“โ€“ Subscribe to our monthly newsletter: https://mltokyo.ai/membership-join โ€“โ€“ FIND MLT RESOURCES โ€“โ€“ Github: https://github.com/Machine-Learning-Tokyo Youtube: https://www.youtube.com/MLTOKYO Slack: https://goo.gl/WnbYUP โ€“โ€“ RECRUITING โ€“โ€“ MLT events are for community building and knowledge sharing. We politely ask that company representatives, recruiters and consultants looking to hire or sell their services do not come to this event. โ€“โ€“ CODE OF CONDUCT โ€“โ€“ MLT promotes an inclusive environment that values integrity, openness and respect. https://github.com/Machine-Learning-Tokyo/MLT_starterkit

    11
  • Machine Learning in Production

    Tokyo Chapter - ninetytwo13

    We're excited to turn our focus to Machine Learning in production environments with a workshop by Adam Gibson, Co-Founder and CTO of Skymind, a SF/Tokyo-based AI Startup specialized in production-level Machine Learning and Deep Learning for industry. โ€“โ€“ SCHEDULE โ€“โ€“ โ— 13:00-13:50: Setup (Pre-install docker if you can) โ— 13:50-14:00: Break โ— 14:00-14:50: Serving models Hands on/Explanation โ— 15:00-15:50: Trade offs/gotchas with traditional applications (packaging, distributed systems,..) and general example scenarios โ— 16:00-16:15: Break โ— 16:15-17:00: Model Serving trade offs with GPUs vs cpus โ€“โ€“ ABSTRACT โ€“โ€“ This workshop will cover the first steps of deploying machine learning models to production. At the core of the workshop will be a new open source library called pipelines: https://github.com/SkymindIO/pipelines - we will use the library to show how to run python scripts in production as well as how to deal with containerized applications. โ€“โ€“ PREREQUISITE โ€“โ€“ โ— Hands-on experience with prototyping Machine Learning and Deep Learning models โ€“โ€“ SPEAKER BIO โ€“โ€“ Adam Gibson is the Co-founder of Skymind, an open source company focused on machine learning infrastructure. Adam is also a book author for Oโ€™Reilly: https://www.amazon.com/Deep-Learning-Practitioners-Josh-Patterson/dp/1491914254 โ€“โ€“ THANK YOU โ€“โ€“ Thank you to Tokyo Chapter for sponsoring the venue. Cradled in Tokyo's most vibrant and multicultural neighborhood directly across Tokyo Midtown and Hinokicho Park, Tokyo Chapter at ninetytwo13 offers inspiring collaborative workspaces and living spaces centered around a community of creators, from writers and artists, to film makers and entrepreneurs. https://tokyochapter.com/ ๐Ÿ’™โ€“โ€“ MLT PATRON โ€“โ€“ Become a MLT Patron and help us to keep MLT meetups like this inclusive and for free. https://www.patreon.com/MLTOKYO โ€“โ€“ MLT RESOURCES & COMMUNITY CHAT โ€“โ€“ Github: https://github.com/Machine-Learning-Tokyo Youtube: https://www.youtube.com/MLTOKYO Slack: https://goo.gl/WnbYUP โ€“โ€“ RECRUITING โ€“โ€“ MLT events are for community building and knowledge sharing. We politely ask that company representatives, recruiters and consultants looking to hire or sell their services do not come to this event.

    16