Along with the development of machine learning or deep learning (deep learning) in recent years, systems using machine learning are rapidly spreading in society. On the other hand, however, various software engineering methods used in conventional IT systems have become unusable enough to say at all before a system incorporating machine learning (machine learning system) . The methodology of development, testing and operation of the machine learning system has not been established yet, and at the development site engineers are managed to surpass it by trial and error.
Based on such a situation, we think that it is necessary to establish and systematize a new paradigm which should be called "machine learning engineering" for machine learning systems. In the first half of the workshop, we will present the position and clarify the position of the participants, and in the second half we will discuss based on Open Space Technology.
Workshop execution chairperson
Kenji Morita (Preferred Networks)
Yoshioka Shinwa (NII)
Program members
Fuyuki Ishikawa (NII)
Taka Imai (LeapMind)
Hiroshi Maruyama (Preferred Networks)
Taku Doi (Level 5)
Mihisa Ota (Brainpad)
Ryosuke Yoshizaki (Kikagaku)
OST Facilitator
Yoshinobu Kawaguchi (Rakuten)
topic
Management method and organization theory to operate machine learning project
Requirement analysis, purpose design, man-hour estimation method for machine learning system
Collection and maintenance of efficient teacher data, pre-processing method
Framework, programming language, development environment for efficiently developing machine learning system
Architecture for designing machine learning system
Testing, verification, debugging, monitoring method of machine learning system
Platform and infrastructure supporting machine learning system, hardware
Jul 01
2018
Jul 02
2018
Draft paper submission deadline
Draft Paper Acceptance Notification
Final Paper Deadline
Registration deadline
Submit Comment