OpenBytes Events

Project OpenBytes global kickoff meeting was held on Dec 8th.

Inspiring and valuable information about the project scope and further details was discussed.
For those who didn't attend the meeting, meeting minutes is right here for you.

Kickoff Meeting Overview

Attendees

Edward Cui, Project Initiator, Founder and CEO of Graviti

Olivia Wu, Executive Director of OpenBytes, COO of Graviti

Linji Xue, Algo Team Leader of Graviti

Scott Nicholas, VP of Project Information, The Linux Foundation

Agenda

11:00am - 11:25am EST | Meeting Opening and Welcome From MC

Project Introduction: mission, scope and main efforts

11:25am - 11:30am EST | How to Get Involved

Action plans, project communication plan and review process

11:30am - 11:45am EST | Data Standards Session from Linji Xue

Dataset formats & standards

11:45am - 12:00pm EST | Legal Session from Scott Nicholas

Dataset licensing

12:00pm - 12:15pm EST | Open mic & Ending

Forum for the community

11:00am - 11:20am EST (20min) ｜Project Introduction

The motivation of Project OpenBytes / Project Scope

View Slides

11:25am - 11:45am EST (20min) ｜Data Standards Session

Building Open Datasets Standards

View Slides

11:45am -12:00pm EST (15min) ｜Data Licensing Session

Community Data License Agreement /Permissive- Version 2.0

Q&A

Q1. Is there any guidelines to follow if I’d like to upload a dataset?

A. Edward Cui: We will publish the guidance documentation on Github for everyone to collaborate and contribute their knowledge to the guidance. The key here is where to host those data. Graviti is currently building a data platform to be the host of open datasets.

Q2. How would the schema be able to integrate with the machine learning framework and models?

A. Linji Xue: We are trying to solve the inconsistent data format problem. If most datasets follow the standard, it will be very easy to connect different datasets and different training codes. So, the key is we must justify the format of datasets. And if dataset format has been clearly defined then we can write the training code and integrate it with machine learning frameworks, according to this predefined format.

Q3. Data license schema is also important, do you plan to work on it?

A. Linji Xue: We need to focus more on researching about the existing language which can use to define data structures. Also we need to work on configuring pre-defined format and solving languages that are not designed for table structures. From this, I included house sharing, the center formats and how to put bounding boxes into the schema languages.

Inspire AI Innovation with

Open Datasets

Learn More

Subscribe to our latest progress,

news and events!

We use cookies to improve your experience and to help us understand how you use our site. Please refer to our cookie notice and privacy policy for more information regarding cookies and other third-party tracking that may be enabled.

Join our Community!

Find us on slack, email or LinkedIn.