Dear ISCA members,

I would like to take this opportunity to thank you for electing me to the ISCA board, where I have assumed the 'communication' position. Facilitating access to our community was at the center of my campaign to join the board and is the motivation behind some of the ideas that I would like to present to you for discussion.

In the following, I would like to share with you a vision that is realizable with the technical infrastructure that we have today. As you read you will see that technology is not the challenge here. The real question is: Are we ready to mobilize and organize a community around an enabling platform that is self-sustainable?

In addition to the website that holds static information and is updated on a regular basis, it would be interesting to create a community that initiates, nourishes and produces researchers in our field, regardless of geographic location. It is a general trend that has been supported by mobile computing and the MOOC movement in recent years. The proposed platform serves as a pipeline to get anyone up to speed on top-notch research relying 'only' on dedication, a computer and the internet. It would guide a student in a sequential manner from basic coursework and mathematical foundations to state-of-the-art papers and research problems and solutions. The goal is to enable participate at the publication level, granting access to datasets, computational power and sample code and scripts so that anyone can become a player.

Vision 1: ISCA MOOC

Such a MOOC platform should provide a path from beginner to expert:

An ordered list of courses, representing a recommended curriculum accepted by the speech community as a guideline for building a sound base in Mathematics, linguistics, signal processing, programming, writing skills (etc.) before diving further into the field of speech science. These courses can be on any platform, such as EDX, Coursera, etc.

Advanced classes on specific areas of interest that prepare the student for research or applications.

Finally, at the graduate level, the platform could provide the following:

Research problem statements,

datasets and

computational resources to be tapped online, supplied by labs.

While (1) is potentially a simple organized list of links to courses that can be found online in one of the MOOCs or other formats, numbers (2) and (3) are equally important in this pipeline to produce cutting edge researchers but are more difficult to realize. Creating a MOOC is labor intensive. Not just in production but also in maintenance. A forum does not run itself and needs tutors to maintain it. These points can only be achieved through crowd sourcing in a vibrant community of industry sponsors, interested graduate students and senior researchers active in the community.

How can such a community be bootstrapped?

For the advanced classes:

ISCA provides an 'empty' MOOC platform such as openEDX

We could provide a simple start with establishing one course per distinguished lecturer

Any training workshops that are sponsored by ISCA could be required to put their

lectures and materials online to add additional courses

Interspeech tutorials could also be placed on this platform for future tutorials

Finally, people may donate additional courses on the platform

Vision 2: The Research Platform:

In order to foster communication about research problems in a transparent manner, between students, labs, data-suppliers and computational resources around the world, we need something different from an asynchronous self-study MOOC. We need competitions that are put up and researchers, including new-bees, to join in and prove themselves to future employers. Such a platform already exists and is called https://www.kaggle.com/: 'Kaggle is the world's largest community of data scientists. They compete with each other to solve complex data science problems, and the top competitors are invited to work on the most interesting and sensitive business problems from some of the world’s biggest companies through Masters competitions.' If you don’t know this site, take a look at it. We could spawn ideas off that platform and create our own or use the same one and build it up for speech science application competitions by creating a sort of classroom for ISCA. Labs could supply datasets, machines for students with or without resources to run their algorithms on donated computational time on remote machines. A central platform could be a simple list of labs that are interested in connecting students to their resources. Later, this platform could develop into something more dynamic and open.

The above ideas are theoretical and pose a number of logistical questions. However, the realization of the vision can grow by taking small first steps. I would be interested to start a discussion on ideas that you might have on how to proceed here.

The persona to have in mind as this vision takes shape is a student, interested in the topic, who has no local advisor or advanced course work at a local institution, nor the computational power to participate at the publishable research level. The goal is to get this person involved in the community. Eventually, they might publish and attend a conference remotely, but the brain, drive and creativity of such a person would not be lost to us simply because that person did not have the luck to walk into a research lab like some of us have done, lucky enough to have had the mentorship to end up where we are today.

Log in and join the discussion on:

http://communicationisca.freeforums.net (a temporary home for a discussion on this topic)

Happy New Year (for those who celebrate end of December),

Kay Berkling