SAVE THE DATE FOR DEEP 2018!
Oct. 10-11, 2018.
OCAD University, 100 McCaul St., Toronto, Ontario, CANADA
We’ll be posting more details here soon!
This year at DEEP 2017:
Learning to Support and Understand Diversity & Complexity
“Let’s not teach our machines our stereotypes & presumptions.”
More and more decisions are being delegated to machines, whether to authorize access, recognize illnesses, predict risks, determine loan and credit worthiness, academic potential, future employment performance or a multitude of other functions. This trend has the potential to amplify existing inequities.
Whether we welcome or fear machine intelligence, it is important that we attend to what we teach machines. Do our machines understand and serve individuals that are different or fail to recognize and ignore anyone that does not conform to the model of an average human?
In understanding how bias against outliers is detrimental to machine-learning we gain insights into how our human stereotypes are embedded in our traditional research and public policies, and damage our society.
Topics at DEEP 2017
Discussions will centre around four areas:
- Government Accessible ICT Procurement Policies—recommendations for federal accessible procurement policies will be shared and discussed. Participants will be able to provide input and respond to the final recommendations.
- Cognitive Access & Learning Differences – an area that is not well developed in accessibility legislation and services is the area of cognitive access, yet it is the most prevalent experience of disability. There is a call for Canada to take bold steps in this issue. Participants will help to establish a bold plan.
- Machine Learning—participants will delve into the challenge of teaching machines and automated systems to recognize and adapt to diversity rather than rely on flawed models of “average humans” that perpetuate social, political, and economic biases.
- Co-operative Platforms—participants will step through development of a cooperative platform and consider social implications, privacy implications, inclusion aspects and potential barriers.
Each group will be working toward concrete policies, inclusion models or recommendations for inclusion that will be published and shared openly under Creative Commons licensing.
In the tradition begun in 2012, this event to plan meaningful change will bring together a diversity of fresh perspectives, new ideas, together with seasoned experience, and insight gained from failures and successes. DEEP participants know that digital inclusion is necessary and urgent. We don’t need to persuade each other of the “why”. This allows us to focus our time together during the think tank on the “how”? We will spend our collective time and talents to design strategies, plans, collaborative projects, cross-sector initiatives and new disruptive ways to foster more inclusive prosperity, privacy, and life-long learning.
Events at DEEP 2017
In cooperation with BIG IDeA
“Start Your Machine Learning Engines and Race to the Edge”
Race to the EDGE
49 McCaul Street | Oct.10th at 10am to 6pm
register now for Race to the Edge
BIG IDeA and DEEP 2017 are looking for teams to help prepare the racecourse for our AI “Race to the Edge”.
Our society is about to hand over many decisions to machines and we need to make sure they understand and can serve human diversity. Artificial Intelligence (AI) systems currently have difficulty recognizing and understanding outlying and edge scenarios or people.
This is bad for people because we are relying on AI systems more and more because soon systems won’t be able to deal with the unexpected, transfer knowledge to new situations, address weak signals, predict changes or deal with our current diverse society. – Jutta Treviranus
This is even worse for people who are different from the models the machines use.
People who are unusual or different:
- will be filtered out of competitive job applications
- won’t be offered limited school placements
- won’t get credit, loans or insurance
- won’t be understood by voice recognition systems
- won’t be recognized by identity systems
- may be flagged as a threat
- may be run over by automated vehicles
- won’t be properly diagnosed by telehealth systems
We need you to prepare the tests for the ultimate challenge.
We are looking for diverse teams that will prepare edge scenarios that might stump the AI systems. Ideally each team should include someone that has experienced being excluded because of difference and someone with a very basic knowledge of machine learning systems and data.
Machine learning and AI have the power to transform our lives and to empower every person and organization around the world. It’s crucial that as these technologies develop they are designed to include people with disabilities.
Jenny Lay-Flurrie, Chief Accessibility Officer, Microsoft.
With machine leaning and cognitive technology, we have the opportunity to provide personalized services to people with all abilities.
Ruoyi Zhou, Ph.D., Director, Accessibility Research, IBM Research