Thanks to the many ISSIP ambassadors who participated in person at Prof David Lee’s talk…
Please Join us at the ISSIP Workshop on “Workshop on AI Biases” Jan 8, 1 pm- 4 pm HST, @ HICSS 52
You are cordially invited to join the International Society of Service Innovation Professionals, ISSIP, in a Workshop on “AI Biases (From humans to algorithm to data, biases in AI)” on January 8, 2019, 1 pm – 4 pm, at the 52nd Annual Hawaii International Conference on System Sciences, Jan 08 – 11, 2019,Grand Wailea, Maui, Hawaii, USA.
With the increase of statistical machine learning models in real-world business applications, questions around undesirable biases that may get introduced into the predictions they make are also on the rise. Some recent academic studies have shown undesirable biases in identifying people of certain race, color from certain parts of the world. Undesirable biases in statistical machine learnt models get introduced due to various reasons. They may creep in via training data, or in the features that the model learns from or due to lack of proper goal setting for the model. Several approaches are being proposed and investigated to deal with these undesirable biases. Two schools of thought are emerging in dealing with these undesirable biases. One school proposes approaches to discover them after they have crept into the models and try to fix them. In this case, models can be machine learnt using statistical approaches. Another school of thought proposes techniques to build machine learning models that are inherently representative of the subject-matter-expert opinion, thereby eliminating the creeping-in of undesirable biases in the first place. These models need to be carefully hand-crafted with the help of tooling machinery.
In this workshop, we intend to debate the topic of ‘AI and Bias’ from various angles including how to prepare training data that is ‘fair’, establishing ‘fairness’ measurement metrics for building AI systems, the role of data preparers and data collection platforms in creating fair AI models, alternative techniques to detecting, measuring, assessing and fixing undesirable biases in AI models. At this workshop, people from academia and industry come together to present their work in this area and debate the merits of various techniques.
Tentative Agenda:
- Welcome & structure of the workshop
- Invited Speakers presenting their work related to the workshop topic.
- Panel: With all speakers. Moderated by Rama Akkiraju
- General Discussion
- Wrap up and action items
Tentative List of Invited Speakers:
(To be provided soon)
All HICSS attendees are invited to attend this symposium. If you would like to participate, sign up when you register for the HICSS Conference. If you already registered for the HICSS 2019, but did not indicate on your registration form that you would like to attend this symposium, please let us know via email (issip@issip.org). We can update your registration.
We look forward to seeing you at the conference and the symposium!
Workshop Chairs – Rama Akkiraju (Distinguished Engineer, IBM; President, ISSIP), Ammar Rayes (Distinguished Engineer, Cisco, Board of Director Member, Past President, ISSIP
P)
About The International Society of Service Innovation Professionals – ISSIP (pronounced iZip), is a 501 (C) (3) nonprofit professional association, where leaders and experts from companies including Cisco, IBM, United Healthcare Group, Accenture, Google and more are partnering with world-leading foundations, including Japan Science and Technology Agency, JST, National Science Foundation, NSF, as well as hundreds of thoughts leaders from universities worldwide to advance Service Innovation for our interconnected world. Our purpose is to help institutions and individuals to grow and be successful in our global service economy. ISSIP is a global organization with over 1000 members representing a number of major companies, research centers, foundations, more than 200 universities, and 42 countries. For more information about ISSIP, please visit www.issip.org. ISSIP has been a proud sponsor of HICSS for the last 5 years.