Machine learning has become an important topic for students outside of computer science and statistics to understand because of its useful applications and its societal impacts. This project explores making machine learning more accessible to those without any computer science or statistics knowledge. The first component of this project interviewed instructors who have taught machine learning to non-computer science major students. I'm now researching what algorithmic literacy means in today's context.
PRESENTATION @ 4S
I will be presenting my work on algorithmic literacy at the Reframing Algorithmic Accountability in Practice session in September.
PAPER in TOCE JOURNAL
What Is Hard About Teaching Machine Learning to Non-Majors? Insights From Classifying Instructors' Learning Goals. Elisabeth Sulmont, Elizabeth Patitsas, and Jeremy R. Cooperstock. ACM Transactions on Computing Education. Accepted, to appear.
PAPER @ SIGCSE 2019
Can You Teach Me To Machine Learn? An Exploration of Pedagogical Content Knowledge for Teaching Machine Learning to Non-Majors. Elisabeth Sulmont, Elizabeth Patitsas, and Jeremy R. Cooperstock. In Technical Symposium on Computer Science Education, Minneapolis, USA, March 2019. ACM (slides ↗)
Sept 2017 - Dec 2018
Today's smartphone notification systems are incapable of determining whether a notification has been successfully perceived without explicit interaction from the user. SweatSponse is a feedback loop using skin conductance responses (SCR) to infer the perception of smartphone notifications just after their presentation. Results from a laboratory study confirm, for the first time, that both vibrotactile and auditory smartphone notifications induce SCR, that the induced responses are different from that of arbitrary stimuli, and that they can be employed to better predict perception of smartphone notifications after their presentation using wearable sensors. I was involved in developing the experiment design and data pipeline.
PAPER @ CHI 2019 WITH HONOURABLE MENTION
Detecting Perception of Smartphone Notifications Using Skin Conductance Responses. Pascal E. Fortin, Elisabeth Sulmont, and Jeremy R. Cooperstock. In Human Factors in Computing Systems (CHI), Glasgow, UK, May 2019. To appear.
POSTER @ UIST 2018
SweatSponse: Closing the Loop on Notification Delivery Using Skin Conductance Responses. Pascal E. Fortin, Elisabeth Sulmont, and Jeremy R. Cooperstock. In User Interface Software and Technology, Berlin, Germany, October 2018. ACM
Sept 2017 - April 2018
In collaboration with Quebec's Ministère de la Sécurité publique (MSP), we developed a Twitter monitoring app to identify tweets of interest related to natural disasters and emergency situations in Quebec (e.g., floodings, blizzards) using machine learning. In a team of two, I was responsible for the machine learning component, the data pipeline, and project management. This app is used daily by MSP personnel.
May 2017 - Jan 2018
As a summer research assistant, I designed and developed an alternate Facebook interface for older adults through a browser extension (see report ↗). Additionally, I helped an on-going lab project to evaluate a digital paper-based email system for older adults. From this work, I was appointed an AGE-WELL Highly Qualified Personnel.
POSTER @ UIST 2017
A Digital Pen and Paper Email System for Older Adults. Taciana Pontual Falcão, Xiaofeng Yong, Elisabeth Sulmont, Robert Douglas Ferguson, and Karyn Moffatt. In User Interface Software and Technology, Quebec City, Canada, October 2017. ACM
DEMO @ AGE-WELL CONFERENCE 2017
"SimpleFB: A Simpler Facebook for Older Adults" presented by Elisabeth Sulmont and Taciana Pontual Falcão at AGE-WELL Annual Conference Winnipeg, Manitoba 2017.
POSTER @ AGE-WELL CONFERENCE 2017
"Accessible Social Engagement for Older Adults" presented by Elisabeth Sulmont at AGE-WELL Annual Conference Winnipeg, Manitoba 2017.