Computation Behavioral Modeling (CBM) Research Lab
Current focus of this lab is on applications of ML on digital crowd-sensing data, along with its biases and societal impact. For more information read our recent publications and project descriptions.
If you are interested to be involved in our lab please contact me or join our Mailing List
Crowd-Sensing
I am interested in understanding behavioral patterns through large scale digital data from the crowds. Below are some of my current existing projects.
Urban Community Detection using Federated Deep Embedded Clustering
Improving the digital representation of our cities through federated crowd-sourcing tasks
In order to learn a fair and inclusive representation of the world, algorithms require access to unbiased and diverse datasets. What if such dataset could be dynamically created by people without the need for centralized data and privacy concerns related to sharing? Relying on Federated Learning Paradigm and Human-Centric AI we discover the possibilities of breaking away from centralized data repositories and rely directly on citizens to collaboratively train the algorithms that have societal impacts on them.
Deep Embedded Clustering of un-lablled acoustic data: Heartbeat Sound
Human Centric AI
In this line of work my focus is understanding how ML models impact society and policy makers decisions. More specifically I am interested in dimensions of privacy and fairness. Below are some of my current projects.
Biases in visitation counts for park and recreational planning (see our ICWSM 2021).
In this work we explore the algorithmic biases and fairness analysis of who is counted and who goes uncounted if we rely on rich social media images to detect usage of public lands and parks.
Privacy Framework for CDR data sharing (T-Mobile Grant)
See my publications with Dr Haddadi.
