Year: 2024
Type: conference-paper
Abstract: While human-centered approaches to Machine Learning investigate diverse human roles within the interaction loop, the concept of Interactive Machine Teaching (IMT) has proposed with an emphasis on leveraging the teaching abilities of humans to build machine learning systems. Yet, most systems and studies in this field are devoted to single users. This doctoral thesis focuses on collaborative interactive machine teaching, exploring how people can collectively structure the teaching process and understand their experience. The research introduces TeachTOK, a collaborative interactive machine learning application that enables groups of users to curate data and train a model together incrementally. The thesis presents an exploratory study in collaborative interactive machine teaching, highlighting different collaboration patterns in different teaching activities and how collaboration helps revise teaching strategies. Based on these findings, we expand the work with novel research questions for the subsequent user study to investigate the potential of collaboration in other forms of involving humans in the loop for improving the quality of ML models, specifically regarding bias, harmful behaviours, and fairness.
Source: HAL
Author: Mohammadzadeh, Behnoosh
Language: French
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
SDG: Quality education
Open Access status: green
OpenAlex ID: W4399991849
