The School of Computer Science is pleased to present…
Spatial Team Formation Using Graph Neural Networks
MSc Thesis Proposal by:
Karan Saxena
Date: Wednesday, July 5th, 2023
Time: 11:30 AM
Location: Essex Hall, Room 122
Reminders: 1. Two-part attendance mandatory (sign-in sheet, QR Code) 2. Arrive 5-10 minutes prior to event starting - LATECOMERS WILL NOT BE ADMITTED. Note that due to demand, if the room has reached capacity, even if you are "early" admission is not guaranteed. 3. Please be respectful of the presenter by NOT knocking on the door for admittance once the door has been closed whether the presentation has begun or not (If the room is at capacity, overflow is not permitted (ie. sitting on floors) as this is a violation of the Fire Safety code). 4. Be respectful of the decision of the advisor/host of the event if you are not given admittance. The School of Computer Science has numerous events occurring soon
Abstract:
Effective team formation is crucial for project success, influenced by skill diversity and geographic proximity. This paper proposes a novel approach for geo-spatial team formation that leverages graph neural network (gnn) embeddings to transfer knowledge from a heterogenous collaborative network among experts, with metapaths and lack thereof, followed by a neural-based multi-label classifier. Our method aims to optimize team composition by considering the interplay between skill compatibility and geographic cohesiveness. Specifically, we construct a heterogeneous graph representation whose nodes are experts, skills, and locations to capture the complex relationships between experts' skills and their geographic locations. We employ gnn to learn vector representations of experts, encoding both their skill profiles and geographic information using metapaths. We propose a multi-objective optimization to guide the team formation process. The objective is to maximize skill diversity while minimizing geographic dispersion, balancing effective collaboration and efficient communication. We utilize classification and information retrieval metrics to evaluate the accuracy of the recommended teams of experts concerning the required skills and geographical distribution. Experimental evaluations of our proposed method on a real-world dataset of patents and computer science publications against baseline methods demonstrated the effectiveness of our approach in forming diverse and geographically cohesive teams. The findings of this study contribute to the field of team formation by highlighting the benefits of incorporating gnn embeddings considering skill and location in tandem for the task of team formation.
Keywords: Team Formation, NLP, Graph Neural Networks
Thesis Committee:
Internal Reader: Dr. Jianguo Lu
External Reader: Dr. Mohammad Hassanzadeh
Advisor: Dr. Hossein Fani
MSc Thesis Proposal Announcement
