MSc Thesis Defense Announcement of Sagar Kaw:"Learning Networks with Attention Layers for Team Recommendation"

 

Date: Thursday April 20, 2023

Time:  12:00pm – 1:30pm

Location: Essex Hall, Room 105

 

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.

 

The Team Formation Problem aims to identify a group of experts who possess the required skills to complete a common goal. Graph-based approaches have been commonly used to solve this problem, but recently, researchers have started exploring this problem from the perspective of social information retrieval and applying neural architectures to recommend teams of experts. However, the learning process of these architectures is faced with several challenges. This includes the inability to handle network modifications after the training process is over as well as the time complexity of the learning process is high, which is proportional to the size of the network. In this study, we propose a new framework called “LANT - Leveraging Graph Attention Network for Team formation” which leverages graph neural networks and variational inference to address the challenges faced by existing approaches. The proposed framework utilizes transfer learning and neural team recommendation, with self-supervised learning of node embeddings achieved using Deep Graph Infomax with Graph Attention Networks as an encoder. We demonstrate empirically how LANT effectively addresses the challenges faced by existing approaches and outperforms state-of-the-art methods on large scale real world datasets. The proposed framework provides an efficient and scalable solution to team formation problems and can be applied in various fields where expert teams are required to achieve a common goal.

 

Keywords: Team Recommendation, Transfer Learning, Graph Neural Networks

 

Internal Reader: Dr. Pooya Moradian Zadeh

External Reader: Dr. Bharat Maheshwari      

Advisor: Dr. Ziad Kobti

Co-Advisor: Dr. Kalyani Selvarajah

Chair: Dr. Adel Abusitta