Anees Kazi | Graph Convolutional Network for Disease Prediction

Virtual London Machine Learning Meetup - 27.10.2021 @ 18:30

We would like to invite you to our next Virtual Machine Learning Meetup.

Agenda:

• 18:25: Virtual doors open
• 18:30: Talk
• 19:10: Q&A session
• 19:30: Close

Sponsors
https://evolution.ai/ : Machines that Read - Intelligent data extraction from corporate and financial documents.

• Title: Graph Convolutional Network for Disease Prediction with Imbalanced Data and multi-modal data analysis (Anees Kazi is a senior research scientist at the chair of Computer Aided Medical Procedure and Augmented Reality (CAMPAR) at Technical University of Munich)

Abstract: Recently, Graph Convolutional Networks (GCNs) have particularly been studied in the field of disease prediction which is a well-known classification problem. Class imbalance is in general a problem in medical datasets. When the class imbalance is present in the data, the existing graph-based classifiers tend to be biased towards the major class(es). Meanwhile, the correct diagnosis of the rare true-positive cases among all the patients is vital. In conventional methods, such imbalance is tackled by assigning appropriate weights to classes in the loss function; however, this solution is still dependent on the relative values of weights, sensitive to outliers and, in some cases, biased towards the minor class(es). In this talk, we will see our method ‘Re-weighted Adversarial Graph Convolutional Network (RA-GCN)’ to enhance the performance of the graph-based classifier and prevent it from emphasizing the samples of any particular class.

The second topic that I will touch on is dealing with multi-modal data.
The increased amount of multi-modal medical data has opened the opportunities to simultaneously process various modalities, such as imaging and non-imaging data to gain a comprehensive insight into the disease prediction domain. Graph Convolutional Networks (GCNs) provide novel semi-supervised approaches for integrating heterogeneous modalities while investigating the patients’ associations for disease prediction. However, when the meta-data used for graph construction is not available at inference time (e.g., coming from a distinct population), the conventional methods exhibit poor performance. To address this issue, we propose a novel semi-supervised approach named GKD based on knowledge distillation. In the second part of my talk, we will look into the details of the same. Please feel free to read the following two papers for more fruitful discussions.

RA-GCN: Graph Convolutional Network for Disease Prediction Problems with Imbalanced Data (https://arxiv.org/pdf/2103.00221.pdf)

GKD: Semi-supervised Graph Knowledge Distillation for Graph-Independent Inference (https://arxiv.org/abs/2104.03597)

Bio: Anees Kazi is currently a senior research scientist at the chair of Computer Aided Medical Procedure and Augmented Reality (CAMPAR) at Technical University of Munich, mainly focusing on multi-modal data analysis using Graph-based AI models. She is the team leader of the Graph Deep Learning group at CAMPAR and recently finished her PhD also at CAMPAR on 'Graph Deep Learning for Healthcare Applications'. Anees has worked towards providing solutions to brain-related disease diagnosis problems by solving technical challenges such as dealing with multiple graph scenarios, graph structure heterogeneity, data imbalance and cross model retrieval. In 2019, Anees was awarded TUM-ICL incentive funding to collaborate with Prof. Michael Bronstein at Imperial College London. The team focused on the challenging problem of graph structure learning.

The discussion will be facilitated by Fabrizio Frasca. Fabrizio is currently a Machine Learning researcher at Twitter Cortex and a PhD candidate at Imperial College London, under the supervision of Prof. Michael Bronstein. His research has been focusing on graph representation learning and network science, encompassing bioinformatics and computational social sciences.