talk-data.com

This project implements a 2024 paper introducing innovative methods for visualization and explainability in graph tasks like node classification and link predication. Performance will be evaluated against baselines on existing and novel datasets, with a focus on explainability in graph tasks.

The aim of the project is to develop an application that analyses hotel reviews and provides valuable information that allows customers to make more informed decisions based on the past experiences of others.

This project develops a prototype application for analyzing logical text structures to generate reasoning-based multiple-choice questions for TMS preparation. Using Graph Neural Networks (GNNs), it captures logical connections within text to create questions that enhance logical reasoning skills. The effectiveness of GNN-based generation will be evaluated by comparing transformer models trained on questions with and without GNN analysis.

This project aims to develop a system for video body language annotation in small social contexts. The video could be a part of a podcast, a movie or a conference presentation. Analyzing gestures can provide insights for a large, but subtle part of day-to-day communication.

The autonomous security robot on 4 wheels is designed to provide security monitoring and threat detection. The robot can map its environment and move around autonomously, patrol specific areas or objects, detect and respond to threats. The robot monitors its surroundings for potential threats, such as intruders or suspicious items. It can be set up to automatically guard a certain area, protect a person, or watch over an object