Exploring Model Fusion with Optimal Transport on Transformers

When it comes to the field of artificial intelligence, the possibilities seem endless. With numerous models developed to achieve high-level performance in various tasks such as object detection, image segmentation, and language translation, the question arises of how we can utilize the knowledge gained from each individual model to improve overall performance. That’s where model fusion comes in. By combining the knowledge of different models, we hope to create a child network that outperforms its parent networks.

Gently

Gently was a project that my team and I built during a 36-hour hackathon called StartHack. We chose Supercell’s case, which was about reducing online vulgarity in their games. The main idea of the project was to promote positive in-game behavior and discourage negative behavior through a behavior score system that we implemented.

VR Horror Maze

VR Horror Maze game is built using Unity and designed for the Oculus Headset. The project took several months to complete, from initial concept to final release. As part of the project, we created a randomly generated maze filled with scary obstacles and a variety of monsters chasing the player.

SWARM

SWARM is an ongoing project that focuses on building a robust interface to control a swarm of drones. Our team has been working on this project since January 2022, and we have made significant progress in developing different interfaces for controlling the drones. The interfaces we have developed include PC, BCI, and Hololens 2. This ensures that users can control the drones using their preferred method, depending on their level of experience and the task at hand.

Low-Power Capacitive Sensing for Biosignal Analyses

Low-Power Capacitive Sensing for Biosignal Analyses was a semester thesis that I worked on from September 2021 to January 2022. The project involved using a novel sensor called the QVAR sensor to measure biosignals such as ECG and EEG.

Guided AR Tour

Guided AR Tour was a project that I worked on in August 2021. The project involved developing an augmented reality (AR) tour app for the opening event of SPH.
The app consisted of 17 different markers, each of which provided a different instruction about SPH. Using Unity and C#, I built the app to be an Android app. I also used Vuforia to track the markers and ensure accurate placement of the AR content.

Comparing Performances of Different NLP Models on Quote Generation

Comparing Performances of Different NLP Models on Quote Generation” is a project that I worked on during the Neural Systems lecture. The goal of the project was to explore different methods for generating quotes and measure their performances in terms of accuracy and fluency.

UniBoost

UniBoost was a peer-to-peer tutoring platform designed specifically for students at ETH Zurich. The platform was created to help students offer private tutoring sessions for various lectures across different departments, giving them the flexibility to set their own pace, availabilities, and price per hour. This allowed students to earn extra income by leveraging their own knowledge and skills, while also helping fellow students who might be struggling with certain subjects.

Nao Robots for RoboCup Soccer Competition

Nao Robots for RoboCup Soccer Competition was a project that we worked on between September 2020 and June 2021. The project aimed to program Nao robots to play soccer in the RoboCup competition.

Holographic Hands for Mixed Reality Instructions

Holographic Hands for Mixed Reality Instructions was a project that I worked on between September 2020 and January 2021. The project aimed to create a mixed reality application that could provide instructional guidance to users in various settings.