A medical Visual Question Answering (VQA) system can provide meaningful references for both doctors and patients during the treatment process. Different from normal images, a learning setting with medical images is more challenging due limited amounts of data, class-imbalance and the presence of label noise for diagnosis tasks. Moreover, little attention is paid to how the images and meta-data is …
Supervisors: Amelia Jiménez-Sánchez
Semester: Fall 2023
Tags: medical imaging, deep learning, machine learning, transfer learning, meta-learning

Machine learning models, especially larger models that are used in for example image or text datasets, can be expensive to train. During development models are usually trained multiple times for example to optimize hyperparameters, which can result in a large carbon footprint. This project specifically focuses specifically on medical data. There are some recent efforts, for example by Selvan et …
Supervisors: Veronika Cheplygina
Semester: Fall 2023
Tags: machine learning, medical imaging, data analysis, resource consumption

There have been several situations where machine learning classifiers, trained to diagnose a particular disease (for example, lung cancer from chest x-rays), overfit on hidden features within the data. Examples include gridlines, surgical markers or evidence of treatment or text present in the images (see references for examples). This causes the classifier to fail on other type of images. …
Supervisors: Veronika Cheplygina, Amelia Jiménez-Sánchez
Semester: Fall 2024
Tags: machine learning, data science, medical imaging

It is common to process data to clean it, filter it, restructure it, get metadata out of it, etc. before feeding the data into a data analysis or machine learning pipeline. There are many tools and libraries out there to aide with this process with different strengths and functionality (DALI, RAPIDS, HoloClean, DAPHNE, DuckDB, etc.). In this project, we would like to analyze pros/cons of some of …
Supervisors: Pınar Tözün
Semester: Fall 2022
Tags: data preprocessing libraries, heterogeneous hardware, machine learning

State-of-the-art machine learning models are known to be compute- and power-hungry. On the other hand, modern servers come equipped with really powerful CPU-GPU co-processors. Not all machine learning models are able to use all the available hardware resources on such servers. Workload collocation is a mechanism to increase hardware utilization when a single workload is not able to utilize all the …
Supervisors: Pınar Tözün
Semester: Fall 2022
Tags: benchmarking, workload collocation, machine learning

Today, there are many compute- and memory-hungry data-intensive workloads from big data analytics applications to deep learning. These workloads increasingly run on shared hardware resources, which requires building hardware resource managers that can both serve the needs of workloads and utilize hardware well. Predicting the resource utilization of applications can aid such resource managers …
Supervisors: Pınar Tözün, Ehsan Yousefzadeh-Asl-Miandoab
Semester: Fall 2022
Tags: benchmarking, hardware resource consumption estimation, machine learning

The Danish Student Cubesat Program is an inter university collaboration that will launch 3 cubesats into Low Earth Orbit over the next 4 years. The satellites will be designed, operated, programmed and built by students and the project offers an opportunity for Master’s students to take part in a live satellite project. ITU is partnering with Aarhus University on DISCOSAT2 which will be an …
Supervisors: Sebastian Büttrich, Julian Priest
Semester: Fall 2021
Tags: Satellite, Cubesat, Image processing, Machine Learning, edge, constrained computing

Machine learning is used extensively in different applications, including medical imaging and natural language processing. As different types of data are involved, it is reasonable to assume that different methods are needed for each application. However, there are also opportunities in translating a method successful in one application, to the other application where it is not widely used. The …
Supervisors: Veronika Cheplygina
Semester: Fall 2021
Tags: machine learning, natural language processing, medical imaging, literature review

Machine learning algorithms for skin lesion classification typically learn from images which have been labeled as malignant (for example, melanoma) or not. Such tasks can still suffer from overfitting due to limited dataset size. In other computer vision tasks, crowdsourcing labels has been effective, but the average person typically does not have the background to classify skin lesions. However, …
Supervisors: Veronika Cheplygina
Semester: Fall 2021
Tags: machine learning, medical imaging, crowdsourcing, similarity

ITU is a partner in the Danish Student Cubesat Program, DISCO which will launch a series of small satellites into orbit, starting with DISCO 1 in 2023 and followed by DISCO2 in 2024. As part of this project ITU is installing a satellite ground station with a range of antenna rotators on the roof of Rued Langaards Vej building and the equipment has been purchased. The ground station will track the …
Supervisors: Julian Priest
Semester: archive
Tags: Satellite, Image processing, Edge, Constrained Computing, Networks, Machine Learning, Embeded, Software Defined Radio