Examensarbeten för masterexamen
Länka till denna samling:
Browse
Senast publicerade
Visar 1 - 5 av 368
- PostDecentralized Deep Learning under Distributed Concept Drift: A Novel Approach to Dealing with Changes in Data Distributions Over Clients and Over Time(2023) Klefbom, Emilie; Örtenberg Toftås, Marcus; Chalmers tekniska högskola / Institutionen för matematiska vetenskaper; Modin, Klas; Dubhashi, DevdattIn decentralized deep learning, clients train local models in a peer-to-peer fashion by sharing model parameters, rather than data. This allows collective model training in cases where data may be sensitive or for other reasons unable to be transferred. In this setting, variations in data distributions across clients have been extensively studied, however, variations over time have received no attention. This project proposes a solution to address decentralized learning where the data distributions vary both across clients and over time. We propose a novel algorithm that can adapt to the evolving concepts in the network without any prior knowledge or estimation of the number of concepts. Evaluation of the algorithm is done using standard benchmarks adapted to the temporal setting, where it outperforms previous methods for decentralized learning.
- PostMachine learning techniques for metastatic tumor prediction(2023) Dahlén , Filip; Chalmers tekniska högskola / Institutionen för matematiska vetenskaper; Beilina, Larisa; Beilina, LarisaAccurate prediction of tumor metastasis is crucial for cancer prognosis and treatment. This study explores the application of artificial intelligence techniques, specif ically neural networks and Principal Component Analysis (PCA), for predicting tumor metastasis based on histopathological images. The aim is to identify an effective approach that balances accuracy and computational efficiency. The study begins with a comparative analysis between PCA and a Convolutional Neural Network (CNN) for image classification, using the ISIC dataset and MNIST dataset. To further investigate potential benefits, a combined model using PCA and CNN is developed. PCA is employed as a dimensionality reduction technique to enhance computational efficiency. The combined model demonstrates promising results, reducing computational time while maintaining accuracy. In the second part of the study, metastatic prediction is addressed using Whole Slide Images (WSI) obtained from Sahlgrenska University Hospital. Preprocessing involves dividing WSIs into smaller tiles and extracting relevant features such as breslow depth and metastatic status. Two models are utilized: a modified ResNet18 and a shallower network (ConvNet2/ConvNet3) with varying inputs. Training and evaluation are performed using a 4-fold cross-validation approach and mini-batches. It was observed that the PCA algorithm performed similar to the neural networks but was outperformed in terms of computational time. The combined model showcased the potential in reducing computational time while maintaining accuracy. The combined model showcased the potential in reducing computational time while maintaining accuracy when the there are a large amount of available data with high complexity. Lastly, the study showed that when the data are limited, the best approach was to utilize a neural network based on image data and clinical data to enhance metastatic prediction. Overall, this thesis emphasizes the potential of neural networks based on image and clinical data for enhancing metastatic prediction, providing valuable insights for improved cancer prognosis and treatment decision-making.
- PostSimulation and analysis of spinodal decomposition using the Allen-Cahn equation(2023) West, William; Chalmers tekniska högskola / Institutionen för matematiska vetenskaper; Gebäck, Tobias; Gebäck, TobiasSpinodal decomposition is a type of phase separation that can occur for certain compositions of mixtures. If one of the phases is then resolved, the result is a labyrinthine structure that can advantageously be used to transport material, such as medicine through tablets. Simulations of spinodal decomposition based on the Allen-Cahn equation were run with a variety of parameters to gather time-based data. Methods were then developed to analyse characteristics from the data, both in 2D and 3D. The results were compared to those from previous physical experiments, with some accuracy. Finally, the effects of boundary conditions on spatial variations and phase patterns were studied.
- PostAdaptive Radar Illuminations with Deep Reinforcement Learning: Illumination Scheduling for Long Range Surveillance Radar with the use of Proximal Policy Optimization(2023) Sandelius, Samuel; Ekelund Karlsson, Albin; Chalmers tekniska högskola / Institutionen för matematiska vetenskaper; Helgesson, Peter; Andersson, AdamA modern radar antenna can direct its energy electronically without inertia or the need for mechanically steering. This opens up several degrees of freedom such as transmission direction and illumination time, and thus also the potential to optimise operation in real-time. Long range surveillance radars solve the trade-off between searching for new targets and tracking known targets. This optimisation is often rule-based. In recent years, Reinforcement Learning (RL) Algorithms have been able to efficiently solve increasingly difficult tasks, such as mastering game strategies or solving complex control tasks. In this thesis we show that reinforcement learning can outperform such rule-based approaches for a simulated radar.
- PostCritical Event Prediction in Logs at Customer Network(2022) Hajizada, Elmar; Chalmers tekniska högskola / Institutionen för matematiska vetenskaper; Jonasson, Johan; Jonasson, Johan; Akrami, RozitaImplementing effective maintenance prognosis for Radio units at Ericsson can result in a number of benefits, including better system safety, improved operational reliability, longer equipment lifespan, and lower maintenance costs. Preventive investigations and repairs on the hardware and software level can be done to avoid the radio unit from failing by forecasting whether or not the radio unit will have an alarm in the near future. The goal of this thesis was to use multiple logs taken from a radio unit to predict whether an alarm would occur in the next one to nine days. The log file contents have been divided into chunks using different approaches like expanding window, independent chunks and time interval chunks where each chunk labeled according to timestamp of the alarm. Ericsson has used a combination of verdicts (features that are defined by subject matter experts) to extract the best features from the log files. This rule-based approach is inefficient since it requires modification of the script using expert knowledge when there is a change in the design of the hardware. The purpose of this thesis project was achieved using data-driven NLP approaches including log parsers and word embeddings. An independent chunks approach with Drain log parser using concatenated bag-of-words representations for each log file fitted on the Xgboost model outperformed other combination of log parsers and word embeddings. LSTM model was used with 1 day interval chunks to see if the complex sequential model can achieve a sufficient score. Experiments using complex sequential model, such as the LSTM many-to-many model with doc2vec embedding, have shown shown that they can predict alerts before they occur. All the tested models were evaluated using cross-validation. The Xgboost model with the independent chunks approach using Drain log parser and BOW embedding achieved an average F1-score of 0.873, LSTM model with time interval chunks approach using doc2vec embedding achieved average 0.853 F1-score across shifting time periods from one to nine days.