This study investigates pronominal anaphora in a Turkish and English translated TED corpus, the TED-MDB (Zeyrek et al., 2020), and provides a heuristic-based resolution mechanism for each language. The corpus comprises 364 English-Turkish sentences aligned. Research has two phases. The annotator annotated the data in the first phase. In the second phase, the knowledge poor method of Mitkov (1998) was tested on the Turkish and English annotated corpuses independently. TED presentations can identify pronominal anaphora with an F1-score of 0.61 in English and 0.63 in Turkish.
Date: 27.01.2023 / 15:00 Place: A-212
In this thesis study, we focus on the construction of an effective network architecture, for which we propose an architecture generation framework and show how it can be used to create an effective terrain classification model. Additionally, we also observe that land cover training data sets on HSI and LiDAR tend to come short in providing training examples with shadow effects. To address this limitation, we additionally propose a generative adversarial network(GAN) driven statistical data augmentation technique that generates synthetic training examples and show its effectiveness in our experimental results.
Date: 26.01.2023 / 14:30 Place: A-108
Among various Agile Software Development methods, Scrum is one of the most widely adopted one. The Scrum Guide clearly describes the Scrum events, artifacts, and roles. However, due to various project characteristics such as team size, team distribution, project domain, technology and requirement stability levels, Scrum practices need to be tailored. In this thesis, we analyze the various adaptations and tailoring choices of companies using Scrum. By incorporating evidence from literature and a survey study that is conducted among participants who use Scrum in their organizations, the impact of Scrum tailoring on technical debt will be analyzed.
Date: 24.01.2023 / 15:00 Place: B-116
This research evaluates the convergence of High-Performance Computing (HPC), Big Data, Artificial Intelligence (AI), and Cloud Computing technologies using bibliometric analysis, including performance and network analysis. The results reveal a rapidly growing literature with a significant increase in research activities in recent years, identifying key trends and patterns in the literature, including top published authors, most productive institutions, cited articles, and influential publications. This thesis provides valuable insights by identifying the bibliometric trends across the concept of technological convergence of HPC-Big Data-AI-Cloud Computing technologies, which is important for both academia and industry.
Date: 24.01.2023 / 11:00 Place: A-212
Catastrophic forgetting is common in the connectionist models while learning from a sequence of tasks. This study aims to investigate different continual learning methods on face analysis tasks involving age estimation, gender recognition, emotion recognition, and face recognition. We analyze face analysis in two stages, which is also very common in Artificial Neural Networks: face detection and face attributes analysis. Firstly, experiments for learning face detection and facial landmark detection are conducted by studying multitask learning. Secondly, some continual learning methods inspired by biological systems are leveraged to overcome catastrophic interference in artificial models.
Date: 26.01.2023 / 09:30 Place: B-116