Behçet Uğur Töreyin

The promising performance increase offered by quantum computing has led to the idea of applying it to neural networks. Studies in this regard can be divided into two main categories: simulating quantum neural networks with the standard quantum circuit model, and implementing them based on hardware. However, the ability to capture the non-linear behavior in neural networks using a computation process that usually involves linear quantum mechanics principles remains a major challenge in both categories. In this study, a hardware-efficient quantum neural network operating as an open quantum system is proposed, which presents non-linear behaviour. The model's compatibility with learning processes is tested through the obtained analytical results. In other words, we show that this dissipative model based on repeated interactions, which allows for easy parametrization of input quantum information, exhibits differentiable, non-linear activation functions.

The proliferation of non-invasive neuroimaging datasets acquired from different modalities has driven advancements in machine learning models for diagnosing brain disorders. While prior studies have primarily focused on feature engineering and model architecture improvements, they often neglect the impact of low-quality samples in training datasets, which can significantly hinder diagnostic performance. To address this, we introduce a novel sample selection framework, Influence-based Detection of Opponent Samples (IDOS), which estimates sample quality using influences approximated by the change in loss relative to a reference point. We utilized Graph Convolutional Networks (GCN) and Differentiable Graph Pooling Modules (DIFFPOOL) in IDOS using an architecture that leverages whole-brain graphs. Excluding low-quality samples identified by IDOS significantly enhanced both models’ performance compared to the baseline, yielding average improvements of 6.89% and 7.15% across accuracy, precision, recall, and specificity, for GCN and DIFFPOOL, respectively. The proposed framework offers a generalizable solution for mitigating the impact of suboptimal samples.

This paper introduces two novel datasets for cell motility and wound healing research: the Wound Healing Assay Dataset (WHAD) and the Cell Adhesion and Motility Assay Dataset (CAMAD). WHAD comprises time-lapse phase-contrast images of wound healing assays using genetically modified MCF10A and MCF7 cells, while CAMAD includes MDA-MB-231 and RAW 264.7 cells cultured on various substrates. These datasets offer diverse experimental conditions, comprehensive annotations, and high-quality imaging data, addressing gaps in existing resources. The collection methods, experimental designs, and annotation processes are detailed, emphasizing the reliability and utility of our datasets. WHAD and CAMAD enable researchers to develop robust algorithms for cell tracking, segmentation, and behavior analysis in phase-contrast microscopy images. The applications of these datasets span cancer research, image processing technique development, and automated analysis of wound healing and cell motility assays. By providing these resources, we aim to accelerate progress in biomedical image analysis and cell biology research.