Cellular-resolution imaging is critical in medical diagnosis because it can reveal fine-scale
structures within tissues. Optical coherence tomography (OCT) is a cellular-resolution imaging
technique with the advantages of non-invasive, painless, fast, real-time imaging, etc. The principle
is based on low-coherence interferometry, which yields depth-resolved imaging with micrometerscale
resolution. In clinical practice, the safety, image quality, scanning speed, and tissue
penetration depth of OCT are all important characteristics. Skin cancer is one of the most common
cancers and is among the costliest of all cancers to treat. In our study, we used a Mirau-based fullfield
optical coherence tomography to measure five in vitro skin cells: keratinocyte (HaCaT cell
line), melanocyte, squamous cell carcinoma cell line (A431), and two melanoma cell lines, i.e.,
A375 and A2058. The light source of the Mirau-based FF-OCT system is a cerium-doped yttrium
aluminum garnet (Ce:YAG) single-clad crystal fiber, and generates a 560-nm central wavelength
with a 99-nm full width at half maximum (FWHM). We adopted CNN-based supervised deeplearning
classifiers to discriminate the skin cells and achieved a 95% classification accuracy. This
demonstrates that CNN deep-learning is a powerful tool for identifying cellular-resolution OCT
images. However, obtaining a large number of labeled samples is time-consuming and laborintensive.
In order to reduce label samples in practical applications, we tried to use data
augmentation, self-supervised learning, and transfer learning. These analyses showed interesting
results and efficiently achieved such a goal.

This research explores the integration of artificial intelligence (AI) into IELTS learning, focusing on its psychological impact and the challenges it presents. AI-driven tools like AI Smart Yu offer personalized, data-driven resources for students, such as interactive speaking simulations and adaptive reading modules, which aim to enhance exam preparation. However, despite these advancements, the psychological aspects of AI in education remain underexplored. Key concerns include AI’s ability to foster intrinsic motivation, reduce exam anxiety, and offer culturally sensitive feedback. AI systems often fail to address the emotional needs of students, particularly middle schoolers, which could hinder motivation and development. Additionally, AI’s lack of adaptability to students’ diverse learning styles and backgrounds poses challenges. To address these limitations, the research suggests incorporating educational psychology to develop AI tools that are not only technically efficient but also emotionally supportive and culturally relevant. A hybrid approach, blending AI with human interaction, is proposed to ensure comprehensive, personalized, and effective learning experiences for IELTS preparation.

The energy sector is a critical contributor to the growth and development of any country’s economy. However, ensuring robust cybersecurity within the context of smart energy services presents persistent usability challenges in an increasingly digital environment. This study explores the intersection of human-computer interaction (HCI), cybersecurity, and usability to identify and address issues that impact the overall security of smart energy management systems. By analyzing the complex relationships between users and security protocols, this research aims to enhance the security framework, promote better user adherence, and improve system usability. The study focuses on three primary objectives: (1) identifying the most prevalent usability issues in current cybersecurity practices; (2) examining the relationship between HCI and user compliance with security measures; and (3) proposing strategies to improve cybersecurity usability by leveraging HCI principles. Hybrid approaches utilizing artificial intelligence facilitate empirical analysis and framework evaluation. Additionally, a comparative study with six existing models has been conducted. By envisioning a future where security measures not only ensure enhanced protection but also integrate seamlessly into user experiences, this approach seeks to provide valuable insights into ongoing cybersecurity discussions and contribute to a more resilient security landscape against evolving digital threats.

Artificial intelligence (AI), including machine learning (ML) and deep learning (DL), learns by training and is restricted by the amount and quality of training data. Training involves a tradeoff between prediction bias and variance controlled by model complexity. Increased model complexity decreases prediction bias, increases variance, and increases overfitting possibilities. Overfitting is a significantly smaller training prediction error relative to the trained model prediction error for an independent validation set. Uncertain data generate risks for AI–ML because they increase overfitting and limit generalization ability. Specious confidence in predictions from overfit models with limited generalization ability, leading to misguided water resource management, is the uncertainty-related negative consequence. Improved data is the way to improve AI–ML models. With uncertain water resource data sets, like stream discharge, there is no quick way to generate improved data. Data assimilation (DA) provides mitigation for uncertainty risks, describes data- and model-related uncertainty, and propagates uncertainty to results using observation error models. A DA-derived mitigation example is provided using a common-sense baseline, derived from an observation error model, for the confirmation of generalization ability and a threshold identifying overfitting. AI–ML models can also be incorporated into DA to provide additional observations for assimilation or as a forward model for prediction and inverse-style calibration or training. The mitigation of uncertain data risks using DA involves a modified bias–variance tradeoff that focuses on increasing solution variability at the expense of increased model bias. Increased variability portrays data and model uncertainty. Uncertainty propagation produces an ensemble of models and a range of predictions.