Freezing of Gait (FoG) is a prevalent motor dysfunction experienced by patients with Parkinson’s disease, and while extensive research has focused on detecting FoG episodes in both clinical and home environments over the past decade, predictive approaches that enable preemptive prompting are still scarce. Most existing studies rely on fully supervised learning conditions, which pose significant challenges. This study introduces a semi-supervised learning framework utilizing a prototype network designed to leverage a FoG prediction model based on impaired gait patterns indicative of pre-FoG, simultaneously harnessing information from unlabeled sensor data for real-time FoG prediction. To validate our framework, we establish pre-FoG state labeling across four datasets: DAPHNet, CPGDD, PDTURN, and BXHC. Our developed real-time detection model demonstrates strong performance, achieving up to \(96\%\) sensitivity, \(99\%\) specificity, and \(91\%\) average accuracy on the cross-validation dataset. Furthermore, we compare the accuracy between semi-supervised and fully supervised modes, revealing that the semi-supervised approach yields improvements in average accuracy ranging from 0.01 to 0.02 across three datasets. Thus, our proposed method represents an effective strategy for the accurate prediction of FoG in real-time algorithmic settings.

Recent advancements in large language models (LLMs) have demonstrated their potential to significantly impact finance trading, particularly through sentiment analysis. The cryptocurrency market, known for its volatility and unpredictability, often renders price-based trading approaches inadequate. This necessitates the adoption of more sophisticated techniques such as market sentiment analysis, which can benefit from the insights provided by LLMs. This study introduces an innovative method that integrates sentiment analysis derived from five distinct LLMs with deep reinforcement learning to devise a cryptocurrency trading strategy. Recognizing that LLM outputs cannot be guaranteed to be infallibly accurate, which contributing to the LLM hallucinations, this paper details the implementation of a stringent outlier detection and removal process. By adopting a “Trust-The-Majority” strategy, the research aims to ensure that trading decisions are informed by reliable sentiment data. In addition, sentiment scores are traditionally timestamped to the publication of news or social media posts. To more accurately reflect the actual impact of such information on market sentiment, this study applies the Ebbinghaus Forgetting Curve to model the waning influence of information over time. This allows for a more nuanced understanding of how news affects market dynamics. The enhanced sentiment scores, in conjunction with traditional market data such as OHLCV (Open, High, Low, Close, Volume), are utilized by a deep reinforcement learning model to make trading decisions. Experimental results demonstrate that the proposed multi-LLM sentiment-driven framework improves trading performance in the fast-paced cryptocurrency market. The methodology outlined in this paper offers a solid foundation for incorporating real-time market sentiment analysis into financial applications.

Accurate measurement of oxygen uptake ( \(\dot{\mathrm{V}}\mathrm{O}_{2}\) ) dynamics and maximal oxygen consumption ( \(\dot{\mathrm{V}}\mathrm{O}_{2}\max\) ), a vital marker of cardiorespiratory fitness and exercise capacity, requires specialized exercise physiology laboratories with costly equipment. This study develops a Temporal Fusion Network (TFN) approach utilizing easily accessible physiological parameters (heart rate, heart rate reserve, tidal volume, and breathing frequency), which can be measured with wearable sensors, anthropometric variables (age, gender, height, and weight), as well as health status to estimate \(\dot{\mathrm{V}}\mathrm{O}_{2}\) dynamics during cardiopulmonary exercise testing (CPET). These input physiological parameters were derived from 140 laboratory CPET of a diverse cohort of adults (90 males, 50 females; 77 healthy, 63 smokers; average age: 26.6 years), to analyze \(\dot{\mathrm{V}}\mathrm{O}_{2}\) dynamics. The TFN model demonstrated high predictive accuracy to estimate \(\dot{\mathrm{V}}\mathrm{O}_{2}\) dynamics, with a Root Mean Square Error (RMSE) of 0.03 L/min and an R-squared ( R 2 ) value of 0.92, indicating robust performance across varied population groups. This TFN model paves the way for practical and cost-effective approach to estimate \(\dot{\mathrm{V}}\mathrm{O}_{2}\) during exercise, with potential integration with consumer health devices to expand accessibility and, enhance its utility for clinical and fitness applications.

Freezing of Gait (FoG) represents a critical and debilitating symptom of Parkinson's disease, posing significant challenges in patient mobility and safety. Numerous research efforts have focused on predicting the onset of FoG using wearable sensors and computational aids. However, the unpredictability and brief nature of FoG episodes complicate the ability to predict their onset in real-time, rendering timely detection a complex goal. In our study, we employed an Autoencoder-Wavenet network architecture designed to effectively utilize data from triaxial accelerometers and tri-axial gyroscopes positioned at the ankle. This approach allowed for a more nuanced analysis of movement patterns associated with FoG. Our findings indicated that this model successfully achieved a high level of accuracy in detecting FoG, with a specificity of 0.81, and sensitivity of 0.87. Furthermore, the model demonstrated the capability to provide real-time warnings of FoG onset, achieving a specificity of 0.68 and a sensitivity of 0.86. And an accuracy of 0.67 within a 1-second timeframe on the test set. Consequently, these results underscore the potential of our model in contributing to the real-time detection and management of FoG in Parkinson's disease patients.

Laser speckle contrast imaging (LSCI) is widely used for in vivo real-time detection and analysis of local blood flow microcirculation due to its non-invasive ability and excellent spatial and temporal resolution. However, vascular segmentation of LSCI images still faces a lot of difficulties due to numerous specific noises caused by the complexity of blood microcirculation's structure and irregular vascular aberrations in diseased regions. In addition, the difficulties of LSCI image data annotation have hindered the application of deep learning methods based on supervised learning in the field of LSCI image vascular segmentation. To tackle these difficulties, we propose a robust weakly supervised learning method, which selects the threshold combinations and processing flows instead of labor-intensive annotation work to construct the ground truth of the dataset, and design a deep neural network, FURNet, based on UNet++ and ResNeXt. The model obtained from training achieves high-quality vascular segmentation and captures multi-scene vascular features on both constructed and unknown datasets with good generalization. Furthermore, we intravital verified the availability of this method on a tumor before and after embolization treatment. This work provides a new approach for realizing LSCI vascular segmentation and also makes a new application-level advance in the field of artificial intelligence-assisted disease diagnosis.

Object detection (OD) is a crucial computer vision task that has seen the development of many algorithms and models over the years. While the performance of current OD models has improved, they have also become more complex, making them impractical for industry applications due to their large parameter size. To tackle this problem, knowledge distillation (KD) technology was proposed in 2015 for image classification and subsequently extended to other visual tasks due to its ability to transfer knowledge learned by complex teacher models to lightweight student models. This paper presents a comprehensive survey of KD-based OD models developed in recent years, with the aim of providing researchers with an overview of recent progress in the field. We conduct an in-depth analysis of existing works, highlighting their advantages and limitations, and explore future research directions to inspire the design of models for related tasks. We summarize the basic principles of designing KD-based OD models, describe related KD-based OD tasks, including performance improvements for lightweight models, catastrophic forgetting in incremental OD, small object detection, and weakly/semi-supervised OD. We also analyze novel distillation techniques, i.e. different types of distillation loss, feature interaction between teacher and student models, etc. Additionally, we provide an overview of the extended applications of KD-based OD models on specific datasets, such as remote sensing images and 3D point cloud datasets. We compare and analyze the performance of different models on several common datasets and discuss promising directions for solving specific OD problems.

Organizations rely heavily on time series metrics to measure and model key aspects of operational and business performance. The ability to reliably detect issues with these metrics is imperative to identifying early indicators of major problems before they become pervasive. It can be very challenging to proactively monitor a large number of diverse and constantly changing time series for anomalies, so there are often gaps in monitoring coverage, disabled or ignored monitors due to false positive alarms, and teams resorting to manual inspection of charts to catch problems. Traditionally, variations in the data generation processes and patterns have required strong modeling expertise to create models that accurately flag anomalies. In this paper, we describe an anomaly detection system that overcomes this common challenge by keeping track of its own performance and making changes as necessary to each model without requiring manual intervention. We demonstrate that this novel approach outperforms available alternatives on benchmark datasets in many scenarios.