OBJECTIVE: As the world's population ages, nursing homes are of increasing importance. In order to care for a growing number of older adults, intelligent technologies are needed. Artificial Intelligence can be utilised to enhance palliative care in nursing homes. However, the data needed to train artificially intelligent agents is lacking within this sensitive domain due to privacy issues. Therefore, it is difficult for researchers to develop technological solutions. With the advent of large language models, such as ChatGPT, new text generation methods are made possible using limited data. In this pilot study, we investigate the use of large language models to generate synthetic data. METHODS: We investigate the feasibility of using GPT-3.5 and GPT-4o models along with one-shot prompting to produce synthetic nurse notes which faithfully describe nursing home residents with met or unmet palliative care needs. We used LangChain to create a repeatable pipeline which can be adapted to different use-cases. We also compare the performance of both models using a set of qualitative and quantitative evaluations to determine which set of notes is more suitable for subsequent research. RESULTS: GPT-3.5 performed slightly better than GPT-4o in our qualitative healthcare professional analysis. Quantitative analysis revealed appropriately heterogenous results across contextual similarity, lexical overlap, sentiment, and readability scores. CONCLUSION: Our work is the first investigation of such a generation method in the nursing home palliative care domain. Further refinement and validation of such data is needed in order to ensure the safe use of our approach.

Deep learning for protein function prediction faces significant challenges in identifying disease-specific proteins. We present Extracellular Matrix Protein Predictor (EPOP), an advanced transfer learning framework leveraging protein language models to decode disease mechanisms. Focusing on pelvic organ prolapse (POP), which affects up to 50% of women worldwide, EPOP demonstrates AI's power to reveal novel therapeutic targets. We developed a sophisticated fine-tuning protocol for the ESM-2 model, optimized for ECM protein prediction. Our architecture integrates specialized attention mechanisms with interpretability modules, trained on expertly curated and balanced datasets totaling 80,000 proteins (40,000 ECM and 40,000 non-ECM). The framework employs a novel validation strategy using a 16,000-sample independent test set and clinical proteomics data. EPOP achieved unprecedented performance (99.40% accuracy) in ECM protein classification, significantly surpassing traditional deep learning architectures (10.81% improvement over Transformer models, 21.71% over Long Short-Term Memory). Applied to clinical samples, our model revealed a previously unknown pattern of ECM remodeling, identifying 24 novel disease-associated proteins. Model interpretability analysis uncovered specific sequence motifs and structural features critical for ECM protein function, providing mechanistic insights into disease progression. EPOP demonstrates how advanced AI bridges molecular analysis and clinical applications, uncovering novel therapeutic targets. Its success suggests broader applications across ECM-related disorders, potentially transforming approaches to diseases affecting connective tissue architecture.

This paper contains the research proposal of Deniz Mevlevioglu that was presented at the MMSys 2022 doctoral symposium. The benefits of real-time anxiety prediction in Virtual Reality are vast, including uses from therapy to entertainment. These anxiety predictions can be made using biosensors by tracking physiological measurements such as heart rate, electrical brain activity and skin conductivity. However, there are multiple challenges when trying to achieve accurate predictions. First of all, defining anxiety in a useful context and getting objective measurements to predict it can be difficult due to different interpretations of the word. Secondly, personal differences can make it difficult to fit everyone into a generalisable model. Lastly, Virtual Reality strives for immersion, and many systems that use objective measures such as on-body sensors to detect anxiety can make it hard for the user to immerse themselves into the virtual world. Our research aims to come up with a system that will address these problems and manage to get accurate and objective predictions of anxiety in real-time while still allowing the users to be immersed in the experience. To this end, we aim to use fast-performing classification models with multi-modal on-body sensor data to maximise comfort and minimise noise and inaccuracies.

Stroop Colour-Word Task has been widely used as a cognitive task. There are computerised and Virtual Reality versions of this task that are commonly used. The emotional version of the task, called the Emotional Stroop Colour-Word task is commonly used to induce certain emotions in a person. We are developing an application that brings the Emotional Stroop Colour-Word task into Virtual Reality. The aim of this application is to elicit different stress levels on the user and to record associated brain, heart and skin activity using wearable sensors. It is an immersive application that includes a tutorial, artificial intelligence generated audio instructions and a logging system for the user activity.

As the use of Virtual Reality (VR) expands across fields, new kinds of interaction methods are introduced. This study presents the Visual Heights VR experience that integrates natural breathing as an input method to provide visual respiratory feedback. Incorporating spatial audio, haptic feedback and breath visualisation, the experience aims to be highly immersive. This experience was made to be used as part of a controlled pilot study to see the effect of respiratory feedback on the user’s anxiety levels. The user’s anxiety is assessed by their heart rate, brain electrical activity, skin conductance and respiratory rate. These biosignals are recorded within the experience; captured by external hardware. The pieces of hardware used were Galvanic Skin Response to measure skin conductance, photoplethysmogram to measure heart rate; Electroencephalogram to measure the electrical activity in the brain, and a prototype device that records airflow on an axis from -1 to 1 for respiratory rate. It was found that the aforementioned prototype was not sufficient for calculating the respiratory rate. Results of the controlled study showed that the Visual Heights VR experience delivered the expected positive correlation between skin conductance and perceived height (r=.491, p < .05, N=1543) which suggests it is plausible to be used as a material for further research. As the integration of user’s physiological signals and breathing for visual feedback can contribute to therapeutic uses of VR, research with bigger sample sizes will be conducted to better investigate the relationship between visual respiratory feedback and anxiety using the Visual Heights VR experience.

Anxiety-inducing and assessment methods in Virtual Reality has been a topic of discussion in recent literature. The importance of the topic is related to the difficulty of getting accurate and timely measurements of anxiety without relying on self-report and breaking the immersion. To this end, the current study utilises the emotional version of a well-established cognitive task; the Stroop Color-Word Task and brings it to Virtual Reality. It consists of three levels; congruent which is used as control and corresponds with no anxiety, incongruent, which corresponds with mild anxiety and emotional, which corresponds with severe anxiety. This pilot serves two functions. The first is to validate the effects of the task using biosignal measurements. The second is to use the bio signal information and the labels to train a machine-learning algorithm. The information collected by the pilot will be used to decide what types of signals and devices to use in the final product, as well as what algorithm and time frame will be better suited for the purpose of accurately determining the user’s anxiety level within Virtual Reality without breaking the immersion.

Due to the pandemic limitations caused by Covid-19, people need to work at home and carry on the meetings virtually. Virtual meeting tools start popularizing and thriving. Those tools allow users to see each other through screen and camera, chat through voice and text, and share content or ideas through screen share. However, screen sharing protein models through virtual meetings is not easy due to the difficulty of viewing protein 3D (Three Dimensional) structures from a 2D (Two Dimensional) screen. Moreover, interactions upon a protein are also limited. ProMVR is a tool the author developed to tackle the issue that protein designers may find limitations working in a traditional 2D or 3D environment and they may find it hard to communicate their ideas with other designers. Since ProMVR is a VR tool, it allows users to “jump into” a virtual environment, take a close look at protein models, and have intuitive interactions.

Cystic Fibrosis is the most common life limiting genetic disease affecting Caucasians. With life expectancy for these patients rising, there is need to investigate interventions for a CF adult that does not impede on their busy lifestyle. This research describes an mHealth game data analytics system. The proposed system incorporates a game where a user must blow into the microphone to control an object through obstacles. This Android app collects game performance data, well-being questionnaire data, and initial blow strength. This data is then analysed and graphically represented for data interpretation on a web tool. The data collected may trigger an alert SMS to be sent to the patient based on three criteria which can be customised through the web tool for a more individualised analysis. The mHealth app is created using Android, and data is from this app posted to a remote server and stored in a database for analysis. The web tool comprises of PHP, MySQL, HTML, JavaScript and Text API. On opening the app, the patient creates a profile and all subsequent data is stored under this profile for analysis. Default alert criteria is saved for each patient which can be edited by a CF multidisciplinary team member. By default, if the user records an increase in sputum volume, or decline in baseline, or decline in game performance five times consecutively, a SMS alert is sent to the patient. This proposed system is currently undergoing pilot testing with a small CF patient cohort. Comparative data for three CF patients who utilised the system over a two month period is also discussed here within. This research concludes that there is a need for mHealth data analysis systems for adults with CF. The system presented is found to be engaging among CF adult patients and can be incorporated into their daily lifestyle with ease. Questionnaire data recorded through the app regarding therapy compliance suggests a correlation with game performance. This correlation can serve as a practical example for the importance of therapy compliance. Additionally, it can demonstrate how non-compliance can lead to possible exacerbations.

The world we live in today is a mobile one, whereby we are surrounded phones, PDA's and laptops. This paper looks at how all these devices can be used together of a collaborative fashion to solve parallel computing problems in a Java based environment. The means by which cross platform Bluetooth enabled applications can be developed and executed is examined, enabling mobile message passing to come into its own. The effectiveness of this parallel architecture is examined, with real world test results being presented to show that cross platform mobile parallel computing is more than a viable option for the world of today.

Mobility especially the flexibility given to us by the mobile phone is the future of computing as we know it. No longer are we restricted to sitting at a desk in front of a powerful desktop machine. Mobile technology of today allows users to work, learn and play no matter where they may be. Wireless technology is becoming more and more a standard feature of computing, so much so that it is expected that approximately two billion Bluetooth enabled devices will have been produced by the end of 2007. This paper examines how Bluetooth application development may be simplified for the programmer by use of the mobile message passing interface (MMPI). It explores a selection of application areas that can benefit from this simplified means of wireless inter-device communication, including: compute intensive tasks, mobile learning and multi-player gaming.

It is estimated that by the end of 2007 approximately two billion Bluetooth enabled devices will have been produced to date. The mobile phone market has a significant share of this. The gaming sector has done little to promote the use of bluetooth for gaming. This reflects the limited number of bluetooth enabled games currently available. Why should the modern day phone user have to compete only against the device, when a far more competitive, and enjoyable game can be achieved by the interaction of two or more human players. The development of such games, typically requires significant bluetooth specific api calls to setup the communications mechanism, and carry out communication. This paper details how the Mobile Message Passing Interface (MMPI) may be applied the the realm of bluetooth gaming. In particular this paper introduces a framework that allows for the transformation of a single player game, to a multiplayer Bluetooth enabled game with the minimum of code changes.

In this paper we investigate a new algorithm for the feedback-guided dynamic loop scheduling (FGDLS) method in the discrete case. The method uses a feedback-guided mechanism to schedule a parallel loop within a sequential outer loop. The execution times and the scheduling bounds for the current outer iteration are used to find the scheduling bounds of the next outer iteration. An O(p+log p) algorithm has been proposed for the discrete case where it was proved to achieve optimal bounds in only a few iterations. This articles introduces an O(log p) algorithm for the discrete case and presents some properties of it.

This paper outlines how the mobile message passing interface (MMPI) may be used for parallel computation. MMPI allows parallel programming of mobile devices over a Bluetooth network. This paper gives an overview of the MMPI library, and demonstrates that mobile devices are capable of parallel computation. An example of matrix multiplication O(n3) is used to show this

In this article we introduce a new iterative method for integral partition called the feedback guided dynamic integral partition (FGDIP) algorithm. The problem to study is the partition of a definite integral into p identical sub-integrals. The method generates iteratively a sequence of integral bounds by re-balancing the previous integral partition to achieve a better one. A simple convergence condition is also proposed. Experimental results show that the proposed method FGDIP achieves better performance than the classical Newton's method