BLE beacon networks are widely adopted for IoT and smart city applications, but they are susceptible to security threats, including piggybacking and spoofing attacks. These attacks can infringe on or even jeopardize the profitability of network owners. To address this issue, BLE beacon packets are often encrypted or updated periodically, which requires frequent synchronization between individual beacons and a centralized server, consuming significant resources. We propose a novel crowd-assisted, secure BLE identifier (ID) updating framework that significantly reduces network overhead. The proposed framework quantifies mobile crowdsourcing participant presence to dynamically adjust the pool size of the required beacon IDs. Our experiments, which factor in real-life user presence information, prove the practicality of our framework, which reduced the network resource consumption by up to 90%.

To address the short-lived battery lifetime of Bluetooth low energy (BLE) beacons, researchers proposed solar-powered designs, equipped with rechargeable energy storage such as a supercapacitor. However, accurately monitoring the energy status - an essential step for device maintenance - has shown to be a major concern. Existing energy status monitoring methods, which are either crowd-assisted or require on-site data collection, suffer from severe losses of energy status information. This paper presents an energy status recovery framework with support vector regression (SVR) to address this issue. The proposed framework leverages recurrence training of SVR with lost energy status information to capture features from discharge behavior, achieving high accuracy while minimizing training and prediction time. Multiple real-life BLE beacon energy level records are evaluated to demonstrate that our proposed framework can recover the energy information with at least 98% accuracy under a data loss rate of up to 99%.

With the current and continuous advancements in the field of text-to-image modeling, it has become critical to design prompts that make the best of these model capabilities and guides them to generate the most desirable images, and thus the field of prompt engineering has emerged. Here, we study a method to use prompt engineering to enhance text-to-image model representation of the Arabic culture. This work proposes a simple, novel approach for prompt engineering that uses the domain knowledge of a state-of-the-art language model, GPT, to perform the task of prompt augmentation, where a simple, initial prompt is used to generate multiple, more detailed prompts related to the Arabic culture from multiple categories through a GPT model through a process known as in-context learning. The augmented prompts are then used to generate images enhanced for the Arabic culture. We perform multiple experiments with a number of participants to evaluate the performance of the proposed method, which shows promising results, specially for generating prompts that are more inclusive of the different Arabic countries and with a wider variety in terms of image subjects, where we find that our proposed method generates image with more variety 85 % of the time and are more inclusive of the Arabic countries more than 72.66 % of the time, compared to the direct approach.

With growing demands for Internet of Things (IoT) applications, BLE beacon networks are rapidly being adopted. Periodic battery replacement operations and onsite maintenance are required to ensure continuous and reliable service. These operations are labor intensive and resource exhaustive. Therefore, Bluetooth gateways/mobile devices are often employed to monitor/collect the energy status. However, the gateways consume a considerable amount of network requests, and the user existence influences the data collection, thus the data accuracy, based on mobile devices. Reducing the number of energy status reports and maintaining the high accuracy of the energy status monitoring service is essential to catalyze a generic adoption of beacon networks and IoT infrastructure of similar nature in more businesses and real-life applications. In this article, we proposed a novel energy status monitoring framework that will dynamically change the energy status report interval based on the discharging rate of the battery, thereby reducing the total number of network requests and maintaining the required accuracy of energy status. The proposed framework identifies the BLE beacons with similar battery discharging rates, suggests a dynamic report interval, and leverages this information to reduce the number of energy status reports. We have experimented with real-life BLE beacon energy status data for 50 days to demonstrate that we could reduce the total number of network requests up to 70% while retaining 99% estimation accuracy.

To address the short-lived battery lifetime of Bluetooth low energy (BLE) beacons, solar-powered designs were proposed, equipped with rechargeable energy storage such as a supercapacitor. However, energy status monitoring, which is essential for device maintenance, proved to be a major concern as the energy status of energy harvesting devices can change quickly due to charging and discharging behaviours. Existing energy status monitoring methods performed in a crowd-assisted manner or by demanding on-site data collections are accompanied by severe loss of energy status information. This paper presents an accurate energy status recovery framework with SVR to address this issue. The proposed framework leverages recurrent training of SVR with lost energy status information to capture features from discharge behaviour to achieve high accuracy while minimizing the training and prediction time. Multiple real-life BLE beacon energy level records are evaluated to demonstrate that our proposed framework can recover the energy information with at least 90% accuracy under a data loss rate of up to 99%.

One key feature of Bluetooth low energy (BLE) beacons is the received signal strength, which can be used to estimate the distance between any Bluetooth-compatible receiver (e.g., smartphone, tablet, etc.) and fixed deployed beacon. Although received signal strength (RSS) can be measured easily with commonly available smart devices, the measurements are unreliable, in which general estimation models are not robust to different hardware and settings for real deployed beacon networks. Furthermore, the lack of consideration for object mobility in these models undermines its practicality. Motivated by the above limitations, this article proposes a novel distance classifier, d-Classifier, to classify the distance with a feature vector constructed with features such as hardware type and deployment environment to improve the robustness. Moreover, comprehensive experiments related to mobility are conducted to study the relationship between packet receiving rate and estimation accuracy. Improved performance can be achieved by providing extra mobility information with a list of RSS values during estimation. The proposed classifier is validated with an extensive data set that includes over 200 k data collected from real beacon networks. Overall, our proposed d-Classifier achieves a significant performance gain,$>25\%$accuracy improvement, over its prior arts.

With the global population ageing rapidly, Alzheimer's disease (AD) is particularly prominent in older adults, which has an insidious onset followed by gradual, irreversible deterioration in cognitive domains (memory, communication, etc). Thus the detection of Alzheimer's disease is crucial for timely intervention to slow down disease progression. This paper presents a comparative study of different acoustic and linguistic features for the AD detection using various classifiers. Experimental results on ADReSS dataset reflect that the proposed models using ComParE, X-vector, Linguistics, TFIDF and BERT features are able to detect AD with high accuracy and sensitivity, and are comparable with the state-of-the-art results reported. While most previous work used manual transcripts, our results also indicate that similar or even better performance could be obtained using automatically recognized transcripts over manually collected ones. This work achieves accuracy scores at 0.67 for acoustic features and 0.88 for linguistic features on either manual or ASR transcripts on the ADReSS Challenge1test set.

Bluetooth Low Energy (BLE) beacon networks are a popular infrastructure for IoT and smart city applications due to their scalability and affordability, as well as the proliferation of Bluetooth-enabled devices. However, BLE beacon networks suffer from short battery lifetime, inducing additional maintenance costs. Previous works have tackled this problem by proposing a more energy-efficient BLE beacon firmware that will change its operating configuration based on user existence information. However, previous efforts could not adapt to varying user traffic conditions and therefore was impractical. To address this issue, this paper proposes a novel neural network-driven framework, User-P, that extends beacon lifetime by changing its operating configuration by predicting user traffic conditions. Furthermore, the paper also presents a novel machine learning method tailored for user traffic prediction. Last but not least, the effectiveness of the proposed framework and methods are proven through a set of simulations. The simulation results show that the proposed framework can extend the beacon lifetime by 180% in comparison to that of the state-of-the-art techniques.

Real-time CNN-based object detection models for applications like surveillance can achieve high accuracy but are computationally expensive. Recent works have shown 10 to 100× reduction in computation cost for inference by using domain-specific networks. However, prior works have focused on inference only. If the domain model requires frequent retraining, training costs can pose a significant bottleneck. To address this, we propose Dataset Culling: a pipeline to reduce the size of the dataset for training, based on the prediction difficulty. Images that are easy to classify are filtered out since they contribute little to improving the accuracy. The difficulty is measured using our proposed confidence loss metric with little computational overhead. Dataset Culling is extended to optimize the image resolution to further improve training and inference costs. We develop fixed-angle, long-duration video datasets across several domains, and we show that the dataset size can be culled by a factor of 300× to reduce the total training time by 47× with no accuracy loss or even with slight improvement.

Bluetooth low energy (BLE) beacon network is one of the most favored IoT infrastructures due to its flexibility and scalability. Monitoring and updating the battery statuses of the on-site BLE beacons is an essential task for reliable operation and timely maintenance of the infrastructure. However, unregulated frequent updates of the battery statuses result in stressing the beacon network management platform, possibly threatening the reliable operation of the infrastructure. Whereas too infrequent updates degrade the freshness and reliability of the updated information. Without a reliable estimation on battery status, management and timely battery replacement operation would be difficult. To address this issue, this paper presents an efficient update method of battery status for BLE beacon network that minimizes the stress on the management platform server. The proposed approach leverages the correlation in battery status information between certain beacons to reduce the number of necessary updates while retaining high accuracy. Necessary reference data estimation, reference data reliability checking, and error correction on the estimation are the three major components in the solution. An estimation model allows accurate estimation in the cold-start stage. Moreover, an error-correction model allows to check the reliability of reference data and make a correction on the estimated value.

A BLE beacon is a small electronic device that has recently been proposed as a building block to construct an infrastructure supporting emerging smart applications. However, due to its simple communication protocol architecture, which broadcasts a static payload, a BLE beacon-based infrastructure is vulnerable to different types of abuses and attacks, in particular free-riding and device spoofing. Many beacon manufacturers propose dynamically randomizing beacon advertisement packets at the device firmware level as a solution. However, this approach is difficult to implement for already deployed beacon nodes as it requires a firmware update on each device. To alleviate these drawbacks, a crowd-assisted architecture for securing BLE beacons is proposed in this paper. A detailed architecture is presented along with experimental results and an implementation to demonstrate its feasibility. It is found that the beacon ID can be changed by user's mobile phone within a 20 m range with probability of almost 100% under both stationary and mobile conditions.

MOTIVATION: The classification of proteins into homologous groups (families) allows their structure and function to be analysed and compared in an evolutionary context. The modular nature of eukaryotic proteins presents a considerable challenge to the delineation of families, as different local regions within a single protein may share common ancestry with distinct, even mutually exclusive, sets of homologs, thereby creating an intricate web of homologous relationships if full-length sequences are taken as the unit of evolution. We attempt to disentangle this web by developing a fully automated pipeline to delineate protein subsequences that represent sensible units for homology inference, and clustering them into putatively homologous families using the Markov clustering algorithm. RESULTS: Using six eukaryotic proteomes as input, we clustered 162,349 protein sequences into 19,697-77,415 subsequence families depending on granularity of clustering. We validated these Markov clusters of homologous subsequences (MACHOS) against the manually curated Pfam domain families, using a quality measure to assess overlap. Our subsequence families correspond well to known domain families and achieve higher quality scores than do groups generated by fully automated domain family classification methods. We illustrate our approach by analysis of a group of proteins that contains the glutamyl/glutaminyl-tRNA synthetase domain, and conclude that our method can produce high-coverage decomposition of protein sequence space into precise homologous families in a way that takes the modularity of eukaryotic proteins into account. This approach allows for a fine-scale examination of evolutionary histories of proteins encoded in eukaryotic genomes. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online. MACHOS for the six proteomes are available as FASTA-formatted files: http://research1t.imb.uq.edu.au/ragan/machos.