Intangible Cultural Heritage (ICH) faces challenges to survival and growth from globalization and economic pressures. Research has explored how Generative AI (GenAI) can capture, represent, and convey cultural meaning to help address these challenges. This paper bridges current GenAI ICH strategies and explores the forms of creativity that emerge when artisans use their embodied craft skills to create with GenAI fine-tuned on their own ICH. Collaborating in-situ with Huayao embroiderers in rural China produced eight novel hand-embroidered pieces from our custom GenAI workflows trained on Huayao ICH. Embroiderers noted that GenAI tools could enhance creativity and had the potential to support innovation within tradition, yet human creativity and manual refinement remained necessary to ensure cultural authenticity. We suggest that our combination of rapid iterative GenAI and in-situ slow making by hand created a space in which discourses on the meaning and value of ICH could be explored.

Once-for-All (OFA) training enables a single super-net to generate multiple sub-nets tailored to diverse deployment scenarios, supporting flexible trade-offs among accuracy, robustness, and model-size without retraining. However, as the number of supported sub-nets increases, excessive parameter sharing in the backbone limits representational capacity, leading to degraded calibration and reduced overall performance. To address this, we propose SOLAR (Switchable Output Layer for Accuracy and Robustness in Once-for-All Training), a simple yet effective technique that assigns each sub-net a separate classification head. By decoupling the logit learning process across sub-nets, the Switchable Output Layer (SOL) reduces representational interference and improves optimization, without altering the shared backbone. We evaluate SOLAR on five datasets (SVHN, CIFAR-10, STL-10, CIFAR-100, and TinyImageNet) using four super-net backbones (ResNet-34, WideResNet-16-8, WideResNet-40-2, and MobileNetV2) for two OFA training frameworks (OATS and SNNs). Experiments show that SOLAR outperforms the baseline methods: compared to OATS, it improves accuracy of sub-nets up to 1.26%, 4.71%, 1.67%, and 1.76%, and robustness up to 9.01%, 7.71%, 2.72%, and 1.26% on SVHN, CIFAR-10, STL-10, and CIFAR-100, respectively. Compared to SNNs, it improves TinyImageNet accuracy by up to 2.93%, 2.34%, and 1.35% using ResNet-34, WideResNet-16-8, and MobileNetV2 backbones (with 8 sub-nets), respectively. The code of SOLAR is publicly available at: https://github.com/NAIL-UH/SOLAR and its website can be accessed at https://saktx.github.io/solar.github.io/.

Generative Artificial Intelligence (GAI), particularly text-to-image (T2I) generation tools, presents new possibilities for preserving and innovating traditional cultural patterns. However, AI-generated images often lack cultural context, which risks cultural bias and the loss of cultural significance. This study explores the use of GAI in generating culturally meaningful patterns, focusing on Chinese intangible cultural heritage Huayao cross-stitch as a case study. By applying Low-Rank Adaptation (LoRA) fine-tuning to optimize T2I tools and using in-situ interviews and focus groups, we collected feedback from 18 Huayao artisans. The results show that while fine-tuned models improved stylistic accuracy, the cultural meaning of the patterns remained insufficient. This research highlights AI's limited role in cultural innovation and emphasizes the necessity for dynamically maintaining cultural authenticity through the daily practices of cultural holders. It also reflects on how AI might have a long-term impact on the creative position of artisan communities.

With the success of wireless sensor networks (WSNs), traditional engineering and infrastructure industries are starting to develop solutions using WSN technologies. One of the main challenges of designing and developing WSNs for industrial monitoring and control is satisfying their strict reliability requirements. In this paper, we present a network-level reliability model, namely, end-to-end data delivery reliability (E2E-DDR), for estimating and optimizing the reliability performance of WSNs. In the E2E-DDR model, a framework is presented for capturing the mapping function between the packet reception ratio, background noise, and received signal strength (RSS). We use an alpha-stable distribution to accurately represent the background noise and a modified log-normal path loss model to more realistically describe the RSS. We also report a comprehensive performance evaluation performed by applying the E2E-DDR model in a real-world case study to estimate the network-level reliability and optimize the WSN deployment parameters.

Accurate estimation of PHY conditions would allow for better cross-layer resource management and provisioning. Unfortunately, most commercial-of-the-shelf (COTS) devices typically provide very limited information. For example, IEEE 802.15.4 Zigbee radios only provide received signal strength indicator (RSSI), link quality indicator (LQI) and noise floor readings, which are stored in frame check sequence (FCS) of MAC frames. In this paper, we revisit the issue of link quality prediction in IEEE 802.15.4 low rate wireless personal area networks (LR-WPAN) analytically and experimentally. By deciphering LQI readings available in Zigbee radios with CC2420 chipset, we demonstrate for the first time that LQI truly reflects the signal-to-noise ratio (SNR) at receiver. We also investigate the chip correlation (CORR) defined in CC2420 data sheet, and verify its relationship with LQI readings through measurement studies. Furthermore, in order to predict the instantaneous link quality for commodity Zigbee radios, we develop an inference model under different channel environments that uses instantaneous LQI readings as input. The proposed model is validated using extensive simulation and experimental study. We believe it will lead to more informed resource management decisions in WPANs.

The foray of wireless networks into time-sensitive applications demands support for emergency services, with guaranteed QoS bounds. The IEEE 802.11e technology could be an attractive option for emergency networking, if the protocol design is enhanced to achieve deterministic QoS for emergency traffic during distributed network operation. We propose channel preemptive EDCA scheme, an in-channel service preemption methodology for the 802.11e EDCA framework where emergency traffic incorporates the privileges to preempt the services of lower priority traffic for better QoS guarantees. The CP-EDCA results for emergency frames depict up to 50% decrease in medium access delays and insensitivity to routine traffic competition.

The focus of the paper is on establishing the mapping between a fixed camera and the PTZ camera using master-slave configuration. The authors proposed and developed a general framework for arbitrary camera topology. The method estimates online the time-variant transformation between a fixed master camera having global view and tracking the target and a slave camera taking close up images. In the experiment, the method transforms the target feet position of the master camera to head position of the slave camera. It takes into account both zooming camera and target uncertainties.

Malignant melanoma has a good prognosis if treated early. Accurate skin lesion segmentation from the background skin is important not only because the shape feature can be directly derived from the process, but also because it can provide a scope for texture analysis. In this paper, we propose an evolutionary strategy based segmentation algorithm to identify the lesion area by an ellipse. It can detect the lesion automatically without setting parameters manually. The method is validated by experiments and comparisons with manually segmentation by an expert and algorithms developed by M. Doshi, et al. (2004).

The immense explosion of geographically referenced data calls for efficient discovery of spatial knowledge. One of the special challenges for spatial data mining is that information is usually not uniformly distributed in spatial datasets. Consequently, the discovery of regional knowledge is of fundamental importance for spatial data mining. This paper centers on discovering regional association rules in spatial datasets. In particular, we introduce a novel framework to mine regional association rules relying on a given class structure. A reward-based regional discovery methodology is introduced, and a divisive, grid-based supervised clustering algorithm is presented that identifies interesting subregions in spatial datasets. Then, an integrated approach is discussed to systematically mine regional rules. The proposed framework is evaluated in a real-world case study that identifies spatial risk patterns from arsenic in the Texas water supply.

There has been increasing interest to design better probability estimation trees, or PETs, for ranking and probability estimation. Capable of generating class membership probabilities, PETs have been shown to be highly accurate and flexible for many difficult problems, such as cost-sensitive learning and matching skewed distributions. There are a large number of PET algorithms available, and about ten of them are well-known. This large number provides an advantage, but it also creates confusion in practice. One would ask "given a new dataset, which algorithm to choose and what performance to expect and not to expect? What are the reasons to explain either good or bad performance under different situations?" In this paper, we systematically, for the first time, answer these important questions by conducting a large-scale empirical comparison of five popular PETs by examining their AUC, MSE and error rate "learning curves" (instead of training-test split based cross-validation). Using the maximum AUC achieved by any of the evaluated probability estimation tree algorithms, we demonstrate that the preference of a probability estimation tree on different evaluation metrics can be accurately characterized by the "signal-noise separability" of the dataset, as well as some other observable statistics of the dataset explained further in the paper. Moreover, in order to understand their relative performance, many important and previously unrevealed properties of each PET's mechanism and heuristics are analyzed and evaluated. Importantly, a practical guide for choosing the most appropriate PET algorithm given a new data mining problem is provided.

Much work on skewed, stochastic, high dimensional, and biased datasets usually implicitly solve each problem separately. Recently, we have been approached by Texas Commission on Environmental Quality (TCEQ) to help them build highly accurate ozone level alarm forecasting models for the Houston area, where these technical difficulties come together in one single problem. Key characteristics of this problem that are challenging and interesting include: 1) the dataset is sparse (72 features, and 2% or 5% positives depending on the criteria of "ozone days"), 2) evolving over time from year to year, 3) limited in collected data size (7 years or around 2500 data entries), 4) contains a large number of irrelevant features, 5) is biased in terms of "sample selection bias", and 6) the true model is stochastic as a function of measurable factors. Besides solving a difficult application problem, this dataset offers a unique opportunity to explore new and existing data mining techniques, and to provide experience and guidance for similar problems. Our main technical focus addresses on how to estimate reliable probability given both sample selection bias and a large number of irrelevant features, and how to choose the most reliable decision threshold to predict the unknown future with different distribution. On the application side, the prediction accuracy of our approach is 20% higher in recall (correctly detects 1 to 3 more ozone days, depending on the year) and 10% higher in precision (15 to 30 fewer false alarm days per year) than state-of-the-art methods used by air quality control scientists, and these results are significant for TCEQ.

We present a distributed vision-based surveillance system. The system acquires and processes grey level images through one or multiple camera units monitoring certain area(s) via a local area network (LAN) and is capable of combining information from multiple camera units to obtain a consensus decision. It can be trained to detect certain type of intrusions, for example pedestrians, a group of pedestrians, vehicles, pets, etc., and minimizes false alerts due to other non-interested intrusions. As a case study, we aim to detect pedestrian/vehicle in an observation area. Our vision-based intrusion detection approach consists of two main steps: background subtraction based hypothesis generation (HG) and appearance-based hypothesis verification (HV). HG hypothesizes possible threats (intrusions), and HV verifies those hypotheses using a Gabor filter for feature extraction and support vector machines (SVMs) for classification. The system has been tested in an unconstrained outdoor environment, illustrating good performance.

Plant monitoring and diagnosis are usually integrated as one process to detect and isolate suspect symptoms and use these symptoms to find the root cause of the failure. The research reported here enables two new plant monitoring and diagnosis frameworks/architectures, each of which employs multiple fuzzy rule-based systems in parallel and in sequence. By employing fuzzy sets and by constructing a decision concerning the normalcy of system behavior in stages, we are able to exploit far more information contained in the signals (one signal per sensor). The paper focuses on a new transient identification scheme using fuzzy logic. A fuzzy rule based transient identification system is implemented in MATLAB. The experiments demonstrate how traditional features such as WOLP and HAS, for transient identification can be fuzzified to improve the efficiency and performance of traditional classifiers used for the same purpose.

We consider the problem of gender classification from frontal facial images using genetic feature subset selection. We argue that feature selection is an important issue in gender classification and demonstrate that Genetic Algorithms (GA) can select good subsets of features (i.e., features that encode mostly gender information), reducing the classification error. First, Principal Component Analysis (PCA) is used to represent each image as a feature vector (i.e., eigen-features) in a low-dimensional space. Genetic Algorithms (GAs) are then employed to select a subset of features from the low-dimensional representation by disregarding certain eigenvectors that do not seem to encode important gender information. Four different classifiers were compared in this study using genetic feature subset selection: a Bayes classifier, a Neural Network (NN) classifier, a Support Vector Machine (SVM) classifier, and a classifier based on Linear Discriminant Analysis (LDA). Our experimental results show a significant error rate reduction in all cases. The best performance was obtained using the SVM classifier. Using only 8.4% of the features in the complete set, the SVM classifier achieved an error rate of 4.7% from an average error rate of 8.9% using manually selected features.

Niching or speciation is a particularly appropriate for multimodal function optimization. An irony in EC research is that genetic algorithms (GAs) are not touted primarily as function optimizers yet all reported niching research is in the GA context. Evolution strategies (ES) and major variants of evolutionary programming are better suited by design for global optimization. Borrowing methods that have been reported for niching in GAs, we have applied them to an EC algorithm that resembles ES in structure. We have found that the selection methods in ES, e.g., (/spl mu/+/spl lambda/), interact satisfactorily with niching strategies used in GAs. On the other hand, adapting selection methods such as SUS that minimize bias does not lead to favorable results. This is counter to expectations but can be reconciled with prevailing theories. We conclude with a conjecture concerning a lower bound on population size for multimodal optimization.

Software quality modeling can be used by a software maintenance project to identify a limited set of software modules that probably need improvement. A model's goal is to recommend a set of modules to receive special treatment. The purpose of the paper is to report our experiences modeling software quality with classification trees, including necessary preprocessing of data. We conducted a case study on two releases of a very large legacy telecommunications system. A module was considered fault-prone if any faults were discovered by customers, and not fault-prone otherwise. Software product, process, and execution metrics were the basis for predictors. The TREEDISC algorithm for building classification trees was investigated, because it emphasizes statistical significance. Numeric data, such as software metrics, are not suitable for TREEDISC. Consequently, we transformed measurements into discrete ordinal predictors by grouping. This case study investigated the sensitivity of modeling results to various groupings. We found that robustness, accuracy, and parsimony of the models were influenced by the maximum number of groups. Models based on two sets of candidate predictors had similar sensitivity.