Search

Hand-crafted features, such as Mel-filterbanks, have traditionally been the choice for many audio processing applications. Recently, there has been a growing interest in learnable front-ends that extract representations directly from the raw audio waveform. \textcolor{black}{However, both hand-crafted filterbanks and current learnable front-ends lead to fixed computation graphs at inference time, failing to dynamically adapt to varying acoustic environments, a key feature of human auditory systems.} To this end, we explore the question of whether audio front-ends should be adaptive by comparing the Ada-FE front-end (a recently developed adaptive front-end that employs a neural adaptive feedback controller to dynamically adjust the Q-factors of its spectral decomposition filters) to established learnable front-ends. Specifically, we systematically investigate learnable front-ends and Ada-FE across two commonly used back-end backbones and a wide range of audio benchmarks including speech, sound event, and music. The comprehensive results show that our Ada-FE outperforms advanced learnable front-ends, and more importantly, it exhibits impressive stability or robustness on test samples over various training epochs., Comment: Accepted by IEEE TASLP

The rapid growth of encryption has significantly enhanced privacy and security while posing challenges for network traffic classification. Recent approaches address these challenges by transforming network traffic into text or image formats to leverage deep-learning models originally designed for natural language processing, and computer vision. However, these transformations often contradict network protocol specifications, introduce noisy features, and result in resource-intensive processes. To overcome these limitations, we propose NetMatrix, a minimalistic tabular representation of network traffic that eliminates noisy attributes and focuses on meaningful features leveraging RFCs (Request for Comments) definitions. By combining NetMatrix with a vanilla XGBoost classifier, we implement a lightweight approach, LiM ("Less is More") that achieves classification performance on par with state-of-the-art methods such as ET-BERT and YaTC. Compared to selected baselines, experimental evaluations demonstrate that LiM improves resource consumption by orders of magnitude. Overall, this study underscores the effectiveness of simplicity in traffic representation and machine learning model selection, paving the way towards resource-efficient network traffic classification., Comment: Accepted to The Web Conference (WWW) 2025 Short Paper Track

Due to reasons of convenience and lack of tech literacy, transliteration (i.e., Romanizing native scripts instead of using localization tools) is eminently prevalent in the context of low-resource languages such as Sinhala, which have their own writing script. In this study, our focus is on Romanized Sinhala transliteration. We propose two methods to address this problem: Our baseline is a rule-based method, which is then compared against our second method where we approach the transliteration problem as a sequence-to-sequence task akin to the established Neural Machine Translation (NMT) task. For the latter, we propose a Transformer-based Encode-Decoder solution. We witnessed that the Transformer-based method could grab many ad-hoc patterns within the Romanized scripts compared to the rule-based method. The code base associated with this paper is available on GitHub - https://github.com/kasunw22/Sinhala-Transliterator/, Comment: 8 pages, 7 tables

The early detection of potential failures in industrial machinery components is paramount for ensuring the reliability and safety of operations, thereby preserving Machine Condition Monitoring (MCM). This research addresses this imperative by introducing an innovative approach to Real-Time Acoustic Anomaly Detection. Our method combines semi-supervised temporal convolution with representation learning and a hybrid model strategy with Temporal Convolutional Networks (TCN) to handle various intricate anomaly patterns found in acoustic data effectively. The proposed model demonstrates superior performance compared to established research in the field, underscoring the effectiveness of this approach. Not only do we present quantitative evidence of its superiority, but we also employ visual representations, such as t-SNE plots, to further substantiate the model's efficacy., Comment: 10 pages, 10 figures, ICMLC2024

The rapid pervasivity of the Internet of Things (IoT) calls for an autonomous and efficient resource management framework to seamlessly register and discover facilities and services. Cloud-Fog-Automation (CFA) standards provide a robust foundation for multi-tiered wireless architectures, enhancing cyber-physical system performance with advanced abstractions. This work is for resource allocation optimization in IoT networks, particularly in power management and time-frequency spreading techniques, ensuring deterministic connectivity, networked computing, and intelligent control systems. Auction game theory is pivotal in managing resource allocation in densely populated, high-demand IoT networks. By employing sealed-bid auctions based on Bayesian game theory, the uncertainties in individual hypotheses and channel states among IoT entities are effectively mitigated. A novel dispersion metric optimization further enhances the coordination of layer-specific IoT uplinks, enabling ultra-reliable, low-latency (URLLC) communication. Numerical results demonstrate the superior performance of this resilient architecture, achieving fair resource allocation with minimal power consumption and robust performance in unsecured scenarios., Comment: 18 pages

This work delves into solving the two dimensional Poisson problem through the Finite Element Method which is relevant in various physical scenarios including heat conduction, electrostatics, gravity potential, and fluid dynamics. However, finding exact solutions to these problems can be complicated and challenging due to complexities in the domains such as re-entrant corners, cracks, and discontinuities of the solution along the boundaries, and due to the singular source function. Our focus in this work is to solve the Poisson equation in the presence of re entrant corners at the vertices of domain where some of the interior angles are greater than 180 degrees. When the domain features a re entrant corner, the numerical solution can display singular behavior near the corners. To address this, we propose a graded mesh algorithm that helps us to tackle the solution near singular points. We derive H1 and L2 error estimate results, and we use MATLAB to present numerical results that validate our theoretical findings. By exploring these concepts, we hope to provide new insights into the Poisson problem and inspire future research into the application of numerical methods to solve complex physical scenarios, Comment: 18 pages,10 figures

The neonatal period is the most vulnerable time for the development of seizures. Seizures in the immature brain lead to detrimental consequences, therefore require early diagnosis. The gold-standard for neonatal seizure detection currently relies on continuous video-EEG monitoring; which involves recording multi-channel electroencephalogram (EEG) alongside real-time video monitoring within a neonatal intensive care unit (NICU). However, video-EEG monitoring technology requires clinical expertise and is often limited to technologically advanced and resourceful settings. Cost-effective new techniques could help the medical fraternity make an accurate diagnosis and advocate treatment without delay. In this work, a novel explainable deep learning model to automate the neonatal seizure detection process with a reduced EEG montage is proposed, which employs convolutional nets, graph attention layers, and fully connected layers. Beyond its ability to detect seizures in real-time with a reduced montage, this model offers the unique advantage of real-time interpretability. By evaluating the performance on the Zenodo dataset with 10-fold cross-validation, the presented model achieves an absolute improvement of 8.31% and 42.86% in area under curve (AUC) and recall, respectively., Comment: Paper is accepted to IEEE International Conference on Systems, Man, and Cybernetics (SMC) 2024. Final Version

The long-distance detection of the presence of elephants is pivotal to addressing the human-elephant conflict. IoT-based solutions utilizing seismic signals originating from the movement of elephants are a novel approach to solving this problem. This study introduces an instrumentation system comprising a specially designed geophone-sensor interface for non-invasive, long-range elephant detection using seismic waves while minimizing the vulnerability of seismic signals to noise. The geophone-sensor interface involves a cascade array of an instrumentation amplifier, a second-order Butterworth filter for signal filtering, and a signal amplifier. The introduced geophone-sensor interface was tested under laboratory conditions, and then real-world experiments were carried out for tamed, partly tamed, and untamed elephants. The experimental results reveal that the system remains stable within the tested frequency range from 1 Hz to 1 kHz and the temperature range of 10C to 40C. The system successfully captured the seismic signals generated by the footfalls of elephants within a maximum detection range of 155.6 m, with an overall detection accuracy of 99.5%., Comment: in IEEE Access

Indium selenide, In2Se3, has recently attracted growing interest due to its novel properties, including room temperature ferroelectricity, outstanding photoresponsivity, and exotic in-plane ferroelectricity, which open up new regimes for next generation electronics. In2Se3 also provides the important advantage of tuning the electrical properties of ultra-thin layers with an external electrical and magnetic field, making it a potential platform to study novel two-dimensional physics. Yet, In2Se3 has many different polymorphs, and it has been challenging to synthesize single-phase material, especially using scalable growth methods, as needed for technological applications. In this paper, we use aberration-corrected scanning transmission electron microscopy to characterize the microstructure of twin-free single-phase ultra-thin layers of beta-In2Se3, prepared by a unique molecular beam epitaxy approach. We emphasize features of the In2Se3 layer and In2Se3/InP interface which provide evidence for understanding the growth mechanism of the single-phase In2Se3. This novel approach for forming high-quality twin-free single phase two-dimensional crystals on InP substrates is likely to be applicable to other technologically important substrates., Comment: 19 pages, 4 figures

Synthetic Aperture Radar (SAR) Automatic Target Recognition (ATR) is a key technique used in military applications like remote-sensing image recognition. Vision Transformers (ViTs) are the current state-of-the-art in various computer vision applications, outperforming their CNN counterparts. However, using ViTs for SAR ATR applications is challenging due to (1) standard ViTs require extensive training data to generalize well due to their low locality; the standard SAR datasets, however, have a limited number of labeled training data which reduces the learning capability of ViTs; (2) ViTs have a high parameter count and are computation intensive which makes their deployment on resource-constrained SAR platforms difficult. In this work, we develop a lightweight ViT model that can be trained directly on small datasets without any pre-training by utilizing the Shifted Patch Tokenization (SPT) and Locality Self-Attention (LSA) modules. We directly train this model on SAR datasets which have limited training samples to evaluate its effectiveness for SAR ATR applications. We evaluate our proposed model, that we call VTR (ViT for SAR ATR), on three widely used SAR datasets: MSTAR, SynthWakeSAR, and GBSAR. Further, we propose a novel FPGA accelerator for VTR, in order to enable deployment for real-time SAR ATR applications., Comment: SPIE DCS 2024

We present the observational mass-radius (M-R) relation for a sample of 47 magnetized white dwarfs (WDs) with the magnetic field strength (B) ranging from 1 to 773 MG, identified from the SDSS data release 7 (DR7). We derive their effective temperature, surface gravity (log g), luminosity, radius, and mass. While atmospheric parameters are derived using a Virtual Observatory Spectral Energy Distribution Analyzer (VOSA), the mass is derived using their location in the HR diagram in comparison with the evolutionary tracks of different masses. We implement this mass measurement instead of a more traditional method of deriving masses from log g, which is unreliable as is based on SED and generates errors from other physical parameters involved. The main disadvantage of this method is that we need to assume a core composition of WDs. As it is complicated to identify the exact composition of these WDs from low-resolution spectra, we use tracks for the masses 0.2 to 0.4 solar mass assuming a He-core, 0.5 to 1.0 solar mass assuming CO core, and above solar mass assuming O-Ne-Mg core. We compare the observed M-R relation with those predicted by the finite temperature model by considering different B, which are well in agreement considering their relatively low surface fields, less than or of the order of 10^9 G. Currently, there is no direct observational detection of magnetized WDs with B > 10^9 G. We propose that our model can be further extrapolated to higher B, which may indicate the existence of super-Chandrasekhar mass (M > 1.4 solar mass) WDs at higher B., Comment: 9 latex pages including 10 figures (11 pdf figures); accepted for publication in MNRAS

Image classifiers for domain-specific tasks like Synthetic Aperture Radar Automatic Target Recognition (SAR ATR) and chest X-ray classification often rely on convolutional neural networks (CNNs). These networks, while powerful, experience high latency due to the number of operations they perform, which can be problematic in real-time applications. Many image classification models are designed to work with both RGB and grayscale datasets, but classifiers that operate solely on grayscale images are less common. Grayscale image classification has critical applications in fields such as medical imaging and SAR ATR. In response, we present a novel grayscale image classification approach using a vectorized view of images. By leveraging the lightweight nature of Multi-Layer Perceptrons (MLPs), we treat images as vectors, simplifying the problem to grayscale image classification. Our approach incorporates a single graph convolutional layer in a batch-wise manner, enhancing accuracy and reducing performance variance. Additionally, we develop a customized accelerator on FPGA for our model, incorporating several optimizations to improve performance. Experimental results on benchmark grayscale image datasets demonstrate the effectiveness of our approach, achieving significantly lower latency (up to $16\times$ less on MSTAR) and competitive or superior performance compared to state-of-the-art models for SAR ATR and medical image classification., Comment: Accepted to IEEE ICIP 2024

Unmanned aerial vehicles (UAVs) are envisioned to provide diverse services from the air. The service quality may rely on the wireless performance which is affected by the UAV's position. In this paper, we focus on the UAV placement problem in the Internet of Vehicles, where the UAV is deployed to monitor the road traffic and sends the monitored videos to vehicles. The studied problem is formulated as video resolution maximization by optimizing over the UAV's position. Moreover, we take into account the maximal transmission delay and impose a probabilistic constraint. To solve the formulated problem, we first leverage the techniques in extreme value theory (EVT) and Gaussian process regression (GPR) to characterize the influence of the UAV's position on the delay performance. Based on this characterization, we subsequently propose a proactive resolution selection and UAV placement approach, which adaptively places the UAV according to the geographic distribution of vehicles. Numerical results justify the joint usage of EVT and GPR for maximal delay characterization. Through investigating the maximal transmission delay, the proposed approach nearly achieves the optimal performance when vehicles are evenly distributed, and reduces 10% and 19% of the 999-th 1000-quantile over two baselines when vehicles are biased distributed., Comment: Accepted in the IEEE Transactions on Intelligent Transportation Systems

Tuning magnetic properties of magnetic topological materials is of interest to realize elusive physical phenomena such as quantum anomalous hall effect (QAHE) at higher temperatures and design topological spintronic devices. However, current topological materials exhibit Curie temperature (TC) values far below room temperature. In recent years, significant progress has been made to control and optimize TC, particularly through defect engineering of these structures. Most recently we showed evidence of TC values up to 80K for (MnSb2Te4)x(Sb2Te3)1-x, where x is greater than or equal to 0.7 and less than or equal to 0.85, by controlling the compositions and Mn content in these structures. Here we show further enhancement of the TC, as high as 100K, by maintaining high Mn content and reducing the growth rate from 0.9 nm/min to 0.5 nm/min. Derivative curves reveal the presence of two TC components contributing to the overall value and propose TC1 and TC2 have distinct origins: excess Mn in SLs and Mn in Sb2-yMnyTe3QLs alloys, respectively. In pursuit of elucidating the mechanisms promoting higher Curie temperature values in this system, we show evidence of structural disorder where Mn is occupying not only Sb sites but also Te sites, providing evidence of significant excess Mn and a new crystal structure:(Mn1+ySb2-yTe4)x(Sb2-yMnyTe3)1-x. Our work shows progress in understanding how to control magnetic defects to enhance desired magnetic properties and the mechanism promoting these high TC in magnetic topological materials such as (Mn1+ySb2-yTe4)x(Sb2-yMnyTe3)1-x.

Since their inception, embeddings have become a primary ingredient in many flavours of Natural Language Processing (NLP) tasks supplanting earlier types of representation. Even though multilingual embeddings have been used for the increasing number of multilingual tasks, due to the scarcity of parallel training data, low-resource languages such as Sinhala, tend to focus more on monolingual embeddings. Then when it comes to the aforementioned multi-lingual tasks, it is challenging to utilize these monolingual embeddings given that even if the embedding spaces have a similar geometric arrangement due to an identical training process, the embeddings of the languages considered are not aligned. This is solved by the embedding alignment task. Even in this, high-resource language pairs are in the limelight while low-resource languages such as Sinhala which is in dire need of help seem to have fallen by the wayside. In this paper, we try to align Sinhala and English word embedding spaces based on available alignment techniques and introduce a benchmark for Sinhala language embedding alignment. In addition to that, to facilitate the supervised alignment, as an intermediate task, we also introduce Sinhala-English alignment datasets. These datasets serve as our anchor datasets for supervised word embedding alignment. Even though we do not obtain results comparable to the high-resource languages such as French, German, or Chinese, we believe our work lays the groundwork for more specialized alignment between English and Sinhala embeddings.

Low-symmetry van der Waals materials have enabled strong confinement of mid-infrared light through hyperbolic phonon polaritons (HPhPs) at the nanoscale. Yet, the bottleneck persists in manipulating the intrinsic polaritonic dispersion to drive further progress in phonon-polaritonics. Here, we present a thermomechanical strategy to manipulate the HPhPs in $\alpha$-MoO$_3$ using high pressure and temperature treatment. The hot pressing engineers the stoichiometry of $\alpha$-MoO$_3$ by controllably introducing oxygen vacancy defects (OVDs), which cause a semiconductor-to-semimetal transition. Our density functional theory and finite-difference time-domain results, combined with experimental studies show that the OVDs induce a metastable metallic state by reducing the bandgap while modifying the intrinsic dielectric permittivity of $\alpha$-MoO$_3$. Photo-induced force microscopy confirms an average dielectric permittivity tunability of $|\Delta\varepsilon/\varepsilon|\approx0.35$ within a Reststrahlen band of $\alpha$-MoO$_3$, resulting in drastic shifts in the HPhP dispersion. The polariton lifetimes for pristine and hot-pressed flakes were measured as $0.92 \pm 0.06$ and $0.86 \pm 0.11$ ps, respectively, exhibiting a loss of only 7%, while the group velocity exhibited an increase of $38.8 \pm 0.2$%. The OVDs in $\alpha$-MoO$_3$ provide a low-loss platform that enables active tuning of mid-infrared HPhPs and have a profound impact on applications in super-resolution imaging, nanoscale thermal manipulation, boosted molecular sensing, and on-chip photonic circuits., Comment: 42 pages, 13 figures, supporting information

White dwarfs, the extremely dense remnants left behind by most stars after their death, are characterised by a mass comparable to that of the Sun compressed into the size of an Earth-like planet. In the resulting strong gravity, heavy elements sink toward the centre and the upper layer of the atmosphere contains only the lightest element present, usually hydrogen or helium. Several mechanisms compete with gravitational settling to change a white dwarf's surface composition as it cools, and the fraction of white dwarfs with helium atmospheres is known to increase by a factor ~2.5 below a temperature of about 30,000 K; therefore, some white dwarfs that appear to have hydrogen-dominated atmospheres above 30,000 K are bound to transition to be helium-dominated as they cool below it. Here we report observations of ZTF J203349.8+322901.1, a transitioning white dwarf with two faces: one side of its atmosphere is dominated by hydrogen and the other one by helium. This peculiar nature is likely caused by the presence of a small magnetic field, which creates an inhomogeneity in temperature, pressure or mixing strength over the surface. ZTF J203349.8+322901.1 might be the most extreme member of a class of magnetic, transitioning white dwarfs -- together with GD 323, a white dwarf that shows similar but much more subtle variations. This new class could help shed light on the physical mechanisms behind white dwarf spectral evolution., Comment: 45 pages, 11 figures

Parallel datasets are vital for performing and evaluating any kind of multilingual task. However, in the cases where one of the considered language pairs is a low-resource language, the existing top-down parallel data such as corpora are lacking in both tally and quality due to the dearth of human annotation. Therefore, for low-resource languages, it is more feasible to move in the bottom-up direction where finer granular pairs such as dictionary datasets are developed first. They may then be used for mid-level tasks such as supervised multilingual word embedding alignment. These in turn can later guide higher-level tasks in the order of aligning sentence or paragraph text corpora used for Machine Translation (MT). Even though more approachable than generating and aligning a massive corpus for a low-resource language, for the same reason of apathy from larger research entities, even these finer granular data sets are lacking for some low-resource languages. We have observed that there is no free and open dictionary data set for the low-resource language, Sinhala. Thus, in this work, we introduce three parallel English-Sinhala word dictionaries (En-Si-dict-large, En-Si-dict-filtered, En-Si-dict-FastText) which help in multilingual Natural Language Processing (NLP) tasks related to English and Sinhala languages. In this paper, we explain the dataset creation pipeline as well as the experimental results of the tests we have carried out to verify the quality of the data sets. The data sets and the related scripts are available at https://github.com/kasunw22/sinhala-para-dict.

This paper introduces an approach to decoupling singularly perturbed boundary value problems for fourth-order ordinary differential equations that feature a small positive parameter $\epsilon$ multiplying the highest derivative. We specifically examine Lidstone boundary conditions and demonstrate how to break down fourth-order differential equations into a system of second-order problems, with one lacking the parameter and the other featuring $\epsilon$ multiplying the highest derivative. To solve this system, we propose a mixed finite element algorithm and incorporate the Shishkin mesh scheme to capture the solution near boundary layers. Our solver is both direct and of high accuracy, with computation time that scales linearly with the number of grid points. We present numerical results to validate the theoretical results and the accuracy of our method., Comment: 15 pages, 7 figures

Efficient resource allocation (RA) strategies within massive and dense Internet of Things (IoT) networks is one of the major challenges in the deployment of IoT-network based smart ecosystems involving heterogeneous power-constrained IoT devices operating in varied radio and environmental conditions. In this paper, we focus on the transmit power minimization problem for IoT devices while maintaining a threshold channel throughput. The established optimization literature is not robust against the fast-fading channel and the interaction among different transmit signals in each instance. Besides, realistically, each IoT node possesses incomplete channel state information (CSI) on its neighbors, such as the channel gain being private information for the node itself. In this work, we resort to Bayesian game theoretic strategies for solving the transmit power optimization problem exploiting incomplete CSIs within massive IoT networks. We provide a steady discussion on the rationale for selecting the game theory, particularly the Bayesian scheme, with a graphical visualization of our formulated problem. We take advantage of the property of the existence and uniqueness of the Bayesian Nash equilibrium (BNE), which exhibits reduced computational complexity while optimizing transmit power and maintaining target throughput within networks comprised of heterogeneous devices.

In disaster-stricken environments, it's vital to assess the damage quickly, analyse the stability of the environment, and allocate resources to the most vulnerable areas where victims might be present. These missions are difficult and dangerous to be conducted directly by humans. Using the complementary capabilities of both the ground and aerial robots, we investigate a collaborative approach of aerial and ground robots to address this problem. With an increased field of view, faster speed, and compact size, the aerial robot explores the area and creates a 3D feature-based map graph of the environment while providing a live video stream to the ground control station. Once the aerial robot finishes the exploration run, the ground control station processes the map and sends it to the ground robot. The ground robot, with its higher operation time, static stability, payload delivery and tele-conference capabilities, can then autonomously navigate to identified high-vulnerability locations. We have conducted experiments using a quadcopter and a hexapod robot in an indoor modelled environment with obstacles and uneven ground. Additionally, we have developed a low-cost drone add-on with value-added capabilities, such as victim detection, that can be attached to an off-the-shelf drone. The system was assessed for cost-effectiveness, energy efficiency, and scalability.

Three-dimensional topological insulators (3D-TIs) are a new generation of materials with insulating bulk and exotic metallic surface states that facilitate a wide variety of ground-breaking applications. However, utilization of the surface channels is often hampered by the presence of crystal defects, such as antisites, vacancies and twin domains. For terahertz device applications, twinning is shown to be highly deleterious. Previous attempts to reduce twins using technologically important InP(111) substrates have been promising, but have failed to completely suppress twin domains while preserving high structural quality. Here we report growth of twin-free molecular beam epitaxial Bi2Se3 and Sb2Te3 structures on ultra-thin In2Se3 layers formed by a novel selenium passivation technique during the oxide desorption of smooth, non-vicinal InP(111)B substrates, without the use of an indium source. The formation of un-twinned In2Se3 provides a favorable template to fully suppress twin domains in 3D-TIs, greatly broadening novel device applications in the terahertz regime., Comment: Main manuscript has 27 pages and 7 figures. Supplementary information has 3 pages and 3 figures

An intervening galaxy acts as a gravitational lens and produces multiple images of a single source such as a remote galaxy. Galaxies have peculiar speeds in addition to the bulk motion arising due to the expansion of the universe. There is a difference in light arrival times between lensed images. We calculate more realistic time delays between lensed images when galaxy peculiar motions, that is the motion of the Lens, the Source and the Observer are taken into consideration neglecting the gravitomagnetic effects., Comment: 5 pages, 9 figures, Submitted to the MNRAS

Recent year has brought considerable advancements in Electric Vehicles (EVs) and associated infrastructures/communications. Intrusion Detection Systems (IDS) are widely deployed for anomaly detection in such critical infrastructures. This paper presents an Interpretable Anomaly Detection System (RX-ADS) for intrusion detection in CAN protocol communication in EVs. Contributions include: 1) window based feature extraction method; 2) deep Autoencoder based anomaly detection method; and 3) adversarial machine learning based explanation generation methodology. The presented approach was tested on two benchmark CAN datasets: OTIDS and Car Hacking. The anomaly detection performance of RX-ADS was compared against the state-of-the-art approaches on these datasets: HIDS and GIDS. The RX-ADS approach presented performance comparable to the HIDS approach (OTIDS dataset) and has outperformed HIDS and GIDS approaches (Car Hacking dataset). Further, the proposed approach was able to generate explanations for detected abnormal behaviors arising from various intrusions. These explanations were later validated by information used by domain experts to detect anomalies. Other advantages of RX-ADS include: 1) the method can be trained on unlabeled data; 2) explanations help experts in understanding anomalies and root course analysis, and also help with AI model debugging and diagnostics, ultimately improving user trust in AI systems.

In this paper, we study the biharmonic equation with Dirichlet boundary conditions in a polygonal domain. In particular, we propose a method that effectively decouples the fourth-order problem into a system of two Poison equations and one Stokes equation, or a system of one Stokes equation and one Poisson equation. It is shown that the solution of each system is equivalent to that of the original fourth-order problem on both convex and non-convex polygonal domains. Two finite element algorithms are in turn proposed to solve the decoupled systems. In addition, we show the regularity of the solutions in each decoupled system in both the Sobolev space and the weighted Sobolev space, and we derive the optimal error estimates for the numerical solutions on both quasi-uniform meshes and graded meshes. Numerical test results are presented to justify the theoretical findings., Comment: 46 pages, 9 figures, 17 tables

The human annotations are imperfect, especially when produced by junior practitioners. Multi-expert consensus is usually regarded as golden standard, while this annotation protocol is too expensive to implement in many real-world projects. In this study, we propose a method to refine human annotation, named Neural Annotation Refinement (NeAR). It is based on a learnable implicit function, which decodes a latent vector into represented shape. By integrating the appearance as an input of implicit functions, the appearance-aware NeAR fixes the annotation artefacts. Our method is demonstrated on the application of adrenal gland analysis. We first show that the NeAR can repair distorted golden standards on a public adrenal gland segmentation dataset. Besides, we develop a new Adrenal gLand ANalysis (ALAN) dataset with the proposed NeAR, where each case consists of a 3D shape of adrenal gland and its diagnosis label (normal vs. abnormal) assigned by experts. We show that models trained on the shapes repaired by the NeAR can diagnose adrenal glands better than the original ones. The ALAN dataset will be open-source, with 1,584 shapes for adrenal gland diagnosis, which serves as a new benchmark for medical shape analysis. Code and dataset are available at https://github.com/M3DV/NeAR., Comment: MICCAI 2022

Shape priors have long been known to be effective when reconstructing 3D shapes from noisy or incomplete data. When using a deep-learning based shape representation, this often involves learning a latent representation, which can be either in the form of a single global vector or of multiple local ones. The latter allows more flexibility but is prone to overfitting. In this paper, we advocate a hybrid approach representing shapes in terms of 3D meshes with a separate latent vector at each vertex. During training the latent vectors are constrained to have the same value, which avoids overfitting. For inference, the latent vectors are updated independently while imposing spatial regularization constraints. We show that this gives us both flexibility and generalization capabilities, which we demonstrate on several medical image processing tasks., Comment: 14 pages, 9 figures, submitted for review

In recent hyperspectral unmixing (HU) literature, the application of deep learning (DL) has become more prominent, especially with the autoencoder (AE) architecture. We propose a split architecture and use a pseudo-ground truth for abundances to guide the `unmixing network' (UN) optimization. Preceding the UN, an `approximation network' (AN) is proposed, which will improve the association between the centre pixel and its neighbourhood. Hence, it will accentuate spatial correlation in the abundances as its output is the input to the UN and the reference for the `mixing network' (MN). In the Guided Encoder-Decoder Architecture for Hyperspectral Unmixing with Spatial Smoothness (GAUSS), we proposed using one-hot encoded abundances as the pseudo-ground truth to guide the UN; computed using the k-means algorithm to exclude the use of prior HU methods. Furthermore, we release the single-layer constraint on MN by introducing the UN generated abundances in contrast to the standard AE for HU. Secondly, we experimented with two modifications on the pre-trained network using the GAUSS method. In GAUSS$_\textit{blind}$, we have concatenated the UN and the MN to back-propagate the reconstruction error gradients to the encoder. Then, in the GAUSS$_\textit{prime}$, abundance results of a signal processing (SP) method with reliable abundance results were used as the pseudo-ground truth with the GAUSS architecture. According to quantitative and graphical results for four experimental datasets, the three architectures either transcended or equated the performance of existing HU algorithms from both DL and SP domains., Comment: 16 pages, 6 figures

Monitoring traffic in computer networks is one of the core approaches for defending critical infrastructure against cyber attacks. Machine Learning (ML) and Deep Neural Networks (DNNs) have been proposed in the past as a tool to identify anomalies in computer networks. Although detecting these anomalies provides an indication of an attack, just detecting an anomaly is not enough information for a user to understand the anomaly. The black-box nature of off-the-shelf ML models prevents extracting important information that is fundamental to isolate the source of the fault/attack and take corrective measures. In this paper, we introduce the Network Transformer (NeT), a DNN model for anomaly detection that incorporates the graph structure of the communication network in order to improve interpretability. The presented approach has the following advantages: 1) enhanced interpretability by incorporating the graph structure of computer networks; 2) provides a hierarchical set of features that enables analysis at different levels of granularity; 3) self-supervised training that does not require labeled data. The presented approach was tested by evaluating the successful detection of anomalies in an Industrial Control System (ICS). The presented approach successfully identified anomalies, the devices affected, and the specific connections causing the anomalies, providing a data-driven hierarchical approach to analyze the behavior of a cyber network.

Hundreds of thousands of malicious domains are created everyday. These malicious domains are hosted on a wide variety of network infrastructures. Traditionally, attackers utilize bullet proof hosting services (e.g. MaxiDed, Cyber Bunker) to take advantage of relatively lenient policies on what content they can host. However, these IP ranges are increasingly being blocked or the services are taken down by law enforcement. Hence, attackers are moving towards utilizing IPs from regular hosting providers while staying under the radar of these hosting providers. There are several practical advantages of accurately knowing the type of IP used to host malicious domains. If the IP is a dedicated IP (i.e. it is leased to a single entity), one may blacklist the IP to block domains hosted on those IPs as welll as use as a way to identify other malicious domains hosted the same IP. If the IP is a shared hosting IP, hosting providers may take measures to clean up such domains and maintain a high reputation for their users.

Modern optical nano-elements pursue ever-smaller sizes and individualized functionalities. Those elements that can efficiently manipulate the magnetic field of light boast promising future applications with a great challenge: the magnetic near field is irretrievable from conventional optical far-field characterization. Here we propose photoinduced magnetic force microscopy to directly and exclusively sense the magnetic field of light at the nanoscale. The proposed instrument exploits a magnetic nanoprobe with exclusive magnetic excitation under structured light illumination. The magnetic nanoprobe detects the photoinduced magnetic force, which is defined as the dipolar Lorentz force exerted on the photoinduced magnetic dipole in the nanoprobe. Since the resulting magnetic force is proportional to the incident magnetic field, the measured force reveals the magnetic near-field distribution at the nanoscale. The proposed instrument represents a fundamental step towards comprehensive electric and magnetic near-field detection and/or manipulation in single nano-element optical devices.

There are many approaches to weakly-supervised training of networks to segment 2D images. By contrast, existing approaches to segmenting volumetric images rely on full-supervision of a subset of 2D slices of the 3D volume. We propose an approach to volume segmentation that is truly weakly-supervised in the sense that we only need to provide a sparse set of 3D points on the surface of target objects instead of detailed 2D masks. We use the 3D points to deform a 3D template so that it roughly matches the target object outlines and we introduce an architecture that exploits the supervision it provides to train a network to find accurate boundaries. We evaluate our approach on Computed Tomography (CT), Magnetic Resonance Imagery (MRI) and Electron Microscopy (EM) image datasets and show that it substantially reduces the required amount of effort., Comment: 12 Pages

A decade on from the devastating Black Saturday bushfires in Victoria, Australia, we are at a point where computer simulations of community evacuations are starting to be used within the emergency services. While fire progression modelling is embedded in strategic and operational settings at all levels of government across Victoria, modelling of community response to such fires is only just starting to be evaluated in earnest. For community response models to become integral to bushfire planning and preparedness, the key question to be addressed is: when faced with a bushfire, what will a community really do? Typically this understanding has come from local experience and expertise within the community and services, however the trend is to move towards more informed data driven approaches. In this paper we report on the latest work within the emergency sector in this space. Particularly, we discuss the application of Strahan et al.'s self-evacuation archetypes to an agent-based model of community evacuation in regional Victoria. This work is part of the consolidated bushfire evacuation modelling collaboration between several emergency management stakeholders.

A main challenge in target localization arises from the lack of reliable distance measures. This issue is especially pronounced in indoor settings due to the presence of walls, floors, furniture, and other dynamically changing conditions such as the movement of people and goods, varying temperature, and airflows. Here, we develop a new computational framework to estimate the location of a target without the need for reliable distance measures. The method, which we term Ordinal UNLOC, uses only ordinal data obtained from comparing the signal strength from anchor pairs at known locations to the target. Our estimation technique utilizes rank aggregation, function learning as well as proximity-based unfolding optimization. As a result, it yields accurate target localization for common transmission models with unknown parameters and noisy observations that are reminiscent of practical settings. Our results are validated by both numerical simulations and hardware experiments., Comment: This manuscript has been accepted by the IEEE Internet of Things Journal

Multidimensional scaling of gene sequence data has long played a vital role in analysing gene sequence data to identify clusters and patterns. However the computation complexities and memory requirements of state-of-the-art dimensional scaling algorithms make it infeasible to scale to large datasets. In this paper we present an autoencoder-based dimensional reduction model which can easily scale to datasets containing millions of gene sequences, while attaining results comparable to state-of-the-art MDS algorithms with minimal resource requirements. The model also supports out-of-sample data points with a 99.5%+ accuracy based on our experiments. The proposed model is evaluated against DAMDS with a real world fungi gene sequence dataset. The presented results showcase the effectiveness of the autoencoder-based dimension reduction model and its advantages.

Qubits on solid state devices could potentially provide the rapid control necessary for developing scalable quantum information processors. Materials innovation and design breakthroughs have increased functionality and coherence of qubits substantially over the past two decades. Here we show by improving interface between InAs as a semiconductor and Al as a superconductor, one can reliably fabricate voltage-controlled Josephson junction field effect transistor (JJ-FET) that can be used as tunable qubits, resonators, and coupler switches. We find that band gap engineering is crucial in realizing a two-dimensional electron gas near the surface. In addition, we show how the coupling between the semiconductor layer and the superconducting contacts can affect qubit properties. We present the anharmonicity and coupling strengths from one and two-photon absorption in a quantum two level system fabricated with a JJ-FET., Comment: This paper is part of the Special Topic Collection: Honoring Dr. Art Gossard's 85th Birthday and his Leadership in the Science and Technology of Molecular Beam Epitaxy

Diagnosing pre-existing heart diseases early in life is important as it helps prevent complications such as pulmonary hypertension, heart rhythm problems, blood clots, heart failure and sudden cardiac arrest. To identify such diseases, phonocardiogram (PCG) and electrocardiogram (ECG) waveforms convey important information. Therefore, effectively using these two modalities of data has the potential to improve the disease screening process. We evaluate this hypothesis on a subset of the PhysioNet Challenge 2016 Dataset which contains simultaneously acquired PCG and ECG recordings. Our novel Dual-Convolutional Neural Network based approach uses transfer learning to tackle the problem of having limited amounts of simultaneous PCG and ECG data that is publicly available, while having the potential to adapt to larger datasets. In addition, we introduce two main evaluation frameworks named record-wise and sample-wise evaluation which leads to a rich performance evaluation for the transfer learning approach. Comparisons with methods which used single or dual modality data show that our method can lead to better performance. Furthermore, our results show that individually collected ECG or PCG waveforms are able to provide transferable features which could effectively help to make use of a limited number of synchronized PCG and ECG waveforms and still achieve significant classification performance., Comment: Paper accepted to IEEE International Symposium on Circuits and Systems (ISCAS) 2021

Active Surface Models have a long history of being useful to model complex 3D surfaces but only Active Contours have been used in conjunction with deep networks, and then only to produce the data term as well as meta-parameter maps controlling them. In this paper, we advocate a much tighter integration. We introduce layers that implement them that can be integrated seamlessly into Graph Convolutional Networks to enforce sophisticated smoothness priors at an acceptable computational cost. We will show that the resulting Deep Active Surface Models outperform equivalent architectures that use traditional regularization loss terms to impose smoothness priors for 3D surface reconstruction from 2D images and for 3D volume segmentation., Comment: 11 pages, 7 figures, 6 tables

In the current era of Big Data, data engineering has transformed into an essential field of study across many branches of science. Advancements in Artificial Intelligence (AI) have broadened the scope of data engineering and opened up new applications in both enterprise and research communities. Aggregations (also termed reduce in functional programming) are an integral functionality in these applications. They are traditionally aimed at generating meaningful information on large data-sets, and today, they are being used for engineering more effective features for complex AI models. Aggregations are usually carried out on top of data abstractions such as tables/ arrays and are combined with other operations such as grouping of values. There are frameworks that excel in the said domains individually. But, we believe that there is an essential requirement for a data analytics tool that can universally integrate with existing frameworks, and thereby increase the productivity and efficiency of the entire data analytics pipeline. Cylon endeavors to fulfill this void. In this paper, we present Cylon's fast and scalable aggregation operations implemented on top of a distributed in-memory table structure that universally integrates with existing frameworks.

Data engineering is becoming an increasingly important part of scientific discoveries with the adoption of deep learning and machine learning. Data engineering deals with a variety of data formats, storage, data extraction, transformation, and data movements. One goal of data engineering is to transform data from original data to vector/matrix/tensor formats accepted by deep learning and machine learning applications. There are many structures such as tables, graphs, and trees to represent data in these data engineering phases. Among them, tables are a versatile and commonly used format to load and process data. In this paper, we present a distributed Python API based on table abstraction for representing and processing data. Unlike existing state-of-the-art data engineering tools written purely in Python, our solution adopts high performance compute kernels in C++, with an in-memory table representation with Cython-based Python bindings. In the core system, we use MPI for distributed memory computations with a data-parallel approach for processing large datasets in HPC clusters., Comment: 9 pages, 11 images, Accepted in 9th Workshop on Python for High-Performance and Scientific Computing (In conjunction with Supercomputing 20)

Surface electromyography (sEMG) is a popular bio-signal used for controlling prostheses and finger gesture recognition mechanisms. Myoelectric prostheses are costly, and most commercially available sEMG acquisition systems are not suitable for real-time gesture recognition. In this paper, a method of acquiring sEMG signals using novel low-cost, active, dry-contact, flexible sensors has been proposed. Since the active sEMG sensor was developed to be used along with a bionic arm, the sensor was tested for its ability to acquire sEMG signals that could be used for real-time classification of five selected gestures. In a study of 4 subjects, the average classification accuracy for real-time gesture classification using the active sEMG sensor system was 85%. The common-mode rejection ratio of the sensor was measured to 59 dB, and thus the sensor's performance was not substantially limited by its active circuitry. The proposed sensors can be interfaced with a variety of amplifiers to perform fully wearable sEMG acquisition. This satisfies the need for a low-cost sEMG acquisition system for prostheses., Comment: Paper accepted to IEEE International Conference on Systems, Man, and Cybernetics (SMC) 2020

We present work on creating a synthetic population from census data for Australia, applied to the greater Melbourne region. We use a sample-free approach to population synthesis that does not rely on a disaggregate sample from the original population. The inputs for our algorithm are joint marginal distributions from census of desired person-level and household-level attributes, and outputs are a set of comma-separated-value (.csv) files containing the full synthetic population of unique individuals in households; with age, gender, relationship status, household type, and size, matched to census data. Our algorithm is efficient in that it can create the synthetic population for Melbourne comprising 4.5 million persons in 1.8 million households within three minutes on a modern computer. Code for the algorithm is hosted on GitHub., Comment: 15 pages, 1 figure with 3 sub-figures. International Workshop on Agent-Based Modelling of Urban Systems (ABMUS, 2019)

The transparent interface in epitaxial Al-InAs heterostructures provides an excellent platform for potential advances in mesoscopic and topological superconductivity. Semiconductor-based Josephson Junction Field Effect Transistors (JJ-FETs) fabricated on these heterostructures have a metallic gate that tunes the supercurrent. Here we report the fabrication and measurement of gate-tunable Al-InAs JJ-FETs in which the gate dielectric in contact with the InAs is produced by mechanically exfoliated hexagonal boron nitride (h-BN) followed by dry transfer using a van der Waals-mediated pick up process. We discuss the fabrication process that enables compatibility between layered material transfer and Al-InAs heterostructures to avoid chemical reactions and unintentional doping that could affect the characteristics of the JJ-FET. We achieve full gate-tunablity of supercurrent by using only 5~nm thick h-BN flakes. We contrast our process with devices fabricated using a conventional AlO$_{\rm x}$ gate dielectric and show that h-BN could be an excellent competing dielectric for JJ-FET devices. We observe that the product of normal resistance and critical current, I$_{\rm c}$R$_{\rm n}$, is comparable for both types of devices, but strikingly higher R$_{\rm n}$ for the h-BN-based devices indicating that the surface is doped less compared to AlO$_{\rm x}$ gate dielectric.

The amazing advances being made in the fields of machine and deep learning are a highlight of the Big Data era for both enterprise and research communities. Modern applications require resources beyond a single node's ability to provide. However this is just a small part of the issues facing the overall data processing environment, which must also support a raft of data engineering for pre- and post-data processing, communication, and system integration. An important requirement of data analytics tools is to be able to easily integrate with existing frameworks in a multitude of languages, thereby increasing user productivity and efficiency. All this demands an efficient and highly distributed integrated approach for data processing, yet many of today's popular data analytics tools are unable to satisfy all these requirements at the same time. In this paper we present Cylon, an open-source high performance distributed data processing library that can be seamlessly integrated with existing Big Data and AI/ML frameworks. It is developed with a flexible C++ core on top of a compact data structure and exposes language bindings to C++, Java, and Python. We discuss Cylon's architecture in detail, and reveal how it can be imported as a library to existing applications or operate as a standalone framework. Initial experiments show that Cylon enhances popular tools such as Apache Spark and Dask with major performance improvements for key operations and better component linkages. Finally, we show how its design enables Cylon to be used cross-platform with minimum overhead, which includes popular AI tools such as PyTorch, Tensorflow, and Jupyter notebooks.

An instrument has been developed for precision controlled exposures of electronic devices and material samples in particle beams. The instrument provides simultaneously a real time record of the profile of the beam and the fluence received. The system is capable of treating devices with dimensional scales ranging from millimeters to extended objects of cross sections measured in tens of square centimeters. The instrument has been demonstrated to operate effectively in integrated fluences ranging up to a few times $10^{16}$ 1-MeV-neutron-equivalent/cm$^{2}$ (n$_{\textrm{eq}}$). The positioner portion of the system comprises a set of remotely controllable sample holders incorporating cooling and interfaces for sample power and readout, all constructed from low activation technologies. The monitoring component of the system samples the current or voltage of radiation tolerant silicon diodes placed directly in the path of the beam.

Josephson junctions hosting Majorana fermions have been predicted to exhibit a 4$\pi$ periodic current phase relation. The experimental consequence of this periodicity is the disappearance of odd steps in Shapiro steps experiments. Experimentally, missing odd Shapiro steps have been observed in a number of materials systems with strong spin-orbit coupling and have been interpreted in the context of topological superconductivity. Here, we report on missing odd steps in topologically trivial Josephson junctions fabricated on InAs quantum wells. We ascribe our observations to the high transparency of our junctions allowing Landau-Zener transitions. The probability of these processes is found to be independent of the drive frequency. We analyze our results using a bi-modal transparency distribution which demonstrates that only few modes carrying 4$\pi$ periodic current are sufficient to describe the disappearance of odd steps. Our findings highlight the elaborate circumstances that have to be considered in the investigation of the 4$\pi$ Josephson junctions in relationship to topological superconductivity.

A series of astronomical observations obtained over the period 1986 to 2018 supports the idea that life is a cosmic rather than a purely terrestrial or planetary phenomenon. These include (1) the detection of biologically relevant molecules in interstellar clouds and in comets, (2) mid-infrared spectra of interstellar grains and the dust from comets, (3) a diverse set of data from comets including the Rosetta mission showing consistency with biology and (4) the frequency of Earth-like or habitable planets in the Galaxy. We argue that the conjunction of all the available data suggests the operation of cometary biology and interstellar panspermia rather than the much weaker hypothesis of comets being only the source of the chemical building blocks of life. We conclude with specific predictions on the properties expected of extra-terrestrial life if it is discovered on Enceladus, Europa or beyond. A radically different biochemistry elsewhere can be considered as a falsification of the theory of interstellar panspermia., Comment: 13 pages, 5 figures, submitted to PASA

Accurate and real-time hand gesture recognition is essential for controlling advanced hand prostheses. Surface Electromyography (sEMG) signals obtained from the forearm are widely used for this purpose. Here, we introduce a novel hand gesture representation called Temporal Muscle Activation (TMA) maps which captures information about the activation patterns of muscles in the forearm. Based on these maps, we propose an algorithm that can recognize hand gestures in real-time using a Convolution Neural Network. The algorithm was tested on 8 healthy subjects with sEMG signals acquired from 8 electrodes placed along the circumference of the forearm. The average classification accuracy of the proposed method was 94%, which is comparable to state-of-the-art methods. The average computation time of a prediction was 5.5ms, making the algorithm ideal for the real-time gesture recognition applications., Comment: Paper accepted to IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP) 2020

The magnetic white dwarfs (MWDs) are found either isolated or in interacting binaries. They divide into two groups: a high field group (0.1-1,000MegaGauss) comprising some 13% of all white dwarfs (WDs), and a low field group (B<0.1MG) whose incidence is currently under investigation. The situation may be similar in magnetic binaries because the bright accretion discs in low field systems hide the photosphere of their WDs thus preventing the study of their magnetic fields' strength and structure. Considerable research has been devoted to the vexed question on the origin of magnetic fields. One hypothesis is that WD magnetic fields are of fossil origin. The other is that magnetic fields arise from binary interaction, through differential rotation, during common envelope evolution. The recently discovered population of hot, carbon-rich WDs exhibiting an incidence of magnetism of up to about 70% and a variability from a few minutes to a couple of days may support the merging binary hypothesis. Several studies have raised the possibility of the detection of planets around MWDs. Rocky planets may be discovered by the detection of anomalous atmospheric heating of the MWD. Planetary remains have recently revealed themselves in the atmospheres of about 25% of WDs that are polluted by elements such as Ca, Si, and often also Mg, Fe, Na. This pollution has been explained by ongoing accretion of planetary debris. The study of isolated and accreting MWDs is likely to continue to yield exciting discoveries for many years to come., Comment: Invited, refereed review paper to appear in the COSPAR 2018 special issue of Advances in Space Research (Elsevier) entitled Nova Eruptions, Cataclysmic Variables and related systems: Observational vs. theoretical challenges in the 2020 era, 98 pages, 30 figures

CNN-based volumetric methods that label individual voxels now dominate the field of biomedical segmentation. However, 3D surface representations are often required for proper analysis. They can be obtained by post-processing the labeled volumes which typically introduces artifacts and prevents end-to-end training. In this paper, we therefore introduce a novel architecture that goes directly from 3D image volumes to 3D surfaces without post-processing and with better accuracy than current methods. We evaluate it on Electron Microscopy and MRI brain images as well as CT liver scans. We will show that it outperforms state-of-the-art segmentation methods.