Search

Alzheimer's disease (AD) poses significant global health challenges due to its increasing prevalence and associated societal costs. Early detection and diagnosis of AD are critical for delaying progression and improving patient outcomes. Traditional diagnostic methods and single-modality data often fall short in identifying early-stage AD and distinguishing it from Mild Cognitive Impairment (MCI). This study addresses these challenges by introducing a novel approach: multImodal enseMble via class BALancing diversity for iMbalancEd Data (IMBALMED). IMBALMED integrates multimodal data from the Alzheimer's Disease Neuroimaging Initiative database, including clinical assessments, neuroimaging phenotypes, biospecimen and subject characteristics data. It employs an ensemble of model classifiers, each trained with different class balancing techniques, to overcome class imbalance and enhance model accuracy. We evaluate IMBALMED on two diagnostic tasks (binary and ternary classification) and four binary early detection tasks (at 12, 24, 36, and 48 months), comparing its performance with state-of-the-art algorithms and an unbalanced dataset method. IMBALMED demonstrates superior diagnostic accuracy and predictive performance in both binary and ternary classification tasks, significantly improving early detection of MCI at 48-month time point. The method shows improved classification performance and robustness, offering a promising solution for early detection and management of AD.

In response to escalating global drinking water scarcity we propose an innovative, automatic system for reusing clean sink water to flush toilets. Existing solutions for water recycle in the houses involve purifiers and complex treatments, leading to high costs and constant maintenance. WASD, utilizing sensors and a solenoid valve, rapidly detects and separates clean water, directing it to toilet tank while sending non-reusable water to the drain. This cost-effective and user-friendly approach aims to establish sustainable water practices in domestic settings, contributing to solve the shortage of drinking water.

In this work, we observe plasmon induced hot electron extraction in a heterojunction between indium tin oxide nanocrystals and monolayer molybdenum disulphide. We study the sample with ultrafast differential transmission exciting the sample at 1750 nm where the intense localized plasmon surface resonance of the indium tin oxide nanocrystals is and where the monolayer molybdenum disulphide does not absorb light. With the excitation at 1750 nm we observe the excitonic features of molybdenum disulphide in the visible range, close to the exciton of molybdenum disulphide. Such phenomenon can be ascribed to a charge transfer between indium tin oxide nanocrystals and monolayer molybdenum disulphide upon plasmon excitation. These results are a first step towards the implementation of near infrared plasmonic materials for photoconversion., Comment: 12 pages, 3 figures

NetFlow data is a popular network log format used by many network analysts and researchers. The advantages of using NetFlow over deep packet inspection are that it is easier to collect and process, and it is less privacy intrusive. Many works have used machine learning to detect network attacks using NetFlow data. The first step for these machine learning pipelines is to pre-process the data before it is given to the machine learning algorithm. Many approaches exist to pre-process NetFlow data; however, these simply apply existing methods to the data, not considering the specific properties of network data. We argue that for data originating from software systems, such as NetFlow or software logs, similarities in frequency and contexts of feature values are more important than similarities in the value itself. In this work, we propose an encoding algorithm that directly takes the frequency and the context of the feature values into account when the data is being processed. Different types of network behaviours can be clustered using this encoding, thus aiding the process of detecting anomalies within the network. We train several machine learning models for anomaly detection using the data that has been encoded with our encoding algorithm. We evaluate the effectiveness of our encoding on a new dataset that we created for network attacks on Kubernetes clusters and two well-known public NetFlow datasets. We empirically demonstrate that the machine learning models benefit from using our encoding for anomaly detection., Comment: 11 pages, 17 figures

These days more companies are shifting towards using cloud environments to provide their services to their client. While it is easy to set up a cloud environment, it is equally important to monitor the system's runtime behaviour and identify anomalous behaviours that occur during its operation. In recent years, the utilisation of \ac{rnn} and \ac{dnn} to detect anomalies that might occur during runtime has been a trending approach. However, it is unclear how to explain the decisions made by these networks and how these networks should be interpreted to understand the runtime behaviour that they model. On the contrary, state machine models provide an easier manner to interpret and understand the behaviour that they model. In this work, we propose an approach that learns state machine models to model the runtime behaviour of a cloud environment that runs multiple microservice applications. To the best of our knowledge, this is the first work that tries to apply state machine models to microservice architectures. The state machine model is used to detect the different types of attacks that we launch on the cloud environment. From our experiment results, our approach can detect the attacks very well, achieving a balanced accuracy of 99.2% and an F1 score of 0.982., Comment: 9 pages, 12 figures, workshop paper

We analyze the spectral properties of correlation matrices between distinct statistical systems. Such matrices are intrinsically non symmetric, and lend themselves to extend the spectral analyses usually performed on standard Pearson correlation matrices to the realm of complex eigenvalues. We employ some recent random matrix theory results on the average eigenvalue density of this type of matrices to distinguish between noise and non trivial correlation structures, and we focus on financial data as a case study. Namely, we employ daily prices of stocks belonging to the American and British stock exchanges, and look for the emergence of correlations between two such markets in the eigenvalue spectrum of their non symmetric correlation matrix. We find several non trivial results, also when considering time-lagged correlations over short lags, and we corroborate our findings by additionally studying the asymmetric correlation matrix of the principal components of our datasets., Comment: Revised version; 11 pages, 13 figures

We report on the June 2000 long (100 ks) BeppoSAX exposure that has unveiled above 10 keV a new very high energy component of the X-ray spectrum of Eta Car extending to at least 50 keV. We find that the 2-150 keV spectrum is best reproduced by a thermal + non-thermal model. The thermal component dominates the 2-10 keV spectral range with kT_h=5.5 +/- 0.3 keV and log NH_h=22.68 +/- 0.01. The spectrum displays a prominent iron emission line centred at 6.70 keV. Its equivalent width of 0.94 keV, if produced by the thermal source, gives a slightly sub-solar iron abundance ([Fe/H]=-0.15 +/- 0.02). The high energy tail above 10 keV is best fitted by a power law with a photon index of 2.42 +/- 0.04. The integrated 13-150 keV luminosity of ~12 L-sun is comparable to that of the 2-10 keV thermal component. The present result can be explained, in the Eta Car binary star scenario, by Comptonisation of low frequency radiation by high energy electrons, probably generated in the colliding wind shock front, or in instabilities in the wind of the S Dor primary star. It is possible that the high energy tail had largely weakened near the minimum of the 5.53 yr cycle. With respect to the thermal component, it probably has a longer recovering time like that of the highest excitation optical emission lines. Both features can be associated with the large absorption measured by BeppoSAX at phase 0.05., Comment: 7 pages, 6 figures, accepted for publication in Astronomy and Astrophysics

As a special contribution to the proceedings of the BeppoSAX workshop dedicated to blazar astrophysics we present a catalog of 157 X-ray spectra and the broad-band Spectral Energy Distribution (SED) of 84 blazars observed by BeppoSAX during its first five years of operations. The SEDs have been built by combining BeppoSAX LECS, MECS and PDS data with (mostly) non-simultaneous multi-frequency photometric data, obtained from NED and from other large databases, including the GSC2 and the 2MASS surveys. All BeppoSAX data have been taken from the public archive and have been analysed in a uniform way. For each source we present a SED plot, and for every BeppoSAX observation we give the best fit parameters of the spectral model that best describes the data. The energy where the maximum of the synchrotron power is emitted spans at least six orders of magnitudes ranging from ~ 0.1 eV to over 100 keV. A wide variety of X-ray spectral slopes have been seen depending on whether the synchrotron or inverse Compton component, or both, are present in the X-ray band. The wide energy bandpass of BeppoSAX allowed us to detect, and measure with good accuracy, continuous spectral curvature in many objects whose synchrotron radiation extends to the X-ray band. This convex curvature, which is described by a logarithmic parabola law better than other models, may be the spectral signature of a particle acceleration process that becomes less and less efficient as the particles energy increases. Finally some brief considerations about other statistical properties of the sample are presented., Comment: 36 pages, 26 ps figures, 4 tables - A web page dedicated to this work is available at http://www.asdc.asi.it/blazars/ - To appear in the proceedings of the workshop "Blazar Astrophysics with BeppoSAX and other Observatories" (Frascati, December 2001), Paolo Giommi, Enrico Massaro and Giorgio Palumbo eds. (in press) - giommi@asdc.asi.it

We report here the results of a 90 ks BeppoSAX observation of the low mass X-ray binary and atoll source KS 1731-260 during a quiescent phase. From this observation we derive a source X-ray luminosity of \sim 10^{33} ergs/s (for a source distance of 7 kpc). If the neutron star is spinning at a period of a few milliseconds, as inferred from the nearly-coherent oscillations observed during type-I X-ray bursts, the quiescent X-ray luminosity constrains the neutron star magnetic field strength to less than \sim 10^9 G. If the quiescent emission originates from the release of thermal energy accumulated through deep crust heating in the previous accretion phase, then the measured quiescent luminosity requires a neutron star mass of more than 1.7 M_\odot or a very long outbust recurrence time (\sim 10^3 yr). The former possibility would be in agreement with the expectation of the recycling scenario, namely that the neutron star has accreted a significant amount of mass, enough to spin it up to millisecond periods., Comment: 14 pages, including 1 figure. Submitted to ApJ

The results of a BeppoSAX TOO observation of the BL Lac object Mkn421 during a high intensity state are reported and compared with monitoring X-ray data collected with the BeppoSAX Wide Field Cameras (WFC) and the RXTE All Sky Monitor(ASM). The 0.1-100 keV spectrum of Mkn421 shows continuous convex curvature that can be interpreted as the high-energy end of the synchrotron emission. The source shows significant short-term temporal and spectral variability, which can be interpreted in terms of synchrotron cooling. The comparison of our results with those of previous observations when the source was a factor 3-5 fainter shows evidence for strong spectral variability, with the maximum of the synchrotron power shifting to higher energy during high states. This behaviour suggest an increase in the number of energetic electrons during high states., Comment: 6 pages, LaTeX, plus 8 .ps figures, accepted by MNRAS

The hep-ex data base was decided not to be an appropriate place to make DELPHI notes public. Sorry for the inconvenience., Comment: the paper should not have been made public

Review of prospects for discovery of new physics signals at LEP2. The areas covered include SUSY, exotic fermions, BESS models, leptoquarks, virtual effects and CP violating observables.