Online Machine Research Articles

SDN-Driven Security Framework for DDOS Attack Detection and Mitigation

The rising complexity and persistence of Distributed Denial of Service (DDoS) attacks, particularly low-rate variants, present significant challenges in detection and mitigation within Software-Defined Networking (SDN) environments. Existing detection systems often flood networks with alerts, burdening security personnel and delaying timely mitigation. Furthermore, many solutions are designed and tested in simulated conditions, limiting their real-world applicability. To address these challenges, we propose an SDN-based security framework enhanced with automated monitoring, detection, and mitigation capabilities, optimized for slow-rate DDoS attacks. Our framework was rigorously evaluated on a physical testbed, achieving mitigation efficiencies between 91.66% and 100% under varied attack conditions, thus proving its robustness in practical settings. Additionally, we introduce the SDN-SlowRate-DDoS dataset, designed to assist researchers and industry professionals in developing and testing intrusion detection solutions in more realistic scenarios. To further improve DDoS defense in SDN, we incorporate an ensemble online machine learning model that dynamically adapts to evolving attack patterns, enhancing accuracy across attack types and outperforming traditional models with a detection rate of 99.2% on benchmark datasets. This dual approach leverages both real-world testing and adaptive machine learning, advancing proactive DDoS threat management in SDN environments

Developing A Digital Twin of a TRUEX Extraction Process that Enables Non-proliferation and Safeguards Monitoring through Optical Spectroscopy and Machine Learning

Background The recycling of spent fuel can increase the longevity and decrease the waste footprint associated with the nuclear fuel cycle. However, nuclear fuel reprocessing introduces the risk of proliferation of fissile material. For example, the recycling of uranium (U) and plutonium (Pu) is generally accomplished via a solvent extraction process. The separation efficiency of the process is dependent on a variety of parameters such as acid concentration, organic to aqueous (O/A) volume ratio, number of separation stages, etc. A change in any of these parameters could result in different extraction conditions of fissile material and lead to a process upset with potential safeguards and proliferation concerns. Thus, the ability to monitor the solvent extraction process in real-time would enable the rapid identification and diagnosis of process upsets. Methods In this work, we discuss the development of a digital twin of a surrogate small-scale nuclear fuel reprocessing solvent extraction process utilizing online optical spectroscopic monitoring, machine learning (ML), and chemical modeling. A cascade of centrifugal contactors was employed with the TRUEX (TransUranic Extraction) solvent and surrogate minor actinides. Spectroscopic sensors were placed at strategic locations within the contactor bank and were used to predict TRUEX surrogate analyte concentrations at those locations. Results A solvent extraction digital twin consisting of the data acquisition, chemical model, and anomaly detection, provided expected concentrations at each contactor during normal operations. The digital twin demonstrated the ability to accurately simulate the extraction behavior under ideal conditions. Furthermore, the digital twin’s ML anomaly detection algorithm was shown to be successful in identifying several non-ideal process upset conditions.

A Modal Type Theory of Expected Cost in Higher-Order Probabilistic Programs

The design of online learning algorithms typically aims to optimise the incurred loss or cost , e.g., the number of classification mistakes made by the algorithm. The goal of this paper is to build a type-theoretic framework to prove that a certain algorithm achieves its stated bound on the cost. Online learning algorithms often rely on randomness, their loss functions are often defined as expectations, precise bounds are often non-polynomial (e.g., logarithmic) and proofs of optimality often rely on potential-based arguments. Accordingly, we present pλ-amor, a type-theoretic graded modal framework for analysing (expected) costs of higher-order probabilistic programs with recursion. pλ-amor is an effect-based framework which uses graded modal types to represent potentials, cost and probability at the type level. It extends prior work (λ-amor) on cost analysis for deterministic programs. We prove pλ-amor sound relative to a Kripke step-indexed model which relates potentials with probabilistic coupling. We use pλ-amor to prove cost bounds of several examples from the online machine learning literature. Finally, we describe an extension of pλ-amor with a graded comonad and describe the relationship between the different modalities.

Computational investigation on the anti-tuberculosis potential of novel thiazinane derivatives

ABSTRACT Heterocyclic molecules, containing nitrogen, sulphur, oxygen, etc., have proven their significant roles in drug discovery. Mycobacterium tuberculosis has been one long-lasting disease for several decades and to date this disease, particularly those with multidrug resistance does not have highly efficient and potent therapeutics. The role of the methylcitrate cycle in detoxifying propionyl-CoA, a product formed from the metabolism of host cholesterol by Mtb is well known. Therefore, inhibiting key enzymes or pathways involved in this cycle could potentially influence the survival of the pathogen. Herein a small library of 16 new (Thiazinan-2-ylidene)pyridine-2-amine derivatives is investigated for their potential as anti-tuberculosis inhibitors via various computational methods like docking, dynamic simulations and ADME (absorption, distribution, metabolism and excretion). Further evaluations were carried out with online machine learning tools – mycoCSM and PAINS.

Students’ Multilingual Experiences With Online Machine Translation: Implications for General Language Use

This article discusses how online machine translators (OMT) can influence the way people use general language or languages. The latter distinction has to do with the fact that contemporary lifestyles turn language users into plurilingual communicators. Online shopping, travelling abroad or social media often require at least some degree of plurilingual text mediation skills. Irrespective of foreign language competences, a lot of plurilingual communicators seek OMT support in making sure their messaging with international partners is correct and effective. The article discusses research results concerning how a selected group of BA and MA students in education use machine translation for day-to-day plurilingual communication tasks. We asked them about the relation between machine translation, language learning and language use. We give some insight into the machine translation engine DeepL Translator, which can effectively be used in support of plurilingual text mediation. The article also contains basic educational implications for why and how to make OMT part of foreign language curricula.

A calculator for musculoskeletal injuries prediction in surgeons: a machine learning approach

BackgroundSurgical specialists experience significant musculoskeletal strain as a consequence of their profession, a domain within the healthcare system often recognized for the pronounced impact of such issues. The aim of this study is to calculate the risk of presenting musculoskeletal injuries in surgeons after surgical practice.MethodsCross-sectional study carried out using an online form (12/2021–03/2022) aimed at members of the Spanish Association of Surgeons. Demographic variables on physical and professional activity were recorded, as well as musculoskeletal pain (MSP) associated with surgical activity. Univariate and multivariate analysis were conducted to identify risk factors associated with the development of MSP based on personalized surgical activity. To achieve this, a risk algorithm was computed and an online machine learning calculator was created to predict them. Physiotherapeutic recommendations were generated to address and alleviate each MSP.ResultsA total of 651 surgeons (112 trainees, 539 specialists). 90.6% reported MSP related to surgical practice, 60% needed any therapeutic measure and 11.7% required a medical leave. In the long term, MSP was most common in the cervical and lumbar regions (52.4, 58.5%, respectively). Statistically significant risk factors (OR CI 95%) were for trunk pain, long interventions without breaks (3.02, 1.65–5.54). Obesity, indicated by BMI, to lumbar pain (4.36, 1.84–12.1), while an inappropriate laparoscopic screen location was associated with cervical and trunk pain (1.95, 1.28–2.98 and 2.16, 1.37–3.44, respectively). A predictive model and an online calculator were developed to assess MSP risk. Furthermore, a need for enhanced ergonomics training was identified by 89.6% of surgeons.ConclusionsThe prevalence of MSP among surgeons is a prevalent but often overlooked health concern. Implementing a risk calculator could enable tailored prevention strategies, addressing modifiable factors like ergonomics.

Model predictive control of switched nonlinear systems using online machine learning

This work introduces an online learning-based model predictive control (MPC) approach for the modeling and control of switched nonlinear systems with scheduled mode transitions. Initially, recurrent neural network (RNN) models are constructed offline, utilizing sufficient historical operational data to capture the nominal system dynamics for each mode. Subsequently, we employ real-time process data to develop online learning RNN models, aiming to approximate the dynamics of switched nonlinear systems in the presence of of bounded disturbances. In cases where the initial RNN model is unavailable for a specific switching mode due to very limited historical data, we use real-time data from closed-loop operations under a proportional–integral (PI) controller to build online learning RNN models. To evaluate the predictive performance of online learning RNNs, a theoretical analysis on their generalization error bound is developed using statistical machine learning theory. Additionally, considering the presence or absence of initial RNN models, two MPC schemes are developed. These schemes employ RNNs as prediction models to stabilize switched nonlinear systems, ensuring closed-loop stability by accounting for the generalization error bound derived for online learning RNNs. Finally, the effectiveness of the proposed MPC schemes is demonstrated through a nonlinear process example with two switching modes.

3-D geometric design of microwaveable food products for optimal heating uniformity based on machine learning-supervised multiphysics models

Multiphysics models can assist in geometric design of frozen microwaveable foods, but the conventional 'parametric-sweeping' strategy is computationally intensive. This study develops an online machine learning (ML)-supervised multiphysics modeling strategy to simultaneously optimize multiple geometrical parameters (e.g., top surface width, top surface length, and the ratio of top-to-bottom dimensions) for optimal microwave heating uniformity. First, one or more paired geometric dimensional parameters and heating uniformity results obtained from multiphysics modeling are used as initial training data for the ML optimization process that integrates Gaussian Process Regression (GPR) and Bayesian optimization. Then, the multiphysics modeling procedure is supervised by an ML process to generate new paired geometry-heating uniformity results to expand the training dataset. The loop of multiphysics modeling and ML optimization is conducted to effectively identify geometry designs with good heating uniformity. Results indicate that the ML-supervised optimization strategy can efficiently identify good geometric designs with low heating uniformity by using much fewer multiphysics models than the parametric sweep approach. The ML-supervised approach also exhibits great robustness, delivering good performance even with small (as little as one model) and randomly selected initial training data. This ML-supervised approach is flexible and can be modified to meet specific needs for broad industrial implementation in the development of microwaveable food.

Addressing Data Distribution Shifts in Online Machine Learning Powered Smart City Applications Using Augmented Test-Time Adaptation

Addressing Data Distribution Shifts in Online Machine Learning Powered Smart City Applications Using Augmented Test-Time Adaptation

Using learner corpus data for grammatical accuracy development in written productions

Abstract This study investigates the differential effect of various noticing activities on grammatical accuracy development in EFL learners’ written productions. We focus on different types of noticing activities based on an error-tagged learner corpus and report on effective practical experiments involving learner corpus data. A pretest/posttest quasi-experimental design is used with three experimental groups (receiving different treatments) and one control group. Error frequencies, at both group and individual levels, and proportions of learners producing errors on three specific error types (articles, verb tense, verb agreement) are compared. Our results suggest that accuracy in the use of articles and verb agreement could be more easily fostered through the comparison of learner output with native data (the BNC, in our case). As for verb tenses, the impact of a more traditional form of corrective feedback seems greater while the use of online machine translation tools does not seem to foster much accuracy development.

Assortment Optimization: An Annotated Reading Assortment

Which varieties or brands of a product should a retailer stock on its shelf? Carrying a large variety caters to more customers' needs, but could cannibalize the sales of high-end brands and also cause an inventory nightmare. Assortment optimization aims to formalize these tradeoffs, with the basic problem being as follows. There is a universe of brands j ∈ U , each with a market-accepted price r j . For any S ⊆ U , a function ϕ ( j, S ) indicates the probability that a representative customer from the population would purchase j when given the choice from assortment S , under the market prices. The optimization problem is to maximize the average revenue per customer, i.e. [EQUATION] (1) possibly with constraints on S due to shelf size. Assortment optimization started out by showing how to efficiently find the optimal S from the exponentially many possibilities, under well-established parametric forms for the function ϕ that are called ( discrete) choice models. Since then, the literature has developed choice models of its own that are specialized for assortment optimization. The basic problem has also been extended, and connected with topics such as online algorithms, machine learning, and mechanism design that are mainstream in the Economics and Computation community, with a vast horizon for future directions. This is an annotated reading list about assortment optimization, that aims to provide broad coverage while facing a "cardinality constraint" on the number of papers in the assortment.

Lightweight Distributed Gaussian Process Regression for Online Machine Learning

Lightweight Distributed Gaussian Process Regression for Online Machine Learning

Online Auction System Using AI

Net has pushed the globalization which addresses the interaction and integration many of people ,one-of-a-kind business institutes, government our bodies, and lots of more. As people are exposed to unlimited variety of quantitative products through use of internet, they seek for the predicted one at affordable or favorable fee and time. On-line biding has become distinguished way to the expectations of on-line customers since it excludes the want of physical presence at the bidding location and the product can be acquired on the low- priced price. When bidding is performed on a bidding floor this machine will permit users to get right of entry to the bidding the usage oa an internet portal. Images can also be considered. Online bidders can vicinity bids at any moment and their bids will be displayed on a display at their own window. Authentication may be key for the online machine and get admission to credentials can be provided simplest to proven users. Users can set standards for automobile bidding including restriction amount, next bid increase quantity and many others. Information can be accrued by using the system to choose which objects evoked the maximum interest in bidders. We purpose to construct a tool that recommends the best acceptable product for person based on consultation.

The Crisis of (MT) From Arabic into Foreign Languages Case Study : Google Translate

Why do machine translation from Arabic into foreign languages and vice versa constitute a challenge compared to other languages? Is it because traditional grammar rules haven’t been computerized yet? Or because of the lack of technological skills in the Arab world? Or is it simply due to the lack of interest in developing Arabic in order to cope with the digital world needs? The aim of the present study is to investigate the effectiveness of the free online machine translation (GoogleTranslate) from Arabic into English or vice versa. It attempts to define machine translation., Automatic processing of Arabic within the scope of the “machine translation and automatic language processing” axis, We will try to display samples of the machine translation crisis through the Google Translate application . It has been concluded that the efforts of Arab researchers are still in the beginning, and the pattern of teaching pure sciences such as mathematics, physics, and technology in a language other than Arabic constitutes a major hindrance.

Intelligent femtosecond laser bone drilling via online monitoring and machine learning

In conventional spinal surgeries, mechanical and thermal injuries frequently usually arise due to improper handling, giving rise to a range of complications including infection, poor wound healing and bleeding. Femtosecond laser ablation offers a promising approach owing to high precision and low thermal damage. In this study, an intelligent femtosecond laser drilling method of human spinal bones has been proposed, and a machine learning method has been employed to determine the optimal laser processing window, ensuring high-quality outcomes. A neural network model has been developed to predict drilling quality, achieving an impressive accuracy rate exceeding 98 %, along with precision and recall rates of 100 % and 92.86 %, respectively. To further monitor the process, a fiber spectrometer and a thermal camera has been employed to monitor the focal status and bone temperature during laser processing to make sure the drilling is in a focal position and temperature in safe range. Subsequently, the drilling efficiency has been predicted using another neural network model within high-quality processing window for the maximum ablation processing parameter. The current research has demonstrated a direct, non-destructive and efficient method for intelligent laser spinal drilling.

Dynamic Industrial Optimization: A Framework Integrates Online Machine Learning for Processing Parameters Design

We develop the online process parameter design (OPPD) framework for efficiently handling streaming data collected from industrial automation equipment. This framework integrates online machine learning, concept drift detection and Bayesian optimization techniques. Initially, concept drift detection mitigates the impact of anomalous data on model updates. Data without concept drift are used for online model training and updating, enabling accurate predictions for the next processing cycle. Bayesian optimization is then employed for inverse optimization and process parameter design. Within OPPD, we introduce the online accelerated support vector regression (OASVR) algorithm for enhanced computational efficiency and model accuracy. OASVR simplifies support vector regression, boosting both speed and durability. Furthermore, we incorporate a dynamic window mechanism to regulate the training data volume for adapting to real-time demands posed by diverse online scenarios. Concept drift detection uses the EI-kMeans algorithm, and the Bayesian inverse design employs an upper confidence bound approach with an adaptive learning rate. Applied to single-crystal fabrication, the OPPD framework outperforms other models, with an RMSE of 0.12, meeting precision demands in production.

Ensemble adaptive online machine learning in data stream: a case study in cyber intrusion detection system

Ensemble adaptive online machine learning in data stream: a case study in cyber intrusion detection system

Enhancing stability and robustness in online machine shop scheduling: A multi-agent system and negotiation-based approach for handling machine downtime in industry 4.0

Autonomous factories require high levels of adaptability, flexibility, and resilience to react to uncertainties on the shop floor, such as machine downtime. This paper proposes a negotiation-based, partial rescheduling method, combined with an existing multi-agent system, to swap jobs between machines. The negotiations are restricted to machines within the same work center, giving rise to a partial reschedule. A learning algorithm is also utilized, allowing machines to individually learn how to evaluate proposed bids from other machines and adapt the bids to their current environment. The main objective is to minimize the mean weighted tardiness of all jobs. Computational results indicate an improvement of 10–30 tardiness, compared to continuous rescheduling and complete rescheduling methods. In addition, a decrease of 70–80 sensitivity analysis and analysis of the partial reschedule.

Over-deposition assessment of Direct Energy Deposition (DED) using melt pool geometric features and Machine Learning

Over-deposition’s influence on the DED process is significant in terms of process stability and final part quality. Most recent studies on quality assessment are based on online monitoring of the melt pool and comparing it with pre-specified thresholds. In this study an online machine learning model is developed, capable of identifying two over-deposition defect types, satellite and wire-oscillation, by using frames captured by an optical camera targeting the melt pool. The features extracted from the images included five geometrical melt pool features and the operating laser power and were successfully used to achieve a 94.4% classification accuracy regardless of the final part geometry.

An Online Support Vector Machine Algorithm for Dynamic Social Network Monitoring

Online monitoring of social networks offers exciting features for platforms, enabling both technical and behavioral analysis. Numerous studies have explored the adaptation of traditional quality control methods for detecting change points within social networks. However, the current research studies face limitations such as an overreliance on case-based attributes, high computational costs, poor scalability with large networks, and low sensitivity in fast change point detection. This paper proposes a novel algorithm for social network monitoring using One-Class Support Vector Machines (OC-SVMs) to address these limitations. Additionally, using both nodal and network-level attributes makes it versatile for diverse social network applications and effectively detecting network disturbances. The algorithm utilizes a well-defined training data dictionary with an updating procedure for evolutionary networks, enhancing memory and time efficiency by reducing the processing of input data. Extensive numerical experiments are conducted using an EpiCNet model to simulate interactions in an online social network, covering six change scenarios to evaluate the proposed methodology. The results show lower Average Run Length (ARL) and Expected Delay Detection (EDD), demonstrating the superior accuracy and effectiveness of the OC-SVM algorithm compared to alternative methods. Applying OC-SVM to the Enron Email network indicates its capability to identify change points, reflecting the tumultuous timeline that led to Enron's downfall. This further validates the substantial advancement of OC-SVM in social network monitoring and opens doors to broader real-world applications.