Due to the increased demand for music streaming/recommender services and the recent developments of music information retrieval frameworks, Music Genre Classification (MGC) has attracted the community's attention. However, convolutional-based approaches are known to lack the ability to efficiently encode and localize temporal features. In this paper, we study the broadcast-based neural networks aiming to improve the localization and generalizability under a small set of parameters (about 180k) and investigate twelve variants of broadcast networks discussing the effect of block configuration, pooling method, activation function, normalization mechanism, label smoothing, channel interdependency, LSTM block inclusion, and variants of inception schemes. Our computational experiments using relevant datasets such as GTZAN, Extended Ballroom, HOMBURG, and Free Music Archive (FMA) show state-of-the-art classification accuracies in Music Genre Classification. Our approach offers insights and the potential to enable compact and generalizable broadcast networks for music and audio classification.

Rail breaks are one of the most common causes of derailments internationally. This is no different for the South African Iron Ore line. Many rail breaks occur as a heavy-haul train passes over a crack, large defect or defective weld. In such cases, it is usually too late for the train to slow down in time to prevent a de-railment. Knowing the risk of a rail break occurring associated with a train passing over a section of rail allows for better implementation of maintenance initiatives and mitigating measures. In this paper the Ore Line's specific challenges are discussed and the currently available data that can be used to create a rail break risk prediction model is reviewed. The development of a basic rail break risk prediction model for the Ore Line is then presented. Finally the insight gained from the model is demonstrated by means of discussing various scenarios of various rail break risk. In future work, we are planning on extending this basic model to allow input from live monitoring systems such as the ultrasonic broken rail detection system.

Driven by the goal of eradicating language barriers on a global scale, machine translation has solidified itself as a key focus of artificial intelligence research today. However, such efforts have coalesced around a small subset of languages, leaving behind the vast majority of mostly low-resource languages. What does it take to break the 200 language barrier while ensuring safe, high quality results, all while keeping ethical considerations in mind? In No Language Left Behind, we took on this challenge by first contextualizing the need for low-resource language translation support through exploratory interviews with native speakers. Then, we created datasets and models aimed at narrowing the performance gap between low and high-resource languages. More specifically, we developed a conditional compute model based on Sparsely Gated Mixture of Experts that is trained on data obtained with novel and effective data mining techniques tailored for low-resource languages. We propose multiple architectural and training improvements to counteract overfitting while training on thousands of tasks. Critically, we evaluated the performance of over 40,000 different translation directions using a human-translated benchmark, Flores-200, and combined human evaluation with a novel toxicity benchmark covering all languages in Flores-200 to assess translation safety. Our model achieves an improvement of 44% BLEU relative to the previous state-of-the-art, laying important groundwork towards realizing a universal translation system.

Widely used deep learning models are found to have poor robustness. Little noises can fool state-of-the-art models into making incorrect predictions. While there is a great deal of high-performance attack generation methods, most of them directly add perturbations to original data and measure them using L_p norms; this can break the major structure of data, thus, creating invalid attacks. In this paper, we propose a black-box attack, which, instead of modifying original data, modifies latent features of data extracted by an autoencoder; then, we measure noises in semantic space to protect the semantics of data. We trained autoencoders on MNIST and CIFAR-10 datasets and found optimal adversarial perturbations using a genetic algorithm. Our approach achieved a 100% attack success rate on the first 100 data of MNIST and CIFAR-10 datasets with less perturbation than FGSM.

People have recently begun communicating their thoughts and viewpoints through user-generated multimedia material on social networking websites. This information can be images, text, videos, or audio. Recent years have seen a rise in the frequency of occurrence of this pattern. Twitter is one of the most extensively utilized social media sites, and it is also one of the finest locations to get a sense of how people feel about events that are linked to the Monkeypox sickness. This is because tweets on Twitter are shortened and often updated, both of which contribute to the platform's character. The fundamental objective of this study is to get a deeper comprehension of the diverse range of reactions people have in response to the presence of this condition. This study focuses on finding out what individuals think about monkeypox illnesses, which presents a hybrid technique based on CNN and LSTM. We have considered all three possible polarities of a user's tweet: positive, negative, and neutral. An architecture built on CNN and LSTM is utilized to determine how accurate the prediction models are. The recommended model's accuracy was 94% on the monkeypox tweet dataset. Other performance metrics such as accuracy, recall, and F1-score were utilized to test our models and results in the most time and resource-effective manner. The findings are then compared to more traditional approaches to machine learning. The findings of this research contribute to an increased awareness of the monkeypox infection in the general population.

Unsupervised visual representation learning offers the opportunity to leverage large corpora of unlabeled trajectories to form useful visual representations, which can benefit the training of reinforcement learning (RL) algorithms. However, evaluating the fitness of such representations requires training RL algorithms which is computationally intensive and has high variance outcomes. To alleviate this issue, we design an evaluation protocol for unsupervised RL representations with lower variance and up to 600x lower computational cost. Inspired by the vision community, we propose two linear probing tasks: predicting the reward observed in a given state, and predicting the action of an expert in a given state. These two tasks are generally applicable to many RL domains, and we show through rigorous experimentation that they correlate strongly with the actual downstream control performance on the Atari100k Benchmark. This provides a better method for exploring the space of pretraining algorithms without the need of running RL evaluations for every setting. Leveraging this framework, we further improve existing self-supervised learning (SSL) recipes for RL, highlighting the importance of the forward model, the size of the visual backbone, and the precise formulation of the unsupervised objective.

Wireless Sensor Networks (WSNs) is a promising technology with enormous applications in almost every walk of life. One of the crucial applications of WSNs is intrusion detection and surveillance at the border areas and in the defense establishments. The border areas are stretched in hundreds to thousands of miles, hence, it is not possible to patrol the entire border region. As a result, an enemy may enter from any point absence of surveillance and cause the loss of lives or destroy the military establishments. WSNs can be a feasible solution for the problem of intrusion detection and surveillance at the border areas. Detection of an enemy at the border areas and nearby critical areas such as military cantonments is a time-sensitive task as a delay of few seconds may have disastrous consequences. Therefore, it becomes imperative to design systems that are able to identify and detect the enemy as soon as it comes in the range of the deployed system. In this paper, we have proposed a deep learning architecture based on a fully connected feed-forward Artificial Neural Network (ANN) for the accurate prediction of the number of k-barriers for fast intrusion detection and prevention. We have trained and evaluated the feed-forward ANN model using four potential features, namely area of the circular region, sensing range of sensors, the transmission range of sensors, and the number of sensor for Gaussian and uniform sensor distribution. These features are extracted through Monte Carlo simulation. In doing so, we found that the model accurately predicts the number of k-barriers for both Gaussian and uniform sensor distribution with correlation coefficient (R = 0.78) and Root Mean Square Error (RMSE = 41.15) for the former and R = 0.79 and RMSE = 48.36 for the latter. Further, the proposed approach outperforms the other benchmark algorithms in terms of accuracy and computational time complexity.

More than fifty years after a research group at Dartmouth University launched work on a new field called "Artificial Intelligence," the United States still lacks a national strategy on artificial intelligence (AI) policy. The growing urgency of this endeavor is made clear by the rapid progress of both U.S. allies and adversaries. The European Union is moving forward with two initiatives of far-reaching consequence. The EU Artificial Intelligence Act will establish a comprehensive, risk-based approach for the regulation of AI when it is adopted in 2023. Many anticipate that the EU AI Act will extend the "Brussels Effect" across the AI sector as the earlier European data privacy law, the General Data Privacy Regulation, did for much of the tech industry.

Federated learning (FL) is a newly emerged branch of AI that facilitates edge devices to collaboratively train a global machine learning model without centralizing data and with privacy by default. However, despite the remarkable advancement, this paradigm comes with various challenges. Specifically, in large-scale deployments, client heterogeneity is the norm which impacts training quality such as accuracy, fairness, and time. Moreover, energy consumption across these battery-constrained devices is largely unexplored and a limitation for wide-adoption of FL. To address this issue, we develop EAFL, an energy-aware FL selection method that considers energy consumption to maximize the participation of heterogeneous target devices. EAFL is a power-aware training algorithm that cherry-picks clients with higher battery levels in conjunction with its ability to maximize the system efficiency. Our design jointly minimizes the time-to-accuracy and maximizes the remaining on-device battery levels. EAFLimproves the testing model accuracy by up to 85\% and decreases the drop-out of clients by up to 2.45$\times$.

The advance of Machine Learning (ML) has led to a strong interest in this technology to support decision making. While complex ML models provide predictions that are often more accurate than those of traditional tools, such models often hide the reasoning behind the prediction from their users, which can lead to lower adoption and lack of insight. Motivated by this tension, research has put forth Explainable Artificial Intelligence (XAI) techniques that uncover patterns discovered by ML. Despite the high hopes in both ML and XAI, there is little empirical evidence of the benefits to traditional businesses. To this end, we analyze data on 220,185 customers of an energy retailer, predict cross-purchases with up to 86% correctness (AUC), and show that the XAI method SHAP provides explanations that hold for actual buyers.