Joint Training and Decoding for Multilingual End-to-End Simultaneous Speech Translation

Speech translation systems usually follow a pipeline approach, using word lattices as an intermediate representation. However, previous work assume access to the original transcriptions used to train the ASR system, which can limit applicability in real scenarios. In this work we propose an approach for speech translation through lattice transformations and neural models based on graph networks. Experimental results show that our approach reaches competitive performance without relying on transcriptions, while also being orders of magnitude faster than previous work.

Signal detection and material identification method for loose particles inside aerospace relays based on overlapping signals

Speech is a complex bio-signal that has the potential to provide a rich bio-marker for health. It enables the development of non-invasive routes to early diagnosis and monitoring of speech affecting diseases, such as the ones studied in this work: Depression, and Parkinson's Disease. However, the major limitation of current speech based diagnosis and monitoring tools is the lack of large and diverse datasets. Existing datasets are small, and collected under very controlled conditions. As such, there is an upper bound in the complexity of the models that can be trained using these datasets. There is also limited applicability in real life scenarios where the channel and noise conditions, among others, are impossible to control. In this work, we show that datasets collected from in-the-wild sources, such as collections of vlogs, can contribute to improve the performance of diagnosis tools both in controlled and in-the-wild conditions, even though the data are noisier. Moreover, we show that it is possible to successfully move away from hand-crafted features (i.e. features that are computed based on predefined algorithms, that based on human expertise) and adopt end-to-end modeling paradigms, such as CNN-LSTMs, that extract data driven features from the raw spectrograms of the speech signal, and capture temporal information from the speech signals.

Simultaneous speech translation aims to maintain translation quality while minimizing the delay between reading input and incrementally producing the output. We propose a new general-purpose prediction action which predicts future words in the input to improve quality and minimize delay in simultaneous translation. We train this agent using reinforcement learning with a novel reward function. Our agent with prediction has better translation quality and less delay compared to an agent-based simultaneous translation system without prediction.

A polyelectrolyte based supramolecular assembly for ratiometric sensing of ATP with very high discrimination from pyrophosphate

World transformation has meaningfully abridged the chief power sources in terms of gas, diesel, and coal. Therefore, substitute power sources based on renewable energy mainly focus on fulfilling the energy demand of the world as well as avoiding global warming. Amid different energy sources, solar energy is the key source of substitute energy used for producing electricity using photovoltaic (PV). Conversely, energy engendering performance is extremely dependent on cyclical and environmental factors. The changeable performance of environment shakes energy productivity and tends to create a disapproving influence on constancy, dependability, and the process of grid. Therefore, a precise prediction of PV productivity is critically required to guarantee constancy and dependability of grid. The detailed study to review the perilous techniques is based on PV forecast using machine learning techniques. In this chapter, different types of renewable energy with their merits and demerits are summarized. This chapter also demonstrates the key challenges existing in the current renewable energy using machine learning and various artificial intelligence applications in real-life scenarios. Finally, an inclusive analysis of machine learning techniques in renewable energy through detailed literature survey is presented for better forecasting of energy production in near future.

Semantic segmentation has always been a very challenging research topic in computer vision and deep learning and has extensive applications in real-life scenarios. With the development of computing hardware and deep learning technology, researchers have a higher research enthusiasm for semantic segmentation. This work briefly introduces several semantic segmentation models, datasets, and the main issues of traditional semantic segmentation. Afterward, the current mainstream semantic segmentation algorithm and the experimental results were compared. Finally, future research directions of semantic segmentation are presented and discussed.

The purpose of this project is to present active learning activities and instructional strategies that will allow counselling students to: (a) become familiar with ethical codes, (b) understand their application to their professional conduct, (c) strive to adhere to the codes' principles and values, (d) translate formal codes and standards into day-to-day professional practice, and (e) critique the codes for applicability in real-life scenarios. This project explores the following ethics topics: multicultural and diversity issues, client rights and counsellor responsibilities, confidentiality, boundary and multiple relationship issues, professional competence and training, supervision and consultation, research ethics, couples and family therapy, group work, and rural counselling. The elements used to explore the ethics topics include: discussion questions, case studies, activities, assignments, and annotated bibliographies.

Overconfidence is a cognitive bias that primarily consists of three components: overestimation, overplacement, and overprecision. Overconfidence bias has a wide range of applications in real-life scenarios like medical, educational, and financial fields. This paper focuses on the impact of overconfidence bias in two economic areas: the stock market and entrepreneurship. In the stock market, overconfident investors can lead to mispricing and overtrading. In entrepreneurship, overconfidence can lead to excess entry into the market and delayed exit from the market. Lastly, the paper discusses ways in which individuals can minimize overconfidence during stock market investing and entrepreneurship, including using objective data and information, seeking relevant professional education, and obtaining advice from diverse perspectives. Based on discussing the existence of overconfidence bias and its related coping strategies, this paper highlights the prevalence of the effects of overconfidence bias and the need to pay more attention to overconfidence bias, not only for stock investors and entrepreneurs.

Parcel lockers are considered an effective solution to address inefficiencies in last mile delivery. However, choosing the appropriate location is crucial to provide optimal service and improve logistic performance. Therefore, this research aims to assess the most suitable locations for parcel lockers in Dublin, the capital of Ireland, using a novel hybrid decision-making model. This study seeks to extend the Combinative Distance based Assessment method and integrate the Analytical Hierarchy Process method to handle parcel locker location evaluation problems through Decomposed Fuzzy Sets. The proposed model is a pioneering application in real-life scenarios, contributing to the existing knowledge in this field. The integrated model we propose offers a suitable approach for selecting parcel locker locations. The outcomes of this study will provide valuable insights for policymakers, aiding them in making strategic decisions regarding the selection of parcel locker locations within Dublin city.

Fraud in automobile odometers is a billion dollar problem worldwide. The technical requirements behind any solution for this problem go through the perennial, secure, and immutable registration of vehicular data. Such requirements immediately raise the question of how blockchain technology could be used for this purpose. In this article, the requirements, design choices, and construction details of a blockchain architecture for vehicular odometers are presented and discussed. The architecture advances the state of the art by adopting a more efficient consensus algorithm than the traditional ones, such as proof of work. Results from a prototype proof ofconcept of the proposed architecture validate the solution, which is able to validate a new block in the blockchain every two seconds. Applicability in real scenarios is one step ahead and very promising.

Super-resolution is a fundamental problem in computer vision which aims to overcome the spatial limitation of camera sensors. While significant progress has been made in single image super-resolution, most algorithms only perform well on synthetic data, which limits their applications in real scenarios. In this paper, we study the problem of real-scene single image super-resolution to bridge the gap between synthetic data and real captured images. We focus on two issues of existing super-resolution algorithms: lack of realistic training data and insufficient utilization of visual information obtained from cameras. To address the first issue, we propose a method to generate more realistic training data by mimicking the imaging process of digital cameras. For the second issue, we develop a two-branch convolutional neural network to exploit the radiance information originally-recorded in raw images. In addition, we propose a dense channel-attention block for better image restoration as well as a learning-based guided filter network for effective color correction. Our model is able to generalize to different cameras without deliberately training on images from specific camera types. Extensive experiments demonstrate that the proposed algorithm can recover fine details and clear structures, and achieve high-quality results for single image super-resolution in real scenes.

Legal Judgment Prediction (LJP) refers to the task of automatically predicting judgment results (e.g., charges, law articles and term of penalty) given the fact description of cases.While SOTA models have achieved high accuracy and F1 scores on public datasets, existing datasets fail to evaluate specific aspects of these models (e.g., legal fairness, which significantly impact their applications in real scenarios).Inspired by functional testing in software engineering, we introduce LJPCHECK, a suite of functional tests for LJP models, to comprehend LJP models' behaviors and offer diagnostic insights.We illustrate the utility of LJPCHECK on five SOTA LJP models.Extensive experiments reveal vulnerabilities in these models, prompting an in-depth discussion into the underlying reasons of their shortcomings.

Robust principal component analysis via optimal mean by joint ℓ2,1 and Schatten p-norms minimization

RFID complex event processing technology, which is proved to be a powerful tool in describing relationship between different events, such as timing, causality, and membership in RFID data streams or events, is a key point of RFID researching in these years. For the trend of great network application of RFID, to manage the tremendous and potentially infinite volumes of raw data generated by RFID systems poses new challenges. This paper presents an RFID complex event processing model in the context of distributed RFID network according to the characteristics of RFID data. Besides, some key methods to realize the modules and the application in real scenario are also discussed.