Due to the mismatch between the source and target domains, how to better utilize the biased word information to improve the performance of the automatic speech recognition model in the target domain becomes a hot research topic. Previous approaches either decode with a fixed external language model or introduce a sizeable biasing module, which leads to poor adaptability and slow inference. In this work, we propose CB-Conformer to improve biased word recognition by introducing the Contextual Biasing Module and the Self-Adaptive Language Model to vanilla Conformer. The Contextual Biasing Module combines audio fragments and contextual information, with only 0.2% model parameters of the original Conformer. The Self-Adaptive Language Model modifies the internal weights of biased words based on their recall and precision, resulting in a greater focus on biased words and more successful integration with the automatic speech recognition model than the standard fixed language model. In addition, we construct and release an open-source Mandarin biased-word dataset based on WenetSpeech. Experiments indicate that our proposed method brings a 15.34% character error rate reduction, a 14.13% biased word recall increase, and a 6.80% biased word F1-score increase compared with the base Conformer.

The simultaneous quantile regression (SQR) technique has been used to estimate uncertainties for deep learning models, but its application is limited by the requirement that the solution at the median quantile ({\tau} = 0.5) must minimize the mean absolute error (MAE). In this article, we address this limitation by demonstrating a duality between quantiles and estimated probabilities in the case of simultaneous binary quantile regression (SBQR). This allows us to decouple the construction of quantile representations from the loss function, enabling us to assign an arbitrary classifier f(x) at the median quantile and generate the full spectrum of SBQR quantile representations at different {\tau} values. We validate our approach through two applications: (i) detecting out-of-distribution samples, where we show that quantile representations outperform standard probability outputs, and (ii) calibrating models, where we demonstrate the robustness of quantile representations to distortions. We conclude with a discussion of several hypotheses arising from these findings.

Machine Learning (ML) software has been widely adopted in modern society, with reported fairness implications for minority groups based on race, sex, age, etc. Many recent works have proposed methods to measure and mitigate algorithmic bias in ML models. The existing approaches focus on single classifier-based ML models. However, real-world ML models are often composed of multiple independent or dependent learners in an ensemble (e.g., Random Forest), where the fairness composes in a non-trivial way. How does fairness compose in ensembles? What are the fairness impacts of the learners on the ultimate fairness of the ensemble? Can fair learners result in an unfair ensemble? Furthermore, studies have shown that hyperparameters influence the fairness of ML models. Ensemble hyperparameters are more complex since they affect how learners are combined in different categories of ensembles. Understanding the impact of ensemble hyperparameters on fairness will help programmers design fair ensembles. Today, we do not understand these fully for different ensemble algorithms. In this paper, we comprehensively study popular real-world ensembles: bagging, boosting, stacking and voting. We have developed a benchmark of 168 ensemble models collected from Kaggle on four popular fairness datasets. We use existing fairness metrics to understand the composition of fairness. Our results show that ensembles can be designed to be fairer without using mitigation techniques. We also identify the interplay between fairness composition and data characteristics to guide fair ensemble design. Finally, our benchmark can be leveraged for further research on fair ensembles. To the best of our knowledge, this is one of the first and largest studies on fairness composition in ensembles yet presented in the literature.

Under the global COVID-19 crisis, accurate diagnosis of COVID-19 from Chest X-ray (CXR) images is critical. To reduce intra- and inter-observer variability, during the radiological assessment, computer-aided diagnostic tools have been utilized to supplement medical decision-making and subsequent disease management. Computational methods with high accuracy and robustness are required for rapid triaging of patients and aiding radiologists in the interpretation of the collected data. In this study, we propose a novel multi-feature fusion network using parallel attention blocks to fuse the original CXR images and local-phase feature-enhanced CXR images at multi-scales. We examine our model on various COVID-19 datasets acquired from different organizations to assess the generalization ability. Our experiments demonstrate that our method achieves state-of-art performance and has improved generalization capability, which is crucial for widespread deployment.

The aim of this research paper is to look into the use of continuous authentication with mobile touch dynamics, using three different algorithms: Neural Network, Extreme Gradient Boosting, and Support Vector Machine. Mobile devices are constantly increasing in popularity in the world, today smartphone subscriptions have surpassed 6 billion. Mobile touch dynamics refer to the distinct patterns of how a user interacts with their mobile device, this includes factors such as touch pressure, swipe speed, and touch duration. Continuous authentication refers to the process of continuously verifying a user's identity while they are using a device, rather than just at the initial login. This research used a dataset of touch dynamics collected from 40 subjects using the LG V30+. The participants played four mobile games, PUBG, Diep.io, Slither, and Minecraft, for 10 minutes each game. The three algorithms were trained and tested on the extracted dataset, and their performance was evaluated based on metrics such as accuracy, precision, false negative rate, and false positive rate. The results of the research showed that all three algorithms were able to effectively classify users based on their individual touch dynamics, with accuracy ranging from 80% to 95%. The Neural Network algorithm performed the best, achieving the highest accuracy and precision scores, followed closely by XGBoost and SVC. The data shows that continuous authentication using mobile touch dynamics has the potential to be a useful method for enhancing security and reducing the risk of unauthorized access to personal devices. This research also notes the importance of choosing the correct algorithm for a given dataset and use case, as different algorithms may have varying levels of performance depending on the specific task.

Before 2020, the way occupants utilized the built environment had been changing slowly towards scenarios in which occupants have more choice and flexibility in where and how they work. The global COVID-19 pandemic accelerated this phenomenon rapidly through lockdowns and hybrid work arrangements. Many occupants and employers are considering keeping some of these flexibility-based strategies due to their benefits and cost impacts. This paper simulates various scenarios related to the operational technologies and policies of a real-world campus using a district-scale City Energy Analyst (CEA) model that is calibrated with measured energy and occupancy profiles extracted from WiFi data. These scenarios demonstrate the energy impact of ramping building operations up and down more rapidly and effectively to the flex-based work strategies that may solidify. The scenarios show a 4-12% decrease in space cooling demand due to occupant absenteeism if centralized building system operation is in place, but as high as 21-68% if occupancy-driven building controls are implemented. The paper discusses technologies and strategies that are important in this paradigm shift of operations.

Almost 50 years after the invention of SQL, injection attacks are still top-tier vulnerabilities of today's ICT systems. Consequently, SQLi detection is still an active area of research, where the most recent works incorporate machine learning techniques into the proposed solutions. In this work, we highlight the shortcomings of the previous ML-based results focusing on four aspects: the evaluation methods, the optimization of the model parameters, the distribution of utilized datasets, and the feature selection. Since no single work explored all of these aspects satisfactorily, we fill this gap and provide an in-depth and comprehensive empirical analysis. Moreover, we cross-validate the trained models by using data from other distributions. This aspect of ML models (trained for SQLi detection) was never studied. Yet, the sensitivity of the model's performance to this is crucial for any real-life deployment.

The discovery of causal relationships from high-dimensional data is a major open problem in bioinformatics. Machine learning and feature attribution models have shown great promise in this context but lack causal interpretation. Here, we show that a popular feature attribution model estimates a causal quantity reflecting the influence of one variable on another, under certain assumptions. We leverage this insight to implement a new tool, CIMLA, for discovering conditiondependent changes in causal relationships. We then use CIMLA to identify differences in gene regulatory networks between biological conditions, a problem that has received great attention in recent years. Using extensive benchmarking on simulated data sets, we show that CIMLA is more robust to confounding variables and is more accurate than leading methods. Finally, we employ CIMLA to analyze a previously published single-cell RNA-seq data set collected from subjects with and without Alzheimer's disease (AD), discovering several potential regulators of AD. INTRODUCTION A key challenge in bioinformatics today is to extract causal relationships from omics data.

Neural networks have been successfully employed in various domains such as classification, regression and clustering, etc. Generally, the back propagation (BP) based iterative approaches are used to train the neural networks, however, it results in the issues of local minima, sensitivity to learning rate and slow convergence. To overcome these issues, randomization based neural networks such as random vector functional link (RVFL) network have been proposed. RVFL model has several characteristics such as fast training speed, direct links, simple architecture, and universal approximation capability, that make it a viable randomized neural network. This article presents the first comprehensive review of the evolution of RVFL model, which can serve as the extensive summary for the beginners as well as practitioners. We discuss the shallow RVFLs, ensemble RVFLs, deep RVFLs and ensemble deep RVFL models. The variations, improvements and applications of RVFL models are discussed in detail. Moreover, we discuss the different hyperparameter optimization techniques followed in the literature to improve the generalization performance of the RVFL model. Finally, we give potential future research directions/opportunities that can inspire the researchers to improve the RVFL's architecture and learning algorithm further.

Language Models (LMs) have been shown to leak information about training data through sentence-level membership inference and reconstruction attacks. Understanding the risk of LMs leaking Personally Identifiable Information (PII) has received less attention, which can be attributed to the false assumption that dataset curation techniques such as scrubbing are sufficient to prevent PII leakage. Scrubbing techniques reduce but do not prevent the risk of PII leakage: in practice scrubbing is imperfect and must balance the trade-off between minimizing disclosure and preserving the utility of the dataset. On the other hand, it is unclear to which extent algorithmic defenses such as differential privacy, designed to guarantee sentence- or user-level privacy, prevent PII disclosure. In this work, we introduce rigorous game-based definitions for three types of PII leakage via black-box extraction, inference, and reconstruction attacks with only API access to an LM. We empirically evaluate the attacks against GPT-2 models fine-tuned with and without defenses in three domains: case law, health care, and e-mails. Our main contributions are (i) novel attacks that can extract up to 10$\times$ more PII sequences than existing attacks, (ii) showing that sentence-level differential privacy reduces the risk of PII disclosure but still leaks about 3% of PII sequences, and (iii) a subtle connection between record-level membership inference and PII reconstruction. Code to reproduce all experiments in the paper is available at https://github.com/microsoft/analysing_pii_leakage.