With the increasing use of Machine Learning (ML) in critical autonomous systems, runtime monitors have been developed to detect prediction errors and keep the system in a safe state during operations. Monitors have been proposed for different applications involving diverse perception tasks and ML models, and specific evaluation procedures and metrics are used for different contexts. This paper introduces three unified safety-oriented metrics, representing the safety benefits of the monitor (Safety Gain), the remaining safety gaps after using it (Residual Hazard), and its negative impact on the system's performance (Availability Cost). To compute these metrics, one requires to define two return functions, representing how a given ML prediction will impact expected future rewards and hazards. Three use-cases (classification, drone landing, and autonomous driving) are used to demonstrate how metrics from the literature can be expressed in terms of the proposed metrics. Experimental results on these examples show how different evaluation choices impact the perceived performance of a monitor. As our formalism requires us to formulate explicit safety assumptions, it allows us to ensure that the evaluation conducted matches the high-level system requirements.

Objectives: The present study evaluated the impact of a commercially available explainable AI algorithm in augmenting the ability of clinicians to identify lung cancer on chest X-rays (CXR). Design: This retrospective study evaluated the performance of 11 clinicians for detecting lung cancer from chest radiographs, with and without assistance from a commercially available AI algorithm (red dot, Behold.ai) that predicts suspected lung cancer from CXRs. Clinician performance was evaluated against clinically confirmed diagnoses. Setting: The study analysed anonymised patient data from an NHS hospital; the dataset consisted of 400 chest radiographs from adult patients (18 years and above) who had a CXR performed in 2020, with corresponding clinical text reports. Participants: A panel of readers consisting of 11 clinicians (consultant radiologists, radiologist trainees and reporting radiographers) participated in this study. Main outcome measures: Overall accuracy, sensitivity, specificity and precision for detecting lung cancer on CXRs by clinicians, with and without AI input. Agreement rates between clinicians and performance standard deviation were also evaluated, with and without AI input. Results: The use of the AI algorithm by clinicians led to an improved overall performance for lung tumour detection, achieving an overall increase of 17.4% of lung cancers being identified on CXRs which would have otherwise been missed, an overall increase in detection of smaller tumours, a 24% and 13% increased detection of stage 1 and stage 2 lung cancers respectively, and standardisation of clinician performance. Conclusions: This study showed great promise in the clinical utility of AI algorithms in improving early lung cancer diagnosis and promoting health equity through overall improvement in reader performances, without impacting downstream imaging resources.

Prediction of individual's race and ethnicity plays an important role in social science and public health research. Examples include studies of racial disparity in health and voting. Recently, Bayesian Improved Surname Geocoding (BISG), which uses Bayes' rule to combine information from Census surname files with the geocoding of an individual's residence, has emerged as a leading methodology for this prediction task. Unfortunately, BISG suffers from two Census data problems that contribute to unsatisfactory predictive performance for minorities. First, the decennial Census often contains zero counts for minority racial groups in the Census blocks where some members of those groups reside. Second, because the Census surname files only include frequent names, many surnames -- especially those of minorities -- are missing from the list. To address the zero counts problem, we introduce a fully Bayesian Improved Surname Geocoding (fBISG) methodology that accounts for potential measurement error in Census counts by extending the naive Bayesian inference of the BISG methodology to full posterior inference. To address the missing surname problem, we supplement the Census surname data with additional data on last, first, and middle names taken from the voter files of six Southern states where self-reported race is available. Our empirical validation shows that the fBISG methodology and name supplements significantly improve the accuracy of race imputation across all racial groups, and especially for Asians. The proposed methodology, together with additional name data, is available via the open-source software WRU.

Complex Event Recognition and Forecasting (CER/F) techniques attempt to detect, or even forecast ahead of time, event occurrences in streaming input using predefined event patterns. Such patterns are not always known in advance, or they frequently change over time, making machine learning techniques, capable of extracting such patterns from data, highly desirable in CER/F. Since many CER/F systems use symbolic automata to represent such patterns, we propose a family of such automata where the transition-enabling conditions are defined by Answer Set Programming (ASP) rules, and which, thanks to the strong connections of ASP to symbolic learning, are directly learnable from data. We present such a learning approach in ASP and an incremental version thereof that trades optimality for efficiency and is capable to scale to large datasets. We evaluate our approach on two CER datasets and compare it to state-of-the-art automata learning techniques, demonstrating empirically a superior performance, both in terms of predictive accuracy and scalability.

Physiological monitoring in intensive care units (ICU) generates data that can be used in clinical research. However, the recording conditions in clinical settings limit the automated extraction of relevant information from physiological signals due to noise and artifacts. Therefore, removing artifacts before clinical research is essential. Manual annotation by experienced researchers, which is the gold standard for removing artifacts, is time-consuming and costly due to the volume of the data generated in the ICU. In this study, we propose a hybrid artifact detection system that combines a Variational Autoencoder with a statistical detection component for the labeling of artifactual samples to automate the costly process of cleaning physiological recordings. The system is applied to minute-by-minute mean blood pressure signals from an intensive care unit dataset. Its performance is verified by manual annotations made by an expert. We benchmark the performance of our system with two other systems that combine an ARIMA or an autoencoder-based model with our statistical detection component. Our results indicate that the system consistently achieves sensitivity and specificity levels of over 90%. Thus, it provides an initial foundation to automate data cleaning in recordings from ICU.

The classification of electroencephalography (EEG) signals is useful in a wide range of applications such as seizure detection/prediction, motor imagery classification, emotion classification and drug effects diagnosis, amongst others. With the large number of EEG channels acquired, it has become vital that efficient data-reduction methods are developed, with varying importance from one application to another. It is also important that online classification is achieved during EEG recording for many applications, to monitor changes as they happen. In this paper we introduce a method based on Mutual Information (MI), for channel selection. Obtained results show that whilst there is a penalty on classification accuracy scores, promising speed-up gains can be achieved using MI techniques. Using MI with signal epochs (3secs) containing signal transitions enhances these speed-up gains. This work is exploratory and we suggest further research to be carried out for validation and development. Benefits to improving classification speed include improving application in clinical or educational settings.

Camera traps, unmanned observation devices, and deep learning-based image recognition systems have greatly reduced human effort in collecting and analyzing wildlife images. However, data collected via above apparatus exhibits 1) long-tailed and 2) open-ended distribution problems. To tackle the open-set long-tailed recognition problem, we propose the Temporal Flow Mask Attention Network that comprises three key building blocks: 1) an optical flow module, 2) an attention residual module, and 3) a meta-embedding classifier. We extract temporal features of sequential frames using the optical flow module and learn informative representation using attention residual blocks. Moreover, we show that applying the meta-embedding technique boosts the performance of the method in open-set long-tailed recognition. We apply this method on a Korean Demilitarized Zone (DMZ) dataset. We conduct extensive experiments, and quantitative and qualitative analyses to prove that our method effectively tackles the open-set long-tailed recognition problem while being robust to unknown classes.

Deep Neural Networks (DNNs) have achieved excellent performance in various fields. However, DNNs' vulnerability to Adversarial Examples (AE) hinders their deployments to safety-critical applications. This paper presents a novel AE detection framework, named BEYOND, for trustworthy predictions. BEYOND performs the detection by distinguishing the AE's abnormal relation with its augmented versions, i.e. neighbors, from two prospects: representation similarity and label consistency. An off-the-shelf Self-Supervised Learning (SSL) model is used to extract the representation and predict the label for its highly informative representation capacity compared to supervised learning models. For clean samples, their representations and predictions are closely consistent with their neighbors, whereas those of AEs differ greatly. Furthermore, we explain this observation and show that by leveraging this discrepancy BEYOND can effectively detect AEs. We develop a rigorous justification for the effectiveness of BEYOND. Furthermore, as a plug-and-play model, BEYOND can easily cooperate with the Adversarial Trained Classifier (ATC), achieving the state-of-the-art (SOTA) robustness accuracy. Experimental results show that BEYOND outperforms baselines by a large margin, especially under adaptive attacks. Empowered by the robust relation net built on SSL, we found that BEYOND outperforms baselines in terms of both detection ability and speed. Our code will be publicly available.

The ability to detect zero-day (novel) attacks has become essential in the network security industry. Due to ever evolving attack signatures, existing network intrusion detection systems often fail to detect these threats. This project aims to solve the task of detecting zero-day DDoS (distributed denial-of-service) attacks by utilizing network traffic that is captured before entering a private network. Modern feature extraction techniques are used in conjunction with neural networks to determine if a network packet is either benign or malicious.

Hail risk assessment is necessary to estimate and reduce damage to crops, orchards, and infrastructure. Also, it helps to estimate and reduce consequent losses for businesses and, particularly, insurance companies. But hail forecasting is challenging. Data used for designing models for this purpose are tree-dimensional geospatial time series. Hail is a very local event with respect to the resolution of available datasets. Also, hail events are rare - only 1% of targets in observations are marked as "hail". Models for nowcasting and short-term hail forecasts are improving. Introducing machine learning models to the meteorology field is not new. There are also various climate models reflecting possible scenarios of climate change in the future. But there are no machine learning models for data-driven forecasting of changes in hail frequency for a given area. The first possible approach for the latter task is to ignore spatial and temporal structure and develop a model capable of classifying a given vertical profile of meteorological variables as favorable to hail formation or not. Although such an approach certainly neglects important information, it is very light weighted and easily scalable because it treats observations as independent from each other. The more advanced approach is to design a neural network capable to process geospatial data. Our idea here is to combine convolutional layers responsible for the processing of spatial data with recurrent neural network blocks capable to work with temporal structure. This study compares two approaches and introduces a model suitable for the task of forecasting changes in hail frequency for ongoing decades.