Lung cancer is a devastating disease with the highest mortality rate among cancer types. Over 60% of non-small cell lung cancer (NSCLC) patients, which accounts for 87% of diagnoses, require radiation therapy. Rapid treatment initiation significantly increases the patient's survival rate and reduces the mortality rate. Accurate tumor segmentation is a critical step in the diagnosis and treatment of NSCLC. Manual segmentation is time and labor-consuming and causes delays in treatment initiation. Although many lung nodule detection methods, including deep learning-based models, have been proposed, there is still a long-standing problem of high false positives (FPs) with most of these methods. Here, we developed an electronic health record (EHR) guided lung tumor auto-segmentation called EXACT-Net (EHR-enhanced eXACtitude in Tumor segmentation), where the extracted information from EHRs using a pre-trained large language model (LLM), was used to remove the FPs and keep the TP nodules only. The auto-segmentation model was trained on NSCLC patients' computed tomography (CT), and the pre-trained LLM was used with the zero-shot learning approach. Our approach resulted in a 250% boost in successful nodule detection using the data from ten NSCLC patients treated in our institution.
Electronic health records (EHR) is an inherently multimodal register of the patient's health status characterized by static data and multivariate time series (MTS). While MTS are a valuable tool for clinical prediction, their fusion with other data modalities can possibly result in more thorough insights and more accurate results. Deep neural networks (DNNs) have emerged as fundamental tools for identifying and defining underlying patterns in the healthcare domain. However, fundamental improvements in interpretability are needed for DNN models to be widely used in the clinical setting. In this study, we present an approach built on a collection of interpretable multimodal data-driven models that may anticipate and understand the emergence of antimicrobial multidrug resistance (AMR) germs in the intensive care unit (ICU) of the University Hospital of Fuenlabrada (Madrid, Spain). The profile and initial health status of the patient are modeled using static variables, while the evolution of the patient's health status during the ICU stay is modeled using several MTS, including mechanical ventilation and antibiotics intake. The multimodal DNNs models proposed in this paper include interpretable principles in addition to being effective at predicting AMR and providing an explainable prediction support system for AMR in the ICU. Furthermore, our proposed methodology based on multimodal models and interpretability schemes can be leveraged in additional clinical problems dealing with EHR data, broadening the impact and applicability of our results.
This paper introduces a new approach to sound source localization using head-related transfer function (HRTF) characteristics, which enable precise full-sphere localization from raw data. While previous research focused primarily on using extensive microphone arrays in the frontal plane, this arrangement often encountered limitations in accuracy and robustness when dealing with smaller microphone arrays. Our model proposes using both time and frequency domain for sound source localization while utilizing Deep Learning (DL) approach. The performance of our proposed model, surpasses the current state-of-the-art results. Specifically, it boasts an average angular error of $0.24 degrees and an average Euclidean distance of 0.01 meters, while the known state-of-the-art gives average angular error of 19.07 degrees and average Euclidean distance of 1.08 meters. This level of accuracy is of paramount importance for a wide range of applications, including robotics, virtual reality, and aiding individuals with cochlear implants (CI).
Vascular structure segmentation plays a crucial role in medical analysis and clinical applications. The practical adoption of fully supervised segmentation models is impeded by the intricacy and time-consuming nature of annotating vessels in the 3D space. This has spurred the exploration of weakly-supervised approaches that reduce reliance on expensive segmentation annotations. Despite this, existing weakly supervised methods employed in organ segmentation, which encompass points, bounding boxes, or graffiti, have exhibited suboptimal performance when handling sparse vascular structure. To alleviate this issue, we employ maximum intensity projection (MIP) to decrease the dimensionality of 3D volume to 2D image for efficient annotation, and the 2D labels are utilized to provide guidance and oversight for training 3D vessel segmentation model. Initially, we generate pseudo-labels for 3D blood vessels using the annotations of 2D projections. Subsequently, taking into account the acquisition method of the 2D labels, we introduce a weakly-supervised network that fuses 2D-3D deep features via MIP to further improve segmentation performance. Furthermore, we integrate confidence learning and uncertainty estimation to refine the generated pseudo-labels, followed by fine-tuning the segmentation network. Our method is validated on five datasets (including cerebral vessel, aorta and coronary artery), demonstrating highly competitive performance in segmenting vessels and the potential to significantly reduce the time and effort required for vessel annotation. Our code is available at: https://github.com/gzq17/Weakly-Supervised-by-MIP.
Power consumption plays an important role in on-device streaming speech recognition, as it has a direct impact on the user experience. This study delves into how weight parameters in speech recognition models influence the overall power consumption of these models. We discovered that the impact of weight parameters on power consumption varies, influenced by factors including how often they are invoked and their placement in memory. Armed with this insight, we developed design guidelines aimed at optimizing on-device speech recognition models. These guidelines focus on minimizing power use without substantially affecting accuracy. Our method, which employs targeted compression based on the varying sensitivities of weight parameters, demonstrates superior performance compared to state-of-the-art compression methods. It achieves a reduction in energy usage of up to 47% while maintaining similar model accuracy and improving the real-time factor.
Text simplification refers to the process of increasing the comprehensibility of texts. Automatic text simplification models are most commonly evaluated by experts or crowdworkers instead of the primary target groups of simplified texts, such as persons with intellectual disabilities. We conducted an evaluation study of text comprehensibility including participants with and without intellectual disabilities reading unsimplified, automatically and manually simplified German texts on a tablet computer. We explored four different approaches to measuring comprehensibility: multiple-choice comprehension questions, perceived difficulty ratings, response time, and reading speed. The results revealed significant variations in these measurements, depending on the reader group and whether the text had undergone automatic or manual simplification. For the target group of persons with intellectual disabilities, comprehension questions emerged as the most reliable measure, while analyzing reading speed provided valuable insights into participants' reading behavior.
The problem of 3D high-resolution imaging in automotive multiple-input multiple-output (MIMO) side-looking radar using a 1D array is considered. The concept of motion-enhanced snapshots is introduced for generating larger apertures in the azimuth dimension. For the first time, 3D imaging capabilities can be achieved with high angular resolution using a 1D MIMO antenna array, which can alleviate the requirement for large radar systems in autonomous vehicles. The robustness to variations in the vehicle's movement trajectory is also considered and addressed with relevant compensations in the steering vector. The available degrees of freedom as well as the Signal to Noise Ratio (SNR) are shown to increase with the proposed method compared to conventional imaging approaches. The performance of the algorithm has been studied in simulations, and validated with experimental data collected in a realistic driving scenario.
Cheating in online exams has become a prevalent issue over the past decade, especially during the COVID-19 pandemic. To address this issue of academic dishonesty, our "Exam Monitoring System: Detecting Abnormal Behavior in Online Examinations" is designed to assist proctors in identifying unusual student behavior. Our system demonstrates high accuracy and speed in detecting cheating in real-time scenarios, providing valuable information, and aiding proctors in decision-making. This article outlines our methodology and the effectiveness of our system in mitigating the widespread problem of cheating in online exams.
Modeling the correlations among errors is closely associated with how accurately the model can quantify predictive uncertainty in probabilistic time series forecasting. Recent multivariate models have made significant progress in accounting for contemporaneous correlations among errors, while a common assumption on these errors is that they are temporally independent for the sake of statistical simplicity. However, real-world observations often deviate from this assumption, since errors usually exhibit substantial autocorrelation due to various factors such as the exclusion of temporally correlated covariates. In this work, we propose an efficient method, based on a low-rank-plus-diagonal parameterization of the covariance matrix, which can effectively characterize the autocorrelation of errors. The proposed method possesses several desirable properties: the complexity does not scale with the number of time series, the resulting covariance can be used for calibrating predictions, and it can seamlessly integrate with any model with Gaussian-distributed errors. We empirically demonstrate these properties using two distinct neural forecasting models -- GPVar and Transformer. Our experimental results confirm the effectiveness of our method in enhancing predictive accuracy and the quality of uncertainty quantification on multiple real-world datasets.
The attention mechanism, a cornerstone of state-of-the-art neural models, faces computational hurdles in processing long sequences due to its quadratic complexity. Consequently, research efforts in the last few years focused on finding more efficient alternatives. Among them, Hyena (Poli et al., 2023) stands out for achieving competitive results in both language modeling and image classification, while offering sub-quadratic memory and computational complexity. Building on these promising results, we propose ConfHyena, a Conformer whose encoder self-attentions are replaced with an adaptation of Hyena for speech processing, where the long input sequences cause high computational costs. Through experiments in automatic speech recognition (for English) and translation (from English into 8 target languages), we show that our best ConfHyena model significantly reduces the training time by 27%, at the cost of minimal quality degradation (~1%), which, in most cases, is not statistically significant.