Search

With the digitization of health data, the growth of electronic health and medical records lowers barriers for using algorithmic techniques for data analysis. While classical machine learning techniques for health data approach commercialization, there is not yet clear evidence whether quantum machine learning (QML) will provide any empirical advantage for digital health data processing. In this systematic literature review we assess whether QML algorithms have the potential to outperform existing classical methods in efficacy or efficiency. We include digital electronic health/medical records (EH/MRs) and data considered to be a reasonable proxy to EH/MRs. Eligible QML algorithms must be designed for quantum computing hardware, as opposed to quantum-inspired techniques. PubMed, Embase, IEEE, Scopus and arXiv yielded 4915 studies between 2015 to 10 June 2024. After screening 169 eligible studies, most studies contained widespread technical misconceptions about QML and we excluded 123 studies for insufficient rigor in analysis. Of the remaining 46 studies, only 16 studies consider realistic QML operating conditions, either by testing algorithms on quantum hardware, or using noisy quantum circuits when assessing QML algorithms. We find QML applications in digital health focus primarily on clinical decision support rather than health service delivery or public health. Nearly all QML models are linear quantum models, and therefore represent a subset of general quantum algorithms. Meanwhile, novel data-encoding strategies do not address scalability issues, except in regimes requiring restrictive assumptions about quantum hardware, rendering these protocols inefficient for the general encoding of large health datasets. By establishing the current state of evidence for QML-based health applications, we pave the way for meaningful dialogue about QML use-case discovery in digital health., Comment: 16 pages, 9 figures. Feedback welcome

Accurate body weight estimation is critical in emergency medicine for proper dosing of weight-based medications, yet direct measurement is often impractical in urgent situations. This paper presents a non-invasive method for estimating body weight by calculating total body volume and height using 3D imaging technology. A RealSense D415 camera is employed to capture high-resolution depth maps of the patient, from which 3D models are generated. The Convex Hull Algorithm is then applied to calculate the total body volume, with enhanced accuracy achieved by segmenting the point cloud data into multiple sections and summing their individual volumes. The height is derived from the 3D model by identifying the distance between key points on the body. This combined approach provides an accurate estimate of body weight, improving the reliability of medical interventions where precise weight data is unavailable. The proposed method demonstrates significant potential to enhance patient safety and treatment outcomes in emergency settings., Comment: 6 pages

Local amplitude aberrations caused by scintillation can impact the reconstruction process of a wavefront sensor (WFS) by inducing a spatially non-uniform intensity at the pupil plane. This effect is especially relevant for the commonly-used Shack-Hartmann WFS (SHWFS), which can lose slope information for portions of the beam where the signal is faint, leading to reduced reconstruction performance and eventually total failure as the level of scintillation increases. An alternative WFS is needed for such conditions. The nonlinear curvature wavefront sensor (nlCWFS) has been shown to achieve better sensitivity compared to the SHWFS under low light levels. Additionally, the nlCWFS has demonstrated the ability to maintain its sensitivity in the presence of scintillation, using amplitude aberrations to help inform the reconstruction process, rather than hinder. Experiments to date have thus far only shown reconstruction results for a single scintillation value. Building upon previous simulations and laboratory experiments, we have built a testbed to quantify the effects of varying scintillation strength on the wavefront reconstruction performance of the nlCWFS compared to an equivalent SHWFS. In this paper, we present results showing the difference in performance between the nlCWFS and SHWFS as a function of relative flux and scintillation strength., Comment: 6 pages, 3 figures, SPIE Astronomical Telescopes + Instrumentation 2024

We present algorithms that substantially accelerate partition-based cross-validation for machine learning models that require matrix products $\mathbf{X}^\mathbf{T}\mathbf{X}$ and $\mathbf{X}^\mathbf{T}\mathbf{Y}$. Our algorithms have applications in model selection for, e.g., principal component analysis (PCA), principal component regression (PCR), ridge regression (RR), ordinary least squares (OLS), and partial least squares (PLS). Our algorithms support all combinations of column-wise centering and scaling of $\mathbf{X}$ and $\mathbf{Y}$, and we demonstrate in our accompanying implementation that this adds only a manageable, practical constant over efficient variants without preprocessing. We prove the correctness of our algorithms under a fold-based partitioning scheme and show that the running time is independent of the number of folds; that is, they have the same time complexity as that of computing $\mathbf{X}^\mathbf{T}\mathbf{X}$ and $\mathbf{X}^\mathbf{T}\mathbf{Y}$ and space complexity equivalent to storing $\mathbf{X}$, $\mathbf{Y}$, $\mathbf{X}^\mathbf{T}\mathbf{X}$, and $\mathbf{X}^\mathbf{T}\mathbf{Y}$. Importantly, unlike alternatives found in the literature, we avoid data leakage due to preprocessing. We achieve these results by eliminating redundant computations in the overlap between training partitions. Concretely, we show how to manipulate $\mathbf{X}^\mathbf{T}\mathbf{X}$ and $\mathbf{X}^\mathbf{T}\mathbf{Y}$ using only samples from the validation partition to obtain the preprocessed training partition-wise $\mathbf{X}^\mathbf{T}\mathbf{X}$ and $\mathbf{X}^\mathbf{T}\mathbf{Y}$. To our knowledge, we are the first to derive correct and efficient cross-validation algorithms for any of the $16$ combinations of column-wise centering and scaling, for which we also prove only $12$ give distinct matrix products., Comment: 31 pages, 2 tables, 1 figure, 7 algorithms

When selecting data for training large-scale models, standard practice is to filter for examples that match human notions of data quality. Such filtering yields qualitatively clean datapoints that intuitively should improve model behavior. However, in practice the opposite can often happen: we find that selecting according to similarity with "high quality" data sources may not increase (and can even hurt) performance compared to randomly selecting data. To develop better methods for selecting data, we start by framing dataset selection as an optimization problem that we can directly solve for: given target tasks, a learning algorithm, and candidate data, select the subset that maximizes model performance. This framework thus avoids handpicked notions of data quality, and instead models explicitly how the learning process uses train datapoints to predict on the target tasks. Our resulting method greatly improves language model (LM) performance on both pre-specified tasks and previously unseen tasks. Specifically, choosing target tasks representative of standard LM problems and evaluating on diverse held-out benchmarks, our selected datasets provide a 2x compute multiplier over baseline methods.

There is an increasing need to comprehensively characterize the kinematic performances of different Micromobility Vehicles (MMVs). This study aims to: 1) characterize the kinematic behaviors of different MMVs during emergency maneuvers; 2) explore the influence of different MMV power sources on the device performances; 3) investigate if piecewise linear models are suitable for modeling MMV trajectories. A test track experiment where 40 frequent riders performed emergency braking and swerving maneuvers riding a subset of electric MMVs, their traditional counterparts, and, in some cases, behaving as running pedestrians. A second experiment was conducted to determine the MMVs swerving lower boundaries. Device power source resulted having a statistically significant influence on kinematic capabilities of the MMVs: while e-MMVs displayed superior braking capabilities compared to their traditional counterparts, the opposite was observed in terms of swerving performance. Furthermore, performances varied significantly across the different MMV typologies, with handlebar-based devices consistently outperforming the handlebar-less devices across the metrics considered. The piecewise linear models used for braking profiles fit well for most MMVs, except for skateboards and pedestrians due to foot-ground engagement. These findings underscore that the effectiveness of steering or braking in preventing collisions may vary depending on the type and power source of the device. This study also demonstrates the applicability of piecewise linear models for generating parameterized functions that accurately model braking trajectories, providing a valuable resource for automated systems developers. The model, however, also reveals that the single brake ramp assumption does not apply for certain types of MMVs or for pedestrians, indicating the necessity for further improvements., Comment: 21 pages, 8 figures

Based on previous work, we assess the use of NIR-HSI images for calibrating models on two datasets, focusing on protein content regression and grain variety classification. Limited reference data for protein content is expanded by subsampling and associating it with the bulk sample. However, this method introduces significant biases due to skewed leptokurtic prediction distributions, affecting both PLS-R and deep CNN models. We propose adjustments to mitigate these biases, improving mean protein reference predictions. Additionally, we investigate the impact of grain-to-background ratios on both tasks. Higher ratios yield more accurate predictions, but including lower-ratio images in calibration enhances model robustness for such scenarios., Comment: 19 pages, 17 figures

We present FFCV, a library for easy and fast machine learning model training. FFCV speeds up model training by eliminating (often subtle) data bottlenecks from the training process. In particular, we combine techniques such as an efficient file storage format, caching, data pre-loading, asynchronous data transfer, and just-in-time compilation to (a) make data loading and transfer significantly more efficient, ensuring that GPUs can reach full utilization; and (b) offload as much data processing as possible to the CPU asynchronously, freeing GPU cycles for training. Using FFCV, we train ResNet-18 and ResNet-50 on the ImageNet dataset with competitive tradeoff between accuracy and training time. For example, we are able to train an ImageNet ResNet-50 model to 75\% in only 20 mins on a single machine. We demonstrate FFCV's performance, ease-of-use, extensibility, and ability to adapt to resource constraints through several case studies. Detailed installation instructions, documentation, and Slack support channel are available at https://ffcv.io/ .

Stack Overflow and other similar forums are used commonly by developers to seek answers for their software development as well as privacy-related concerns. Recently, ChatGPT has been used as an alternative to generate code or produce responses to developers' questions. In this paper, we aim to understand developers' privacy challenges by evaluating the types of privacy-related questions asked on Stack Overflow. We then conduct a comparative analysis between the accepted responses given by Stack Overflow users and the responses produced by ChatGPT for those extracted questions to identify if ChatGPT could serve as a viable alternative. Our results show that most privacy-related questions are related to choice/consent, aggregation, and identification. Furthermore, our findings illustrate that ChatGPT generates similarly correct responses for about 56% of questions, while for the rest of the responses, the answers from Stack Overflow are slightly more accurate than ChatGPT., Comment: Submitted to the 10th International Workshop on Evolving Security & Privacy Requirements Engineering (ESPRE'23) co-located with the 31st IEEE International Requirements Engineering Conference September 4-8, 2023, Leibniz Universit\"at, Hannover, Germany

This paper presents an overview of Waymo's approach to building a reliable case for safety - a novel and thorough blueprint for use by any company building fully autonomous driving systems. A safety case for fully autonomous operations is a formal way to explain how a company determines that an AV system is safe enough to be deployed on public roads without a human driver, and it includes evidence to support that determination. It involves an explanation of the system, the methodologies used to develop it, the metrics used to validate it and the actual results of validation tests. Yet, in order to develop a worthwhile safety case, it is first important to understand what makes one credible and well crafted, and align on evaluation criteria. This paper helps enabling such alignment by providing foundational thinking into not only how a system is determined to be ready for deployment but also into justifying that the set of acceptance criteria employed in such determination is sufficient and that their evaluation (and associated methods) is credible. The publication is structured around three complementary perspectives on safety that build upon content published by Waymo since 2020: a layered approach to safety; a dynamic approach to safety; and a credible approach to safety. The proposed approach is methodology-agnostic, so that anyone in the space could employ portions or all of it.

The quantitative measurement of how and when we experience surprise has mostly remained limited to laboratory studies, and its extension to naturalistic settings has been challenging. Here we demonstrate, for the first time, how computational models of surprise rooted in cognitive science and neuroscience combined with state-of-the-art machine learned generative models can be used to detect surprising human behavior in complex, dynamic environments like road traffic. In traffic safety, such models can support the identification of traffic conflicts, modeling of road user response time, and driving behavior evaluation for both human and autonomous drivers. We also present novel approaches to quantify surprise and use naturalistic driving scenarios to demonstrate a number of advantages over existing surprise measures from the literature. Modeling surprising behavior using learned generative models is a novel concept that can be generalized beyond traffic safety to any dynamic real-world environment., Comment: 25 pages, 7 figures

Driver process models play a central role in the testing, verification, and development of automated and autonomous vehicle technologies. Prior models developed from control theory and physics-based rules are limited in automated vehicle applications due to their restricted behavioral repertoire. Data-driven machine learning models are more capable than rule-based models but are limited by the need for large training datasets and their lack of interpretability, i.e., an understandable link between input data and output behaviors. We propose a novel car following modeling approach using active inference, which has comparable behavioral flexibility to data-driven models while maintaining interpretability. We assessed the proposed model, the Active Inference Driving Agent (AIDA), through a benchmark analysis against the rule-based Intelligent Driver Model, and two neural network Behavior Cloning models. The models were trained and tested on a real-world driving dataset using a consistent process. The testing results showed that the AIDA predicted driving controls significantly better than the rule-based Intelligent Driver Model and had similar accuracy to the data-driven neural network models in three out of four evaluations. Subsequent interpretability analyses illustrated that the AIDA's learned distributions were consistent with driver behavior theory and that visualizations of the distributions could be used to directly comprehend the model's decision making process and correct model errors attributable to limited training data. The results indicate that the AIDA is a promising alternative to black-box data-driven models and suggest a need for further research focused on modeling driving style and model training with more diverse datasets.

Deep neural networks have been widely studied for predicting a medical condition, such as total knee replacement (TKR). It has shown that data of different modalities, such as imaging data, clinical variables and demographic information, provide complementary information and thus can improve the prediction accuracy together. However, the data sources of various modalities may not always be of high quality, and each modality may have only partial information of medical condition. Thus, predictions from different modalities can be opposite, and the final prediction may fail in the presence of such a conflict. Therefore, it is important to consider the reliability of each source data and the prediction output when making a final decision. In this paper, we propose an evidence-aware multi-modal data fusion framework based on the Dempster-Shafer theory (DST). The backbone models contain an image branch, a non-image branch and a fusion branch. For each branch, there is an evidence network that takes the extracted features as input and outputs an evidence score, which is designed to represent the reliability of the output from the current branch. The output probabilities along with the evidence scores from multiple branches are combined with the Dempster's combination rule to make a final prediction. Experimental results on the public OA initiative (OAI) dataset for the TKR prediction task show the superiority of the proposed fusion strategy on various backbone models.

Semantic medical image segmentation using deep learning has recently achieved high accuracy, making it appealing to clinical problems such as radiation therapy. However, the lack of high-quality semantically labelled data remains a challenge leading to model brittleness to small shifts to input data. Most works require extra data for semi-supervised learning and lack the interpretability of the boundaries of the training data distribution during training, which is essential for model deployment in clinical practice. We propose a fully supervised generative framework that can achieve generalisable segmentation with only limited labelled data by simultaneously constructing an explorable manifold during training. The proposed approach creates medical image style paired with a segmentation task driven discriminator incorporating end-to-end adversarial training. The discriminator is generalised to small domain shifts as much as permissible by the training data, and the generator automatically diversifies the training samples using a manifold of input features learnt during segmentation. All the while, the discriminator guides the manifold learning by supervising the semantic content and fine-grained features separately during the image diversification. After training, visualisation of the learnt manifold from the generator is available to interpret the model limits. Experiments on a fully semantic, publicly available pelvis dataset demonstrated that our method is more generalisable to shifts than other state-of-the-art methods while being more explainable using an explorable manifold.

Distribution shift is a major source of failure for machine learning models. However, evaluating model reliability under distribution shift can be challenging, especially since it may be difficult to acquire counterfactual examples that exhibit a specified shift. In this work, we introduce the notion of a dataset interface: a framework that, given an input dataset and a user-specified shift, returns instances from that input distribution that exhibit the desired shift. We study a number of natural implementations for such an interface, and find that they often introduce confounding shifts that complicate model evaluation. Motivated by this, we propose a dataset interface implementation that leverages Textual Inversion to tailor generation to the input distribution. We then demonstrate how applying this dataset interface to the ImageNet dataset enables studying model behavior across a diverse array of distribution shifts, including variations in background, lighting, and attributes of the objects. Code available at https://github.com/MadryLab/dataset-interfaces.

In medical image analysis, automated segmentation of multi-component anatomical structures, which often have a spectrum of potential anomalies and pathologies, is a challenging task. In this work, we develop a multi-step approach using U-Net-based neural networks to initially detect anomalies (bone marrow lesions, bone cysts) in the distal femur, proximal tibia and patella from 3D magnetic resonance (MR) images of the knee in individuals with varying grades of osteoarthritis. Subsequently, the extracted data are used for downstream tasks involving semantic segmentation of individual bone and cartilage volumes as well as bone anomalies. For anomaly detection, the U-Net-based models were developed to reconstruct the bone profiles of the femur and tibia in images via inpainting so anomalous bone regions could be replaced with close to normal appearances. The reconstruction error was used to detect bone anomalies. A second anomaly-aware network, which was compared to anomaly-na\"ive segmentation networks, was used to provide a final automated segmentation of the femoral, tibial and patellar bones and cartilages from the knee MR images containing a spectrum of bone anomalies. The anomaly-aware segmentation approach provided up to 58% reduction in Hausdorff distances for bone segmentations compared to the results from the anomaly-na\"ive segmentation networks. In addition, the anomaly-aware networks were able to detect bone lesions in the MR images with greater sensitivity and specificity (area under the receiver operating characteristic curve [AUC] up to 0.896) compared to the anomaly-na\"ive segmentation networks (AUC up to 0.874).

In medical image analysis, the subtle visual characteristics of many diseases are challenging to discern, particularly due to the lack of paired data. For example, in mild Alzheimer's Disease (AD), brain tissue atrophy can be difficult to observe from pure imaging data, especially without paired AD and Cognitively Normal ( CN ) data for comparison. This work presents Disease Discovery GAN ( DiDiGAN), a weakly-supervised style-based framework for discovering and visualising subtle disease features. DiDiGAN learns a disease manifold of AD and CN visual characteristics, and the style codes sampled from this manifold are imposed onto an anatomical structural "blueprint" to synthesise paired AD and CN magnetic resonance images (MRIs). To suppress non-disease-related variations between the generated AD and CN pairs, DiDiGAN leverages a structural constraint with cycle consistency and anti-aliasing to enforce anatomical correspondence. When tested on the Alzheimer's Disease Neuroimaging Initiative ( ADNI) dataset, DiDiGAN showed key AD characteristics (reduced hippocampal volume, ventricular enlargement, and atrophy of cortical structures) through synthesising paired AD and CN scans. The qualitative results were backed up by automated brain volume analysis, where systematic pair-wise reductions in brain tissue structures were also measured

Extremely precise radial velocity (EPRV) measurements are critical for characterizing nearby terrestrial worlds. EPRV instrument precisions of $\sigma_{\mathrm{RV}} = 1-10\,\mathrm{cm/s}$ are required to study Earth-analog systems, imposing stringent, sub-mK, thermo-mechanical stability requirements on Doppler spectrograph designs. iLocater is a new, high-resolution ($R=190,500$ median) near infrared (NIR) EPRV spectrograph under construction for the dual 8.4 m diameter Large Binocular Telescope (LBT). The instrument is one of the first to operate in the diffraction-limited regime enabled by the use of adaptive optics and single-mode fibers. This facilitates affordable optomechanical fabrication of the spectrograph using intrinsically stable materials. We present the final design and performance of the iLocater cryostat and thermal control system which houses the instrument spectrograph. The spectrograph is situated inside an actively temperature-controlled radiation shield mounted inside a multi-layer-insulation (MLI) lined vacuum chamber. The radiation shield provides sub-mK thermal stability, building on the existing heritage of the Habitable-zone Planet Finder (HPF) and NEID instruments. The instrument operating temperature ($T=80-100\,\mathrm{K}$) is driven by the requirement to minimize detector background and instantaneous coefficient of thermal expansion (CTE) of the materials used for spectrograph fabrication. This combination allows for a reduced thermomechanical impact on measurement precision, improving the scientific capabilities of the instrument., Comment: 10 pages, In proceedings of SPIE 12184

In some parameter and solution regimes, a minimally coupled nonrelativistic quantum particle in 1d is isomorphic to a much heavier, vibrating, very thin Euler-Bernoulli rod in 3d, with ratio of bending modulus to linear density $(\hbar/2m)^2$. For $m=m_e$, this quantity is comparable to that of a microtubule. Axial forces and torques applied to the rod play the role of scalar and vector potentials, respectively, and rod inextensibility plays the role of normalization. We show how an uncertainty principle $\Delta x\Delta p_x\gtrsim\hbar$ governs transverse deformations propagating down the inextensible, force and torque-free rod, and how orbital angular momentum quantized in units of $\hbar$ or $\hbar/2$ (depending on calculation method) emerges when the force and torque-free inextensible rod is formed into a ring. For torqued rings with large wavenumbers, a ``twist quantum'' appears that is somewhat analogous to the magnetic flux quantum. These and other results are obtained from a purely classical treatment of the rod, i.e., without quantizing any classical fields., Comment: 7 pages, 1 table, 1 figure, accepted for publication in Phys. Rev. E

Traditional equivalent circuit models (ECMs) have difficulties in estimating battery internal states due to the lack of relevant physics, such as the lithium diffusion in active particles. Here we configure a circuit network to describe the lithium diffusion and define it as a new high-level circuit element called diffusion-aware voltage source. The circuit representation is proven equivalent to the discretized diffusion equation. The new voltage source gives the electrode potential as a function of the surface concentration and thus automatically incorporates the diffusion overpotential. We show that an ECM with the proposed diffusion-aware voltage sources (called "shell ECM") can reproduce the single particle model simulation results, making it a trustworthy easy-to-implement substitute. Furthermore, the simplest shell ECM consisting of a single diffusion-aware voltage source and a resistor is validated against experimental constant-current discharges at various rates. The diffusion-aware voltage source can be used to measure diffusivity by fitting the diffusion resistance against experimental data. The viability of the shell ECM for onboard usage is confirmed by implementation into a battery management system of WAE Technologies. By tracking the internal concentration states, the shell ECM demonstrates robustness to dynamic applied-current profiles., Comment: 35 pages, 14 figures

Using transfer learning to adapt a pre-trained "source model" to a downstream "target task" can dramatically increase performance with seemingly no downside. In this work, we demonstrate that there can exist a downside after all: bias transfer, or the tendency for biases of the source model to persist even after adapting the model to the target class. Through a combination of synthetic and natural experiments, we show that bias transfer both (a) arises in realistic settings (such as when pre-training on ImageNet or other standard datasets) and (b) can occur even when the target dataset is explicitly de-biased. As transfer-learned models are increasingly deployed in the real world, our work highlights the importance of understanding the limitations of pre-trained source models. Code is available at https://github.com/MadryLab/bias-transfer