Search

The INTENTAS project aims to develop an atomic sensor utilizing entangled Bose-Einstein condensates (BECs) in a microgravity environment. This key achievement is necessary to advance the capability for measurements that benefit from both entanglement-enhanced sensitivities and extended interrogation times. The project addresses significant challenges related to size, weight, and power management (SWaP) specific to the experimental platform at the Einstein-Elevator in Hannover. The design ensures a low-noise environment essential for the creation and detection of entanglement. Additionally, the apparatus features an innovative approach to the all-optical creation of BECs, providing a flexible system for various configurations and meeting the requirements for rapid turnaround times. Successful demonstration of this technology in the Einstein-Elevator will pave the way for a future deployment in space, where its potential applications will unlock high-precision quantum sensing.

Current strategies for solving image-based inverse problems apply latent diffusion models to perform posterior sampling.However, almost all approaches make no explicit attempt to explore the solution space, instead drawing only a single sample from a Gaussian distribution from which to generate their solution. In this paper, we introduce a particle-filtering-based framework for a nonlinear exploration of the solution space in the initial stages of reverse SDE methods. Our proposed particle-filtering-based latent diffusion (PFLD) method and proposed problem formulation and framework can be applied to any diffusion-based solution for linear or nonlinear inverse problems. Our experimental results show that PFLD outperforms the SoTA solver PSLD on the FFHQ-1K and ImageNet-1K datasets on inverse problem tasks of super resolution, Gaussian debluring and inpainting., Comment: Mohammad Hadi Sepanj, Nicholas Pellegrino, and Chris Czarnecki contributed equally

As part of an ongoing worldwide effort to comprehend and monitor insect biodiversity, this paper presents the BIOSCAN-5M Insect dataset to the machine learning community and establish several benchmark tasks. BIOSCAN-5M is a comprehensive dataset containing multi-modal information for over 5 million insect specimens, and it significantly expands existing image-based biological datasets by including taxonomic labels, raw nucleotide barcode sequences, assigned barcode index numbers, geographical, and size information. We propose three benchmark experiments to demonstrate the impact of the multi-modal data types on the classification and clustering accuracy. First, we pretrain a masked language model on the DNA barcode sequences of the BIOSCAN-5M dataset, and demonstrate the impact of using this large reference library on species- and genus-level classification performance. Second, we propose a zero-shot transfer learning task applied to images and DNA barcodes to cluster feature embeddings obtained from self-supervised learning, to investigate whether meaningful clusters can be derived from these representation embeddings. Third, we benchmark multi-modality by performing contrastive learning on DNA barcodes, image data, and taxonomic information. This yields a general shared embedding space enabling taxonomic classification using multiple types of information and modalities. The code repository of the BIOSCAN-5M Insect dataset is available at https://github.com/bioscan-ml/BIOSCAN-5M.

Machine comprehension of visual information from images and videos by neural networks faces two primary challenges. Firstly, there exists a computational and inference gap in connecting vision and language, making it difficult to accurately determine which object a given agent acts on and represent it through language. Secondly, classifiers trained by a single, monolithic neural network often lack stability and generalization. To overcome these challenges, we introduce MoE-VRD, a novel approach to visual relationship detection utilizing a mixture of experts. MoE-VRD identifies language triplets in the form of < subject, predicate, object> tuples to extract relationships from visual processing. Leveraging recent advancements in visual relationship detection, MoE-VRD addresses the requirement for action recognition in establishing relationships between subjects (acting) and objects (being acted upon). In contrast to single monolithic networks, MoE-VRD employs multiple small models as experts, whose outputs are aggregated. Each expert in MoE-VRD specializes in visual relationship learning and object tagging. By utilizing a sparsely-gated mixture of experts, MoE-VRD enables conditional computation and significantly enhances neural network capacity without increasing computational complexity. Our experimental results demonstrate that the conditional computation capabilities and scalability of the mixture-of-experts approach lead to superior performance in visual relationship detection compared to state-of-the-art methods.

Single-image super-resolution (SISR) has seen significant advancements through the integration of deep learning. However, the substantial computational and memory requirements of existing methods often limit their practical application. This paper introduces a new Efficient Pyramid Network (EPNet) that harmoniously merges an Edge Split Pyramid Module (ESPM) with a Panoramic Feature Extraction Module (PFEM) to overcome the limitations of existing methods, particularly in terms of computational efficiency. The ESPM applies a pyramid-based channel separation strategy, boosting feature extraction while maintaining computational efficiency. The PFEM, a novel fusion of CNN and Transformer structures, enables the concurrent extraction of local and global features, thereby providing a panoramic view of the image landscape. Our architecture integrates the PFEM in a manner that facilitates the streamlined exchange of feature information and allows for the further refinement of image texture details. Experimental results indicate that our model outperforms existing state-of-the-art methods in image resolution quality, while considerably decreasing computational and memory costs. This research contributes to the ongoing evolution of efficient and practical SISR methodologies, bearing broader implications for the field of computer vision.

Dietary intake data are routinely drawn upon to explore diet-health relationships. However, these data are often subject to measurement error, distorting the true relationships. Beyond measurement error, there are likely complex synergistic and sometimes antagonistic interactions between different dietary components, complicating the relationships between diet and health outcomes. Flexible models are required to capture the nuance that these complex interactions introduce. This complexity makes research on diet-health relationships an appealing candidate for the application of machine learning techniques, and in particular, neural networks. Neural networks are computational models that are able to capture highly complex, nonlinear relationships so long as sufficient data are available. While these models have been applied in many domains, the impacts of measurement error on the performance of predictive modeling has not been systematically investigated. However, dietary intake data are typically collected using self-report methods and are prone to large amounts of measurement error. In this work, we demonstrate the ways in which measurement error erodes the performance of neural networks, and illustrate the care that is required for leveraging these models in the presence of error. We demonstrate the role that sample size and replicate measurements play on model performance, indicate a motivation for the investigation of transformations to additivity, and illustrate the caution required to prevent model overfitting. While the past performance of neural networks across various domains make them an attractive candidate for examining diet-health relationships, our work demonstrates that substantial care and further methodological development are both required to observe increased predictive performance when applying these techniques, compared to more traditional statistical procedures.

Recent state-of-the-art semi-supervised Video Object Segmentation (VOS) methods have shown significant improvements in target object segmentation accuracy when information from preceding frames is used in segmenting the current frame. In particular, such memory-based approaches can help a model to more effectively handle appearance changes (representation drift) or occlusions. Ideally, for maximum performance, Online VOS methods would need all or most of the preceding frames (or their extracted information) to be stored in memory and be used for online learning in later frames. Such a solution is not feasible for long videos, as the required memory size grows without bound, and such methods can fail when memory is limited and a target object experiences repeated representation drifts throughout a video. We propose two novel techniques to reduce the memory requirement of Online VOS methods while improving modeling accuracy and generalization on long videos. Motivated by the success of continual learning techniques in preserving previously-learned knowledge, here we propose Gated-Regularizer Continual Learning (GRCL), which improves the performance of any Online VOS subject to limited memory, and a Reconstruction-based Memory Selection Continual Learning (RMSCL), which empowers Online VOS methods to efficiently benefit from stored information in memory. We also analyze the performance of a hybrid combination of the two proposed methods. Experimental results show that the proposed methods are able to improve the performance of Online VOS models by more than 8%, with improved robustness on long-video datasets while maintaining comparable performance on short-video datasets such as DAVIS16, DAVIS17, and YouTube-VOS18.

In an effort to catalog insect biodiversity, we propose a new large dataset of hand-labelled insect images, the BIOSCAN-Insect Dataset. Each record is taxonomically classified by an expert, and also has associated genetic information including raw nucleotide barcode sequences and assigned barcode index numbers, which are genetically-based proxies for species classification. This paper presents a curated million-image dataset, primarily to train computer-vision models capable of providing image-based taxonomic assessment, however, the dataset also presents compelling characteristics, the study of which would be of interest to the broader machine learning community. Driven by the biological nature inherent to the dataset, a characteristic long-tailed class-imbalance distribution is exhibited. Furthermore, taxonomic labelling is a hierarchical classification scheme, presenting a highly fine-grained classification problem at lower levels. Beyond spurring interest in biodiversity research within the machine learning community, progress on creating an image-based taxonomic classifier will also further the ultimate goal of all BIOSCAN research: to lay the foundation for a comprehensive survey of global biodiversity. This paper introduces the dataset and explores the classification task through the implementation and analysis of a baseline classifier.

Accurate and fast histological staining is crucial in histopathology, impacting diagnostic precision and reliability. Traditional staining methods are time-consuming and subjective, causing delays in diagnosis. Digital pathology plays a vital role in advancing and optimizing histology processes to improve efficiency and reduce turnaround times. This study introduces a novel deep learning-based framework for virtual histological staining using photon absorption remote sensing (PARS) images. By extracting features from PARS time-resolved signals using a variant of the K-means method, valuable multi-modal information is captured. The proposed multi-channel cycleGAN (MC-GAN) model expands on the traditional cycleGAN framework, allowing the inclusion of additional features. Experimental results reveal that specific combinations of features outperform the conventional channels by improving the labeling of tissue structures prior to model training. Applied to human skin and mouse brain tissue, the results underscore the significance of choosing the optimal combination of features, as it reveals a substantial visual and quantitative concurrence between the virtually stained and the gold standard chemically stained hematoxylin and eosin (H&E) images, surpassing the performance of other feature combinations. Accurate virtual staining is valuable for reliable diagnostic information, aiding pathologists in disease classification, grading, and treatment planning. This study aims to advance label-free histological imaging and opens doors for intraoperative microscopy applications., Comment: 20 pages, 6 figures

Modern histopathology relies on the microscopic examination of thin tissue sections stained with histochemical techniques, typically using brightfield or fluorescence microscopy. However, the staining of samples can permanently alter their chemistry and structure, meaning an individual tissue section must be prepared for each desired staining contrast. This not only consumes valuable tissue samples but also introduces delays in essential diagnostic timelines. In this work, virtual histochemical staining is developed using label-free photon absorption remote sensing (PARS) microscopy. We present a method that generates virtually stained histology images that are indistinguishable from the gold standard hematoxylin and eosin (H&E) staining. First, PARS label-free ultraviolet absorption images are captured directly within unstained tissue specimens. The radiative and non-radiative absorption images are then preprocessed, and virtually stained through the presented pathway. The preprocessing pipeline features a self-supervised Noise2Void denoising convolutional neural network (CNN) as well as a novel algorithm for pixel-level mechanical scanning error correction. These developments significantly enhance the recovery of sub-micron tissue structures, such as nucleoli location and chromatin distribution. Finally, we used a cycle-consistent generative adversarial network CycleGAN architecture to virtually stain the preprocessed PARS data. Virtual staining is applied to thin unstained sections of malignant human skin and breast tissue samples. Clinically relevant details are revealed, with comparable contrast and quality to gold standard H&E-stained images. This work represents a crucial step to deploying label-free microscopy as an alternative to standard histopathology techniques., Comment: 18 Pages, 8 figures

Unsupervised domain adaptation (UDA) aims to bridge the gap between source and target domains in the absence of target domain labels using two main techniques: input-level alignment (such as generative modeling and stylization) and feature-level alignment (which matches the distribution of the feature maps, e.g. gradient reversal layers). Motivated from the success of generative modeling for image classification, stylization-based methods were recently proposed for regression tasks, such as pose estimation. However, use of input-level alignment via generative modeling and stylization incur additional overhead and computational complexity which limit their use in real-world DA tasks. To investigate the role of input-level alignment for DA, we ask the following question: Is generative modeling-based stylization necessary for visual domain adaptation in regression? Surprisingly, we find that input-alignment has little effect on regression tasks as compared to classification. Based on these insights, we develop a non-parametric feature-level domain alignment method -- Implicit Stylization (ImSty) -- which results in consistent improvements over SOTA regression task, without the need for computationally intensive stylization and generative modeling. Our work conducts a critical evaluation of the role of generative modeling and stylization, at a time when these are also gaining popularity for domain generalization.

Continual learning in real-world scenarios is a major challenge. A general continual learning model should have a constant memory size and no predefined task boundaries, as is the case in semi-supervised Video Object Segmentation (VOS), where continual learning challenges particularly present themselves in working on long video sequences. In this article, we first formulate the problem of semi-supervised VOS, specifically online VOS, as a continual learning problem, and then secondly provide a public VOS dataset, CLVOS23, focusing on continual learning. Finally, we propose and implement a regularization-based continual learning approach on LWL, an existing online VOS baseline, to demonstrate the efficacy of continual learning when applied to online VOS and to establish a CLVOS23 baseline. We apply the proposed baseline to the Long Videos dataset as well as to two short video VOS datasets, DAVIS16 and DAVIS17. To the best of our knowledge, this is the first time that VOS has been defined and addressed as a continual learning problem.

Mobile microscopic bodies, such as motile cells, can be modelled phenomenologically as ``active particles'' which can move against external forces by depleting an internal energy depot. The microscopic mechanisms underlying such ``active'' behavior must ultimately obey fundamental physics: energy depots must actually consist of dynamical degrees of freedom, such as chemical reaction coordinates, which in some way couple to the particle's motional degrees of freedom. As a step towards connecting phenomenological models with microscopic dynamical mechanisms, recent papers have studied the minimalistic dynamical mechanism of a ``dynamical active particle'', and shown how nonlinear couplings can allow steady energy transfer from depot to motion, even in the presence of weak dissipation. Most real active particles move through viscous environments, however, and are strongly damped. Here we therefore generalize the dynamical active particle into the overdamped regime. We find that its mechanism still operates, and in particular allows the overdamped active particle to travel just as far against friction as the undamped model, by moving at a slower average speed. Our results suggest that active particle phenomenology can indeed be consistent with comprehensible dynamical mechanisms, even in strongly dissipative environments.

The BIOSCAN project, led by the International Barcode of Life Consortium, seeks to study changes in biodiversity on a global scale. One component of the project is focused on studying the species interaction and dynamics of all insects. In addition to genetically barcoding insects, over 1.5 million images per year will be collected, each needing taxonomic classification. With the immense volume of incoming images, relying solely on expert taxonomists to label the images would be impossible; however, artificial intelligence and computer vision technology may offer a viable high-throughput solution. Additional tasks including manually weighing individual insects to determine biomass, remain tedious and costly. Here again, computer vision may offer an efficient and compelling alternative. While the use of computer vision methods is appealing for addressing these problems, significant challenges resulting from biological factors present themselves. These challenges are formulated in the context of machine learning in this paper., Comment: 4 pages, 3 figures. Submitted to the Journal of Computational Vision and Imaging Systems

Useful dynamical processes often begin through barrier-crossing dynamical transitions; engineering system dynamics in order to make such transitions reliably is therefore an important task for biological or artificial microscopic machinery. Here we first show by example that adding even a small amount of back-reaction to a control parameter, so that it responds to the system's evolution, can significantly increase the fraction of trajectories that cross a separatrix. We then explain how a post-adiabatic theorem due to Neishtadt can quantitatively describe this kind of enhancement without having to solve the equations of motion, allowing systematic understanding and design of a class of self-controlling dynamical systems.

Egocentric 3D human pose estimation (HPE) from images is challenging due to severe self-occlusions and strong distortion introduced by the fish-eye view from the head mounted camera. Although existing works use intermediate heatmap-based representations to counter distortion with some success, addressing self-occlusion remains an open problem. In this work, we leverage information from past frames to guide our self-attention-based 3D HPE estimation procedure -- Ego-STAN. Specifically, we build a spatio-temporal Transformer model that attends to semantically rich convolutional neural network-based feature maps. We also propose feature map tokens: a new set of learnable parameters to attend to these feature maps. Finally, we demonstrate Ego-STAN's superior performance on the xR-EgoPose dataset where it achieves a 30.6% improvement on the overall mean per-joint position error, while leading to a 22% drop in parameters compared to the state-of-the-art., Comment: 4 pages, Extended abstract, Joint International Workshop on Egocentric Perception, Interaction and Computing (EPIC) and Ego4D, IEEE/CVF Computer Vision and Pattern Recognition Conference (CVPR), 2022

Librational motion, whereby a rigid body undergoes angular oscillation around a preferred direction, can be observed in optically trapped, silica microspheres. We demonstrate the cooling of one librational degree of freedom for $\sim 5~\mu$m diameter spheres that have been induced to rotate with an external electric field coupled to their electric dipole moment. Cooling is accomplished by adding a phase modulation to the rotating field. The degree of cooling is quantified by applying a $\pi/2$ shift to the phase of the electric field and fitting the resulting exponential decay of the librational motion to obtain a damping time, as well as estimating a mode temperature from the observed libration in equilibrium. The result is an important step in the study of the dynamics of trapped microspheres, crucial to cooling the mechanical motion to its ground state, as well as providing insights regarding the charge mobility in the material at microscopic scales., Comment: 8 pages, 7 figures

Photoacoustic Remote Sensing (PARS) microscopy is an emerging label-free optical absorption imaging modality. PARS operates by capturing nanosecond-scale optical perturbations generated by photoacoustic pressures. These time-domain (TD) modulations are usually projected by amplitude to determine absorption magnitude. However, significant information on the target's material properties is contained within the TD signals. This work proposes a novel clustering method to learn TD features which relate to underlying biomolecule characteristics. This technique identifies features related to constituent biomolecules, enabling single-acquisition virtual tissue labelling. Colorized visualizations of tissue are produced, highlighting specific tissue components. This is demonstrated on freshly resected murine brain tissue, clearly discerning structures including myelinated and unmyelinated neurons (white and gray matter) and nuclear structures., Comment: 4 pages, 4 figures

Histopathological visualizations are a pillar of modern medicine and biological research. Surgical oncology relies exclusively on post-operative histology to determine definitive surgical success and guide adjuvant treatments. The current histology workflow is based on bright-field microscopic assessment of histochemical stained tissues and has some major limitations. For example, the preparation of stained specimens for brightfield assessment requires lengthy sample processing, delaying interventions for days or even weeks. Hence, there is a pressing need for improved histopathology methods. In this paper, we present a deep-learning-based approach for virtual label-free histochemical staining of total-absorption photoacoustic remote sensing (TA-PARS) images of unstained tissue. TA-PARS provides an array of directly measured label-free contrasts such as scattering and total absorption (radiative and non-radiative), ideal for developing H&E colorizations without the need to infer arbitrary tissue structures. We use a Pix2Pix generative adversarial network (GAN) to develop visualizations analogous to H&E staining from label-free TA-PARS images. Thin sections of human skin tissue were first virtually stained with the TA-PARS, then were chemically stained with H&E producing a one-to-one comparison between the virtual and chemical staining. The one-to-one matched virtually- and chemically- stained images exhibit high concordance validating the digital colorization of the TA-PARS images against the gold standard H&E. TA-PARS images were reviewed by four dermatologic pathologists who confirmed they are of diagnostic quality, and that resolution, contrast, and color permitted interpretation as if they were H&E. The presented approach paves the way for the development of TA-PARS slide-free histology, which promises to dramatically reduce the time from specimen resection to histological imaging., Comment: 16 pages, 8 figures

The nature of dark matter and properties of neutrinos are among the most pressing issues in contemporary particle physics. The dual-phase xenon time-projection chamber is the leading technology to cover the available parameter space for Weakly Interacting Massive Particles (WIMPs), while featuring extensive sensitivity to many alternative dark matter candidates. These detectors can also study neutrinos through neutrinoless double-beta decay and through a variety of astrophysical sources. A next-generation xenon-based detector will therefore be a true multi-purpose observatory to significantly advance particle physics, nuclear physics, astrophysics, solar physics, and cosmology. This review article presents the science cases for such a detector., Comment: 77 pages, 40 figures, 1262 references

Many measurement modalities which perform imaging by probing an object pixel-by-pixel, such as via Photoacoustic Microscopy, produce a multi-dimensional feature (typically a time-domain signal) at each pixel. In principle, the many degrees of freedom in the time-domain signal would admit the possibility of significant multi-modal information being implicitly present, much more than a single scalar "brightness", regarding the underlying targets being observed. However, the measured signal is neither a weighted-sum of basis functions (such as principal components) nor one of a set of prototypes (K-means), which has motivated the novel clustering method proposed here. Signals are clustered based on their shape, but not amplitude, via angular distance and centroids are calculated as the direction of maximal intra-cluster variance, resulting in a clustering algorithm capable of learning centroids (signal shapes) that are related to the underlying, albeit unknown, target characteristics in a scalable and noise-robust manner., Comment: Under consideration at Pattern Recognition Letters. 6 pages (excluding references), 5 figures

Optically levitated macroscopic objects are a powerful tool in the field of force sensing, owing to high sensitivity, absolute force calibration, environmental isolation and the advanced degree of control over their dynamics that have been achieved. However, limitations arise from the spurious forces caused by electrical polarization effects that, even for nominally neutral objects, affect the force sensing because of the interaction of dipole moments with gradients of external electric fields. In this paper we introduce a new technique to model and eliminate dipole moment interactions limiting the performance of sensors employing levitated objects. This process leads to the first noise-limited measurement with a sensitivity of $3.3\times10^{-5}e$. As a demonstration, this is applied to the search for unknown charges of a magnitude much below that of an electron or for exceedingly small unbalances between electron and proton charges. The absence of remaining systematic biases, enables true discovery experiments, with sensitivities that are expected to improve as the system noise is brought down to or beyond the quantum limit. As a by-product of the technique, the electromagnetic properties of the levitated objects can also be measured on an individual basis.

In this paper we investigate a variety of deep learning strategies for solving inverse problems. We classify existing deep learning solutions for inverse problems into three categories of Direct Mapping, Data Consistency Optimizer, and Deep Regularizer. We choose a sample of each inverse problem type, so as to compare the robustness of the three categories, and report a statistical analysis of their differences. We perform extensive experiments on the classic problem of linear regression and three well-known inverse problems in computer vision, namely image denoising, 3D human face inverse rendering, and object tracking, selected as representative prototypes for each class of inverse problems. The overall results and the statistical analyses show that the solution categories have a robustness behaviour dependent on the type of inverse problem domain, and specifically dependent on whether or not the problem includes measurement outliers. Based on our experimental results, we conclude by proposing the most robust solution category for each inverse problem class.

We have developed a multimodal photoacoustic remote sensing (PARS) microscope combined with swept source optical coherence tomography for in vivo, non-contact retinal imaging. Building on the proven strength of multiwavelength PARS imaging, the system is applied for estimating retinal oxygen saturation in the rat retina. The capability of the technology is demonstrated by imaging both microanatomy and the microvasculature of the retina in vivo. To our knowledge this is the first time a non-contact photoacoustic imaging technique is employed for in vivo oxygen saturation measurement in the retina.

We examine a Hamiltonian system which represents an active Brownian particle that can move against an external force by drawing energy from an internal depot while immersed in a noisy and dissipative environment. The Hamiltonian consists of two subsystems, one representing the active particle's motion and the other its depot of `fuel'. We show that although the active particle loses some of its energy to dissipation from the environment, dissipation can also help to stabilize the dynamical process that makes the particle active. Additionally we show how a Hamiltonian active particle can harvest energy from its environment.

Silicon photomultipliers are regarded as a very promising technology for next-generation, cutting-edge detectors for low-background experiments in particle physics. This work presents systematic reflectivity studies of Silicon Photomultipliers (SiPM) and other samples in liquid xenon at vacuum ultraviolet (VUV) wavelengths. A dedicated setup at the University of M\"unster has been used that allows to acquire angle-resolved reflection measurements of various samples immersed in liquid xenon with 0.45{\deg} angular resolution. Four samples are investigated in this work: one Hamamatsu VUV4 SiPM, one FBK VUV-HD SiPM, one FBK wafer sample and one Large-Area Avalanche Photodiode (LA-APD) from EXO-200. The reflectivity is determined to be 25-36% at an angle of incidence of 20{\deg} for the four samples and increases to up to 65% at 70{\deg} for the LA-APD and the FBK samples. The Hamamatsu VUV4 SiPM shows a decline with increasing angle of incidence. The reflectivity results will be incorporated in upcoming light response simulations of the nEXO detector., Comment: 18 pages, 11 figures

Optimal performance is critical for decision-making tasks from medicine to autonomous driving, however common performance measures may be too general or too specific. For binary classifiers, diagnostic tests or prognosis at a timepoint, measures such as the area under the receiver operating characteristic curve, or the area under the precision recall curve, are too general because they include unrealistic decision thresholds. On the other hand, measures such as accuracy, sensitivity or the F1 score are measures at a single threshold that reflect an individual single probability or predicted risk, rather than a range of individuals or risk. We propose a method in between, deep ROC analysis, that examines groups of probabilities or predicted risks for more insightful analysis. We translate esoteric measures into familiar terms: AUC and the normalized concordant partial AUC are balanced average accuracy (a new finding); the normalized partial AUC is average sensitivity; and the normalized horizontal partial AUC is average specificity. Along with post-test measures, we provide a method that can improve model selection in some cases and provide interpretation and assurance for patients in each risk group. We demonstrate deep ROC analysis in two case studies and provide a toolkit in Python., Comment: 14 pages, 6 Figures, submitted to IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI), currently under review

We report on a search for non-Newtonian forces that couple to mass, with a characteristic scale of ${\sim}10~\mu$m, using an optically levitated microsphere as a precision force sensor. A silica microsphere trapped in an upward-propagating, single-beam, optical tweezer is utilized to probe for interactions sourced from a nanofabricated attractor mass with a density modulation brought into close proximity to the microsphere and driven along the axis of periodic density in order to excite an oscillating response. We obtain force sensitivity of ${\lesssim}10^{-16}~\rm{N}/\sqrt{\rm{Hz}}$. Separately searching for attractive and repulsive forces results in the constraint on a new Yukawa interaction of $|\alpha| \gtrsim 10^8$ for $\lambda > 10~\mu$m. This is the first test of the inverse-square law using an optically levitated test mass of dimensions comparable to $\lambda$, a complementary method subject to a different set of systematic effects compared to more established techniques.

In recent years a vast amount of visual content has been generated and shared from many fields, such as social media platforms, medical imaging, and robotics. This abundance of content creation and sharing has introduced new challenges, particularly that of searching databases for similar content-Content Based Image Retrieval (CBIR)-a long-established research area in which improved efficiency and accuracy are needed for real-time retrieval. Artificial intelligence has made progress in CBIR and has significantly facilitated the process of instance search. In this survey we review recent instance retrieval works that are developed based on deep learning algorithms and techniques, with the survey organized by deep network architecture types, deep features, feature embedding and aggregation methods, and network fine-tuning strategies. Our survey considers a wide variety of recent methods, whereby we identify milestone work, reveal connections among various methods and present the commonly used benchmarks, evaluation results, common challenges, and propose promising future directions., Comment: IEEE Transactions on Pattern Analysis and Machine Intelligence

Uncertainty quantification (UQ) plays a pivotal role in reduction of uncertainties during both optimization and decision making processes. It can be applied to solve a variety of real-world applications in science and engineering. Bayesian approximation and ensemble learning techniques are two most widely-used UQ methods in the literature. In this regard, researchers have proposed different UQ methods and examined their performance in a variety of applications such as computer vision (e.g., self-driving cars and object detection), image processing (e.g., image restoration), medical image analysis (e.g., medical image classification and segmentation), natural language processing (e.g., text classification, social media texts and recidivism risk-scoring), bioinformatics, etc. This study reviews recent advances in UQ methods used in deep learning. Moreover, we also investigate the application of these methods in reinforcement learning (RL). Then, we outline a few important applications of UQ methods. Finally, we briefly highlight the fundamental research challenges faced by UQ methods and discuss the future research directions in this field.

Malignant brain tumors are among the deadliest neoplasms with the lowest survival rates of any cancer type. In considering surgical tumor resection, suboptimal extent of resection is linked to poor clinical outcomes and lower overall survival rates. Currently available tools for intraoperative histopathological assessment require an average of 20 minutes processing and are of limited diagnostic quality for guiding surgeries. Consequently, there is an unaddressed need for a rapid imaging technique to guide maximal resection of brain tumors. Working towards this goal, presented here is an all optical non-contact label-free reflection mode photoacoustic remote sensing (PARS) microscope. By using a tunable excitation laser, PARS takes advantage of the endogenous optical absorption peaks of DNA and cytoplasm to achieve virtual contrast analogous to standard hematoxylin and eosin (H and E) staining. In conjunction, a fast 266 nm excitation is used to generate large grossing scans and rapidly assess small fields in real-time with hematoxylin-like contrast. Images obtained using this technique show comparable quality and contrast to the current standard for histopathological assessment of brain tissues. Using the proposed method, rapid, high-throughput, histological-like imaging was achieved in unstained brain tissues, indicating PARS utility for intraoperative guidance to improve extent of surgical resection., Comment: 17 pages, 3 figures

Detection and recognition of text in natural images are two main problems in the field of computer vision that have a wide variety of applications in analysis of sports videos, autonomous driving, industrial automation, to name a few. They face common challenging problems that are factors in how text is represented and affected by several environmental conditions. The current state-of-the-art scene text detection and/or recognition methods have exploited the witnessed advancement in deep learning architectures and reported a superior accuracy on benchmark datasets when tackling multi-resolution and multi-oriented text. However, there are still several remaining challenges affecting text in the wild images that cause existing methods to underperform due to there models are not able to generalize to unseen data and the insufficient labeled data. Thus, unlike previous surveys in this field, the objectives of this survey are as follows: first, offering the reader not only a review on the recent advancement in scene text detection and recognition, but also presenting the results of conducting extensive experiments using a unified evaluation framework that assesses pre-trained models of the selected methods on challenging cases, and applies the same evaluation criteria on these techniques. Second, identifying several existing challenges for detecting or recognizing text in the wild images, namely, in-plane-rotation, multi-oriented and multi-resolution text, perspective distortion, illumination reflection, partial occlusion, complex fonts, and special characters. Finally, the paper also presents insight into the potential research directions in this field to address some of the mentioned challenges that are still encountering scene text detection and recognition techniques.

A high sensitivity force sensor based on dielectric microspheres in vacuum, optically trapped by a single, upward-propagating laser beam, is described. Off-axis parabolic mirrors are used both to focus the 1064~nm trapping beam and to recollimate it to provide information on the horizontal position of the microsphere. The vertical degree of freedom is readout by forming an interferometer between the light retroreflected by the microsphere and a reference beam, hence eliminating the need for auxiliary beams. The focus of the trapping beam has a 1/e$^2$ radius of 3.2~$\mu$m and small non-Gaussian tails, suitable for bringing devices close to the trapped microsphere without disturbing the optical field. Electrodes surrounding the trapping region provide excellent control of the electric field, which can be used to drive the translational degrees of freedom of a charged microsphere and the rotational degrees of freedom of a neutral microsphere, coupling to its electric dipole moment. With this control, the charge state can be determined with single electron precision, the mass of individual microspheres can be measured, and empirical calibrations of the force sensitivity can be made for each microsphere. A force noise of $<1\times10^{-17}$~N/$\sqrt{\rm Hz}$, which is comparable to previous reports, is measured on all three degrees of freedom for 4.7~$\mu$m diameter, 84~pg silica microspheres. Various devices have been brought within $1.6~\mu$m of the surface of a trapped microsphere. Metrology in the trapping region is provided by two custom-designed microscopes providing views in the horizontal and one of the vertical planes. The apparatus opens the way to performing high sensitivity three-dimensional force measurements at short distance., Comment: 11 pages, 10 figures

This paper analyzes the robustness of deep learning models in autonomous driving applications and discusses the practical solutions to address that.

We investigate the detection prospects for two-neutrino and neutrinoless second order weak decays of $^{124}$Xe -- double electron capture ($0/2\nu\text{ECEC}$), electron capture with positron emission ($0/2\nu\text{EC}\beta^+$) and double-positron emission ($0/2\nu\beta^+\beta^+$) -- in multi-tonne xenon time projection chambers. We simulate the decays in a liquid xenon medium and develop a reconstruction algorithm which uses the multi-particle coincidence in these decays to separate signal from background. This is used to compute the expected detection efficiencies as a function of position resolution and energy threshold for planned experiments. In addition, we consider an exhaustive list of possible background sources and find that they are either negligible in rate or can be greatly reduced using our topological reconstruction criteria. In particular, we draw two conclusions: First, with a half-life of $T_{1/2}^{2\nu\text{EC}\beta^+} = (1.7 \pm 0.6)\cdot 10^{23}\,\text{yr}$, the $2\nu\text{EC}\beta^+$ decay of $^{124}$Xe will likely be detected in upcoming Dark Matter experiments (e.g. XENONnT or LZ), and their major background will be from gamma rays in the detector construction materials. Second, searches for the $0\nu\text{EC}\beta^+$ decay mode are likely to be background-free, and new parameter space may be within the reach. To this end we investigate three different scenarios of existing experimental constraints on the effective neutrino mass. The necessary 500 kg-year exposure of $^{124}$Xe could be achieved by the baseline design of the DARWIN observatory, or by extracting and using the $^{124}$Xe from the tailings of the nEXO experiment. We demonstrate how a combination of $^{124}$Xe results with those from $0\nu\beta^-\beta^-$ searches in $^{136}$Xe could help to identify the neutrinoless decay mechanism., Comment: 17 pages, 7 figures

Deep convolutional neural networks can be highly vulnerable to small perturbations of their inputs, potentially a major issue or limitation on system robustness when using deep networks as classifiers. In this paper we propose a low-cost method to explore marginal sample data near trained classifier decision boundaries, thus identifying potential adversarial samples. By finding such adversarial samples it is possible to reduce the search space of adversarial attack algorithms while keeping a reasonable successful perturbation rate. In our developed strategy, the potential adversarial samples represent only 61% of the test data, but in fact cover more than 82% of the adversarial samples produced by iFGSM and 92% of the adversarial samples successfully perturbed by DeepFool on CIFAR10.