Search

Adapting vision transformer foundation models through parameter-efficient fine-tuning (PEFT) methods has become increasingly popular. These methods optimize a limited subset of parameters, enabling efficient adaptation without the need to fine-tune the entire model while still achieving competitive performance. However, traditional PEFT methods may limit the model's capacity to capture complex patterns, especially those associated with high-frequency spectra. This limitation becomes particularly problematic as existing research indicates that high-frequency features are crucial for distinguishing subtle image structures. To address this issue, we introduce FreqFit, a novel Frequency Fine-tuning module between ViT blocks to enhance model adaptability. FreqFit is simple yet surprisingly effective, and can be integrated with all existing PEFT methods to boost their performance. By manipulating features in the frequency domain, our approach allows models to capture subtle patterns more effectively. Extensive experiments on 24 datasets, using both supervised and self-supervised foundational models with various state-of-the-art PEFT methods, reveal that FreqFit consistently improves performance over the original PEFT methods with performance gains ranging from 1% to 16%. For instance, FreqFit-LoRA surpasses the performances of state-of-the-art baselines on CIFAR100 by more than 10% even without applying regularization or strong augmentation. For reproducibility purposes, the source code is available at https://github.com/tsly123/FreqFiT., Comment: 24 pages

Developing an interpretable system for generating reports in chest X-ray (CXR) analysis is becoming increasingly crucial in Computer-aided Diagnosis (CAD) systems, enabling radiologists to comprehend the decisions made by these systems. Despite the growth of diverse datasets and methods focusing on report generation, there remains a notable gap in how closely these models' generated reports align with the interpretations of real radiologists. In this study, we tackle this challenge by initially introducing Fine-Grained CXR (FG-CXR) dataset, which provides fine-grained paired information between the captions generated by radiologists and the corresponding gaze attention heatmaps for each anatomy. Unlike existing datasets that include a raw sequence of gaze alongside a report, with significant misalignment between gaze location and report content, our FG-CXR dataset offers a more grained alignment between gaze attention and diagnosis transcript. Furthermore, our analysis reveals that simply applying black-box image captioning methods to generate reports cannot adequately explain which information in CXR is utilized and how long needs to attend to accurately generate reports. Consequently, we propose a novel explainable radiologist's attention generator network (Gen-XAI) that mimics the diagnosis process of radiologists, explicitly constraining its output to closely align with both radiologist's gaze attention and transcript. Finally, we perform extensive experiments to illustrate the effectiveness of our method. Our datasets and checkpoint is available at https://github.com/UARK-AICV/FG-CXR., Comment: ACCV 2024

Medical eye-tracking data is an important information source for understanding how radiologists visually interpret medical images. This information not only improves the accuracy of deep learning models for X-ray analysis but also their interpretability, enhancing transparency in decision-making. However, the current eye-tracking data is dispersed, unprocessed, and ambiguous, making it difficult to derive meaningful insights. Therefore, there is a need to create a new dataset with more focus and purposeful eyetracking data, improving its utility for diagnostic applications. In this work, we propose a refinement method inspired by the target-present visual search challenge: there is a specific finding and fixations are guided to locate it. After refining the existing eye-tracking datasets, we transform them into a curated visual search dataset, called GazeSearch, specifically for radiology findings, where each fixation sequence is purposefully aligned to the task of locating a particular finding. Subsequently, we introduce a scan path prediction baseline, called ChestSearch, specifically tailored to GazeSearch. Finally, we employ the newly introduced GazeSearch as a benchmark to evaluate the performance of current state-of-the-art methods, offering a comprehensive assessment for visual search in the medical imaging domain. Code is available at \url{https://github.com/UARK-AICV/GazeSearch}., Comment: Aceepted WACV 2025

The Probe far-Infrared Mission for Astrophysics (PRIMA) is a proposed space observatory which will use arrays of thousands of kinetic inductance detectors (KIDs) to perform low- and moderate-resolution spectroscopy throughout the far-infrared. The detectors must have noise equivalent powers (NEPs) at or below 0.1 aW/sqrt(Hz) to be subdominant to noise from sky backgrounds and thermal noise from PRIMA's cryogenically cooled primary mirror. Using a Radio Frequency System on a Chip for multitone readout, we measure the NEPs of detectors on a flight-like array designed to observe at a wavelength of 210 microns. We find that 92% of the KIDs measured have an NEP below 0.1 aW/sqrt(Hz) at a noise frequency of 10 Hz., Comment: 7 pages, 7 figures

Predicting human gaze behavior within computer vision is integral for developing interactive systems that can anticipate user attention, address fundamental questions in cognitive science, and hold implications for fields like human-computer interaction (HCI) and augmented/virtual reality (AR/VR) systems. Despite methodologies introduced for modeling human eye gaze behavior, applying these models to medical imaging for scanpath prediction remains unexplored. Our proposed system aims to predict eye gaze sequences from radiology reports and CXR images, potentially streamlining data collection and enhancing AI systems using larger datasets. However, predicting human scanpaths on medical images presents unique challenges due to the diverse nature of abnormal regions. Our model predicts fixation coordinates and durations critical for medical scanpath prediction, outperforming existing models in the computer vision community. Utilizing a two-stage training process and large publicly available datasets, our approach generates static heatmaps and eye gaze videos aligned with radiology reports, facilitating comprehensive analysis. We validate our approach by comparing its performance with state-of-the-art methods and assessing its generalizability among different radiologists, introducing novel strategies to model radiologists' search patterns during CXR image diagnosis. Based on the radiologist's evaluation, MedGaze can generate human-like gaze sequences with a high focus on relevant regions over the CXR images. It sometimes also outperforms humans in terms of redundancy and randomness in the scanpaths., Comment: Submitted to the Journal

Human-AI collaboration to identify and correct perceptual errors in chest radiographs has not been previously explored. This study aimed to develop a collaborative AI system, CoRaX, which integrates eye gaze data and radiology reports to enhance diagnostic accuracy in chest radiology by pinpointing perceptual errors and refining the decision-making process. Using public datasets REFLACX and EGD-CXR, the study retrospectively developed CoRaX, employing a large multimodal model to analyze image embeddings, eye gaze data, and radiology reports. The system's effectiveness was evaluated based on its referral-making process, the quality of referrals, and performance in collaborative diagnostic settings. CoRaX was tested on a simulated error dataset of 271 samples with 28% (93 of 332) missed abnormalities. The system corrected 21% (71 of 332) of these errors, leaving 7% (22 of 312) unresolved. The Referral-Usefulness score, indicating the accuracy of predicted regions for all true referrals, was 0.63 (95% CI 0.59, 0.68). The Total-Usefulness score, reflecting the diagnostic accuracy of CoRaX's interactions with radiologists, showed that 84% (237 of 280) of these interactions had a score above 0.40. In conclusion, CoRaX efficiently collaborates with radiologists to address perceptual errors across various abnormalities, with potential applications in the education and training of novice radiologists., Comment: Under Review in Journal

The search for the polarized imprint of primordial gravitational waves in the cosmic microwave background (CMB) as direct evidence of cosmic inflation requires exquisite sensitivity and control over systematics. The next-generation CMB-S4 project intends to improve upon current-generation experiments by deploying a significantly greater number of highly-sensitive detectors, combined with refined instrument components based on designs from field-proven instruments. The Precursor Small Aperture Telescope (PreSAT) is envisioned as an early step to this next generation, which will test prototype CMB-S4 components and technologies within an existing BICEP Array receiver, with the aim of enabling full-stack laboratory testing and early risk retirement, along with direct correlation of laboratory component-level performance measurements with deployed system performance. The instrument will utilize new 95/155GHz dichroic dual-linear-polarization prototype detectors developed for CMB-S4, cooled to 100mK via the installation of an adiabatic demagnetization refrigerator, along with a prototype readout chain and prototype optics manufactured with wide-band anti-reflection coatings. The experience gained by integrating, deploying, and calibrating PreSAT will also help inform planning for CMB-S4 small aperture telescope commissioning, calibration, and operations well in advance of the fabrication of CMB-S4 production hardware., Comment: 12 pages, 4 figures, submitted to Proc. SPIE

In the realm of chest X-ray (CXR) image analysis, radiologists meticulously examine various regions, documenting their observations in reports. The prevalence of errors in CXR diagnoses, particularly among inexperienced radiologists and hospital residents, underscores the importance of understanding radiologists' intentions and the corresponding regions of interest. This understanding is crucial for correcting mistakes by guiding radiologists to the accurate regions of interest, especially in the diagnosis of chest radiograph abnormalities. In response to this imperative, we propose a novel system designed to identify the primary intentions articulated by radiologists in their reports and the corresponding regions of interest in CXR images. This system seeks to elucidate the visual context underlying radiologists' textual findings, with the potential to rectify errors made by less experienced practitioners and direct them to precise regions of interest. Importantly, the proposed system can be instrumental in providing constructive feedback to inexperienced radiologists or junior residents in the hospital, bridging the gap in face-to-face communication. The system represents a valuable tool for enhancing diagnostic accuracy and fostering continuous learning within the medical community., Comment: Accepted in ISBI 2024

Many problems in statistics and machine learning can be formulated as model selection problems, where the goal is to choose an optimal parsimonious model among a set of candidate models. It is typical to conduct model selection by penalizing the objective function via information criteria (IC), as with the pioneering work by Akaike and Schwarz. Via recent work, we propose a generalized IC framework to consistently estimate general loss-based learning problems. In this work, we propose a consistent estimation method for Generalized Linear Model (GLM) regressions by utilizing the recent IC developments. We advance the generalized IC framework by proposing model selection problems, where the model set consists of a potentially uncountable set of models. In addition to theoretical expositions, our proposal introduces a computational procedure for the implementation of our methods in the finite sample setting, which we demonstrate via an extensive simulation study.

We consider the problem of estimating probability density functions based on sample data, using a finite mixture of densities from some component class. To this end, we introduce the $h$-lifted Kullback--Leibler (KL) divergence as a generalization of the standard KL divergence and a criterion for conducting risk minimization. Under a compact support assumption, we prove an $\mathcal{O}(1/{\sqrt{n}})$ bound on the expected estimation error when using the $h$-lifted KL divergence, which extends the results of Rakhlin et al. (2005, ESAIM: Probability and Statistics, Vol. 9) and Li and Barron (1999, Advances in Neural Information ProcessingSystems, Vol. 12) to permit the risk bounding of density functions that are not strictly positive. We develop a procedure for the computation of the corresponding maximum $h$-lifted likelihood estimators ($h$-MLLEs) using the Majorization-Maximization framework and provide experimental results in support of our theoretical bounds.

SPHEREx, the Spectro-Photometer for the History of the Universe, Epoch of Reionization, and ices Explorer, is a NASA MIDEX mission planned for launch in 2024. SPHEREx will carry out the first all-sky spectral survey at wavelengths between 0.75 micron and 5 micron with spectral resolving power ~40 between 0.75 and 3.8 micron and ~120 between 3.8 and 5 micron At the end of its two-year mission, SPHEREx will provide 0.75-to-5 micron spectra of each 6.2"x6.2" pixel on the sky - 14 billion spectra in all. This paper updates an earlier description of SPHEREx presenting changes made during the mission's Preliminary Design Phase, including a discussion of instrument integration and test and a summary of the data processing, analysis, and distribution plans.

The PRobe far-Infrared Mission for Astrophysics (PRIMA) is working to develop kinetic inductance detectors (KIDs) that can meet the sensitivity targets of a far-infrared spectrometer on a cryogenically cooled space telescope. An important ingredient for achieving high sensitivity is increasing the fractional-frequency responsivity. Here we present a study of the responsivity of aluminum KIDs fabricated at the Jet Propulsion Laboratory. Specifically, we model the KID's temporal response to pair-breaking excitations in the framework of the Mattis-Bardeen theory, incorporating quasiparticle recombination dynamics and the pair-breaking efficiency. Using a near-IR laser, we measure time-resolved photon pulses and fit them to our model, extracting the time-resolved quasiparticle density and the quasiparticle recombination lifetime. Comparing the fit to the known energy of the laser provides a measurement of the pair-breaking efficiency. In addition to photon-sourced excitations, it is important to understand the KID's response to phonon-sourced excitations from cosmic rays. We measure the rate of secondary cosmic rays detected by our devices, and predict the dead time due to cosmic rays for an array in L2 orbit. This work provides confidence in KIDs' robustness to cosmic ray events in the space environment., Comment: 8 pages, 4 figures

Far-infrared (far-IR) astrophysics missions featuring actively cooled telescopes will offer orders of magnitude observing speed improvement at wavelengths where galaxies and forming planetary systems emit most of their light. The PRobe far-Infrared Mission for Astrophysics (PRIMA), which is currently under study, emphasizes low and moderate resolution spectroscopy throughout the far-IR. Full utilization of PRIMA's cold telescope requires far-IR detector arrays with per-pixel noise equivalent powers (NEPs) at or below 1 x 10-19 W/rtHz. We are developing low-volume Aluminum kinetic inductance detector (KID) arrays to reach these sensitivities. We will present on the development of our long-wavelength (210 um) array approach, with a focus on multitone measurements of our 1,008-pixel arrays. We measure an NEP below 1 x 10-19 W/rtHz for 73 percent of our pixels., Comment: 9 pages, 5 figures, 20th International Workshop on Low Temperature Detectors, submitted to the Journal of Low Temperature Physics

Music-driven group choreography poses a considerable challenge but holds significant potential for a wide range of industrial applications. The ability to generate synchronized and visually appealing group dance motions that are aligned with music opens up opportunities in many fields such as entertainment, advertising, and virtual performances. However, most of the recent works are not able to generate high-fidelity long-term motions, or fail to enable controllable experience. In this work, we aim to address the demand for high-quality and customizable group dance generation by effectively governing the consistency and diversity of group choreographies. In particular, we utilize a diffusion-based generative approach to enable the synthesis of flexible number of dancers and long-term group dances, while ensuring coherence to the input music. Ultimately, we introduce a Group Contrastive Diffusion (GCD) strategy to enhance the connection between dancers and their group, presenting the ability to control the consistency or diversity level of the synthesized group animation via the classifier-guidance sampling technique. Through intensive experiments and evaluation, we demonstrate the effectiveness of our approach in producing visually captivating and consistent group dance motions. The experimental results show the capability of our method to achieve the desired levels of consistency and diversity, while maintaining the overall quality of the generated group choreography. The source code can be found at https://aioz-ai.github.io/GCD

In the field of chest X-ray (CXR) diagnosis, existing works often focus solely on determining where a radiologist looks, typically through tasks such as detection, segmentation, or classification. However, these approaches are often designed as black-box models, lacking interpretability. In this paper, we introduce Interpretable Artificial Intelligence (I-AI) a novel and unified controllable interpretable pipeline for decoding the intense focus of radiologists in CXR diagnosis. Our I-AI addresses three key questions: where a radiologist looks, how long they focus on specific areas, and what findings they diagnose. By capturing the intensity of the radiologist's gaze, we provide a unified solution that offers insights into the cognitive process underlying radiological interpretation. Unlike current methods that rely on black-box machine learning models, which can be prone to extracting erroneous information from the entire input image during the diagnosis process, we tackle this issue by effectively masking out irrelevant information. Our proposed I-AI leverages a vision-language model, allowing for precise control over the interpretation process while ensuring the exclusion of irrelevant features. To train our I-AI model, we utilize an eye gaze dataset to extract anatomical gaze information and generate ground truth heatmaps. Through extensive experimentation, we demonstrate the efficacy of our method. We showcase that the attention heatmaps, designed to mimic radiologists' focus, encode sufficient and relevant information, enabling accurate classification tasks using only a portion of CXR. The code, checkpoints, and data are at https://github.com/UARK-AICV/IAI, Comment: Accepted at WACV 2024

We report on efforts to improve the speed of low-G far-infrared transition-edged-sensed bolometers. We use a fabrication process that does not require any dry etch steps to reduce heat capacity on the suspended device and measure a reduction in the detector time constant. However, we also measure an increase in the temperature-normalized thermal conductance (G), and a corresponding increase in the noise-equivalent power (NEP). We employ a new near-IR photon-noise technique using a near-IR laser to calibrate the frequency-domain multiplexed AC system and compare the results to a well-understood DC circuit. We measure an NEP white noise level of 0.8 aW/rtHz with a 1/f knee below 0.1 Hz and a time constant of 3.2 ms., Comment: 27 pages, 16 figures. This article may be downloaded for personal use only. Any other use requires prior permission of the author and AIP Publishing. This article appeared in J. Appl. Phys. 134 (9) and may be found at https://doi.org/10.1063/5.0157208

The BReast CAncer type 2 susceptibility protein (BRCA2) responds to DNA damage by participating in homology-directed repair. BRCA2 deficiency culminates in defective DNA damage repair (DDR) that when prolonged leads to the accumulation of DNA damage causing cancer or apoptosis. Oxidative stress promotes DNA damage and apoptosis and is a common mechanism through which cardiovascular risk factors lead to endothelial dysfunction and atherosclerosis. Herein, we show that endothelial BRCA2 plays a protective role against atherosclerosis under hypercholesterolemic stress. We successfully generated and characterized endothelial cell (EC)-specific BRCA2 knockout (BRCA2endo) mice. To study the effect of EC-specific BRCA2-loss in atherosclerosis, we generated and characterized BRCA2endo mice on apolipoprotein E null background (ApoE-/-), fed them with high-fat diet (HFD) and evaluated atherosclerosis. Baseline phenotyping of BRCA2endo mice did not show any adverse effects in terms of DNA damage and apoptosis as well as cardiac and metabolic function. However, using HFD-fed apolipoprotein E knockout (ApoE-/-) background, we demonstrated that EC-specific loss of BRCA2 resulted in aortic plaque deposition and splenomegaly. Comparison of RNA sequencing data from aortas of EC-specific BRCA2-deficient ApoE-/- and BRCA2-intact ApoE-/- mice revealed a total of 530 significantly differentially expressed genes with Protein Folding Response and Lipid Metabolism as the most affected pathways. This study provides foundational knowledge regarding BRCA2 status and function in the cardiovascular system, and highlights the potential of BRCA2 as a novel therapeutic target in prevention and treatment of atherosclerosis. Our data indicate that BRCA2 mutation carriers may be at a previously unrecognized risk of atherosclerosis in addition to breast and ovarian cancer.

The definition of anomaly detection is the identification of an unexpected event. Real-time detection of extreme events such as wildfires, cyclones, or floods using satellite data has become crucial for disaster management. Although several earth-observing satellites provide information about disasters, satellites in the geostationary orbit provide data at intervals as frequent as every minute, effectively creating a video from space. There are many techniques that have been proposed to identify anomalies in surveillance videos; however, the available datasets do not have dynamic behavior, so we discuss an anomaly framework that can work on very high-frequency datasets to find very fast-moving anomalies. In this work, we present a diffusion model which does not need any motion component to capture the fast-moving anomalies and outperforms the other baseline methods.

Wildlife camera trap images are being used extensively to investigate animal abundance, habitat associations, and behavior, which is complicated by the fact that experts must first classify the images manually. Artificial intelligence systems can take over this task but usually need a large number of already-labeled training images to achieve sufficient performance. This requirement necessitates human expert labor and poses a particular challenge for projects with few cameras or short durations. We propose a label-efficient learning strategy that enables researchers with small or medium-sized image databases to leverage the potential of modern machine learning, thus freeing crucial resources for subsequent analyses. Our methodological proposal is two-fold: (1) We improve current strategies of combining object detection and image classification by tuning the hyperparameters of both models. (2) We provide an active learning (AL) system that allows training deep learning models very efficiently in terms of required human-labeled training images. We supply a software package that enables researchers to use these methods directly and thereby ensure the broad applicability of the proposed framework in ecological practice. We show that our tuning strategy improves predictive performance. We demonstrate how the AL pipeline reduces the amount of pre-labeled data needed to achieve a specific predictive performance and that it is especially valuable for improving out-of-sample predictive performance. We conclude that the combination of tuning and AL increases predictive performance substantially. Furthermore, we argue that our work can broadly impact the community through the ready-to-use software package provided. Finally, the publication of our models tailored to European wildlife data enriches existing model bases mostly trained on data from Africa and North America.

Federated learning has been widely applied in autonomous driving since it enables training a learning model among vehicles without sharing users' data. However, data from autonomous vehicles usually suffer from the non-independent-and-identically-distributed (non-IID) problem, which may cause negative effects on the convergence of the learning process. In this paper, we propose a new contrastive divergence loss to address the non-IID problem in autonomous driving by reducing the impact of divergence factors from transmitted models during the local learning process of each silo. We also analyze the effects of contrastive divergence in various autonomous driving scenarios, under multiple network infrastructures, and with different centralized/distributed learning schemes. Our intensive experiments on three datasets demonstrate that our proposed contrastive divergence loss significantly improves the performance over current state-of-the-art approaches.

Model selection is a ubiquitous problem that arises in the application of many statistical and machine learning methods. In the likelihood and related settings, it is typical to use the method of information criteria (IC) to choose the most parsimonious among competing models by penalizing the likelihood-based objective function. Theorems guaranteeing the consistency of IC can often be difficult to verify and are often specific and bespoke. We present a set of results that guarantee consistency for a class of IC, which we call PanIC (from the Greek root 'pan', meaning 'of everything'), with easily verifiable regularity conditions. The PanIC are applicable in any loss-based learning problem and are not exclusive to likelihood problems. We illustrate the verification of regularity conditions for model selection problems regarding finite mixture models, least absolute deviation and support vector regression, and principal component analysis, and we demonstrate the effectiveness of the PanIC for such problems via numerical simulations. Furthermore, we present new sufficient conditions for the consistency of BIC-like estimators and provide comparisons of the BIC to PanIC.

In recent years, empirical Bayesian (EB) inference has become an attractive approach for estimation in parametric models arising in a variety of real-life problems, especially in complex and high-dimensional scientific applications. However, compared to the relative abundance of available general methods for computing point estimators in the EB framework, the construction of confidence sets and hypothesis tests with good theoretical properties remains difficult and problem specific. Motivated by the universal inference framework of Wasserman et al. (2020), we propose a general and universal method, based on holdout likelihood ratios, and utilizing the hierarchical structure of the specified Bayesian model for constructing confidence sets and hypothesis tests that are finite sample valid. We illustrate our method through a range of numerical studies and real data applications, which demonstrate that the approach is able to generate useful and meaningful inferential statements in the relevant contexts.

In this paper, an ontology-based approach is used to organize the knowledge base of legal documents in road traffic law. This knowledge model is built by the improvement of ontology Rela-model. In addition, several searching problems on traffic law are proposed and solved based on the legal knowledge base. The intelligent search system on Vietnam road traffic law is constructed by applying the method. The searching system can help users to find concepts and definitions in road traffic law. Moreover, it can also determine penalties and fines for violations in the traffic. The experiment results show that the system is efficient for users' typical searching and is emerging for usage in the real-world.

Odd-even harmonics emitted from a laser-target system imprint rich, subtle information characterizing the system's dynamical asymmetry, which is desirable to decipher. In this Letter, we discover a simple universal relation between the odd-even harmonics and the asymmetry of the THz-assisted laser-atomic system -- atoms in a fundamental mid-IR laser pulse combined with a THz laser. First, we demonstrate numerically and then analytically formulize the harmonic even-to-odd ratio as a function of the THz electric field, the source of the system's asymmetry. Notably, we suggest a scaling that makes the obtained rule universal, independent of the parameters of both the fundamental pulse and atomic target. This universality facilitates us to propose a general pump-probe scheme for THz waveform sampling from the even-to-odd ratio, measurable within a conventional compact setup.