Search

The goal of this project is to construct an estimator for the masses of supermassive black holes in active galactic nuclei (AGNs) based on the broad Halpha emission line. We make use of published reverberation mapping data. We remeasure all Halpha time lags from the original data as we find that often the reverberation measurements are improved by detrending the light curves. We produce mass estimators that require only the Halpha luminosity and the width of the Halpha emission line as characterized by either the FWHM or the line dispersion. It is possible, on the basis of a single spectrum covering the Halpha emission line, to estimate the mass of the central supermassive black hole in AGNs, taking into account all three parameters believed to affect mass measurement: luminosity, line width, and Eddington ratio. The typical formal accuracy in such estimates is of order 0.2-0.3 dex., Comment: 15 pages, 9 figures, submitted to A&A

Pairwise interactions between perturbations to a system can provide evidence for the causal dependencies of the underlying underlying mechanisms of a system. When observations are low dimensional, hand crafted measurements, detecting interactions amounts to simple statistical tests, but it is not obvious how to detect interactions between perturbations affecting latent variables. We derive two interaction tests that are based on pairwise interventions, and show how these tests can be integrated into an active learning pipeline to efficiently discover pairwise interactions between perturbations. We illustrate the value of these tests in the context of biology, where pairwise perturbation experiments are frequently used to reveal interactions that are not observable from any single perturbation. Our tests can be run on unstructured data, such as the pixels in an image, which enables a more general notion of interaction than typical cell viability experiments, and can be run on cheaper experimental assays. We validate on several synthetic and real biological experiments that our tests are able to identify interacting pairs effectively. We evaluate our approach on a real biological experiment where we knocked out 50 pairs of genes and measured the effect with microscopy images. We show that we are able to recover significantly more known biological interactions than random search and standard active learning baselines.

Foundation models have emerged as robust models with label efficiency in diverse domains. In medical imaging, these models contribute to the advancement of medical diagnoses due to the difficulty in obtaining labeled data. However, it is unclear whether using a large amount of unlabeled data, biased by the presence of sensitive attributes during pre-training, influences the fairness of the model. This research examines the bias in the Foundation model (RetFound) when it is applied to fine-tune the Brazilian Multilabel Ophthalmological Dataset (BRSET), which has a different population than the pre-training dataset. The model evaluation, in comparison with supervised learning, shows that the Foundation Model has the potential to reduce the gap between the maximum AUC and minimum AUC evaluations across gender and age groups. However, in a data-efficient generalization, the model increases the bias when the data amount decreases. These findings suggest that when deploying a Foundation Model in real-life scenarios with limited data, the possibility of fairness issues should be considered., Comment: Preprint of paper to be presented at Fairness and Ethics Towards Transparent AI: Facing the Challenge through Model Debiasing (FAILED) during ECCV 2024

Ensuring consistent performance across diverse populations and incorporating fairness into machine learning models are crucial for advancing medical image diagnostics and promoting equitable healthcare. However, many databases do not provide protected attributes or contain unbalanced representations of demographic groups, complicating the evaluation of model performance across different demographics and the application of bias mitigation techniques that rely on these attributes. This study aims to investigate the effectiveness of using the backbone of Foundation Models as an embedding extractor for creating groups that represent protected attributes, such as gender and age. We propose utilizing these groups in different stages of bias mitigation, including pre-processing, in-processing, and evaluation. Using databases in and out-of-distribution scenarios, it is possible to identify that the method can create groups that represent gender in both databases and reduce in 4.44% the difference between the gender attribute in-distribution and 6.16% in out-of-distribution. However, the model lacks robustness in handling age attributes, underscoring the need for more fundamentally fair and robust Foundation models. These findings suggest a role in promoting fairness assessment in scenarios where we lack knowledge of attributes, contributing to the development of more equitable medical diagnostics., Comment: Preprint of paper to be presented at Fairness of AI in Medical Imaging (FAIMI) during MICCAI 2024

Mesh-based scene representation offers a promising direction for simplifying large-scale hierarchical visual localization pipelines, combining a visual place recognition step based on global features (retrieval) and a visual localization step based on local features. While existing work demonstrates the viability of meshes for visual localization, the impact of using synthetic databases rendered from them in visual place recognition remains largely unexplored. In this work we investigate using dense 3D textured meshes for large-scale Visual Place Recognition (VPR). We identify a significant performance drop when using synthetic mesh-based image databases compared to real-world images for retrieval. To address this, we propose MeshVPR, a novel VPR pipeline that utilizes a lightweight features alignment framework to bridge the gap between real-world and synthetic domains. MeshVPR leverages pre-trained VPR models and is efficient and scalable for city-wide deployments. We introduce novel datasets with freely available 3D meshes and manually collected queries from Berlin, Paris, and Melbourne. Extensive evaluations demonstrate that MeshVPR achieves competitive performance with standard VPR pipelines, paving the way for mesh-based localization systems. Data, code, and interactive visualizations are available at https://meshvpr.github.io/, Comment: Website: https://mesh-vpr.github.io/

In the past five years, six quasi-periodic X-ray eruption (QPE) sources have been discovered in the nuclei of nearby galaxies. Their origin remains an open question. We present MUSE integral field spectroscopy of five QPE host galaxies to characterize their properties. We find that 3/5 galaxies host extended emission line regions (EELRs) up to 10 kpc in size. The EELRs are photo-ionized by a non-stellar continuum, but the current nuclear luminosity is insufficient to power the observed emission lines. The EELRs are decoupled from the stars both kinematically and in projected sky position, and the low velocities and velocity dispersions ($<$ 100 km s$^{-1}$ and $\lesssim 75$ km s$^{-1}$ respectively) are inconsistent with being AGN- or shock-driven. The origin of the EELRs is likely a previous phase of nuclear activity. The QPE host galaxy properties are strikingly similar to those of tidal disruption events (Wevers et al. submitted). The preference for a very short-lived (the typical EELR lifetime is $\sim$15000 years), gas-rich phase where the nucleus has recently faded significantly suggests that TDEs and QPEs may share a common formation channel, disfavoring AGN accretion disk instabilities as the origin of QPEs. In the assumption that QPEs are related to extreme mass ratio inspiral systems (EMRIs; stellar-mass objects on bound orbits about massive black holes), the high incidence of EELRs and recently faded nuclear activity can be used to aid in the localization of the host galaxies of EMRIs discovered by low frequency gravitational wave observatories., Comment: Submitted to ApJ Letters. In a companion paper we present an IFU analysis of TDE host galaxies (Wevers & French)

In computer vision, keypoint detection is a fundamental task, with applications spanning from robotics to image retrieval; however, existing learning-based methods suffer from scale dependency and lack flexibility. This paper introduces a novel approach that leverages Morse theory and persistent homology, powerful tools rooted in algebraic topology. We propose a novel loss function based on the recent introduction of a notion of subgradient in persistent homology, paving the way toward topological learning. Our detector, MorseDet, is the first topology-based learning model for feature detection, which achieves competitive performance in keypoint repeatability and introduces a principled and theoretically robust approach to the problem., Comment: Accepted to ICML 2024

The term 'active galactic nuclei' (AGN) subtends a huge variety of objects, classified on their properties at different wavelengths. Peaked sources (PS) represent a class of AGN at the first stage of evolution, characterised by a peaked radio spectrum. Among these radio sources, low-luminosity compact (LLC) sources can be identified as PS accreting with a high Eddington rate, harbouring low-power jets and hosting low-mass black holes. These properties are also shared by narrow-line Seyfert 1 galaxies (NLS1s). In 2016, LLCs were hypothesised to be the parent population of NLS1s with a flat radio spectrum (F-NLS1s), suggesting the former to be the same objects as the latter, seen at higher inclination. Based on radio luminosity functions and optical spectra analysis, 10 LLCs were identified as valid candidates for F-NLS1s. To account for the missing puzzle piece, verifying if these LLCs could be hosted in late-type galaxies as NLS1s, we performed the photometric decomposition of their Pan-STARRS1 images in all five filters. We used the 2D fitting algorithm GALFIT for the single-band analysis, and its extension GALFITM for the multi-band analysis. Considering that the morphological type and the structural parameters of the host can be dependent on the wavelength, we found six out of ten LLCs hosted in late-type galaxies, probably with pseudo-bulges, three point-like sources and one object of uncertain classification. Although this study is based on a small sample, it represents the first morphological analysis of LLC host galaxies. The results confirm the trend observed in NLS1s, indicating late-type/disc-like host galaxies for LLCs and supporting the validity of the parent population scenario., Comment: 15 pages, 21 figures, accepted for publication in A&A

Precise, pixel-wise geolocalization of astronaut photography is critical to unlocking the potential of this unique type of remotely sensed Earth data, particularly for its use in disaster management and climate change research. Recent works have established the Astronaut Photography Localization task, but have either proved too costly for mass deployment or generated too coarse a localization. Thus, we present EarthMatch, an iterative homography estimation method that produces fine-grained localization of astronaut photographs while maintaining an emphasis on speed. We refocus the astronaut photography benchmark, AIMS, on the geolocalization task itself, and prove our method's efficacy on this dataset. In addition, we offer a new, fair method for image matcher comparison, and an extensive evaluation of different matching models within our localization pipeline. Our method will enable fast and accurate localization of the 4.5 million and growing collection of astronaut photography of Earth. Webpage with code and data at https://earthloc-and-earthmatch.github.io, Comment: CVPR 2024 IMW - webpage: https://earthloc-and-earthmatch.github.io

Visual Place Recognition (VPR) aims to estimate the location of an image by treating it as a retrieval problem. VPR uses a database of geo-tagged images and leverages deep neural networks to extract a global representation, called descriptor, from each image. While the training data for VPR models often originates from diverse, geographically scattered sources (geo-tagged images), the training process itself is typically assumed to be centralized. This research revisits the task of VPR through the lens of Federated Learning (FL), addressing several key challenges associated with this adaptation. VPR data inherently lacks well-defined classes, and models are typically trained using contrastive learning, which necessitates a data mining step on a centralized database. Additionally, client devices in federated systems can be highly heterogeneous in terms of their processing capabilities. The proposed FedVPR framework not only presents a novel approach for VPR but also introduces a new, challenging, and realistic task for FL research, paving the way to other image retrieval tasks in FL., Comment: 13 pages, 7 figures, CVPR - The 3rd International Workshop on Federated Learning for Computer Vision (FedVision-2024)

Featurizing microscopy images for use in biological research remains a significant challenge, especially for large-scale experiments spanning millions of images. This work explores the scaling properties of weakly supervised classifiers and self-supervised masked autoencoders (MAEs) when training with increasingly larger model backbones and microscopy datasets. Our results show that ViT-based MAEs outperform weakly supervised classifiers on a variety of tasks, achieving as much as a 11.5% relative improvement when recalling known biological relationships curated from public databases. Additionally, we develop a new channel-agnostic MAE architecture (CA-MAE) that allows for inputting images of different numbers and orders of channels at inference time. We demonstrate that CA-MAEs effectively generalize by inferring and evaluating on a microscopy image dataset (JUMP-CP) generated under different experimental conditions with a different channel structure than our pretraining data (RPI-93M). Our findings motivate continued research into scaling self-supervised learning on microscopy data in order to create powerful foundation models of cellular biology that have the potential to catalyze advancements in drug discovery and beyond., Comment: CVPR 2024 Highlight. arXiv admin note: text overlap with arXiv:2309.16064

Visual Place Recognition aims at recognizing previously visited places by relying on visual clues, and it is used in robotics applications for SLAM and localization. Since typically a mobile robot has access to a continuous stream of frames, this task is naturally cast as a sequence-to-sequence localization problem. Nevertheless, obtaining sequences of labelled data is much more expensive than collecting isolated images, which can be done in an automated way with little supervision. As a mitigation to this problem, we propose a novel Joint Image and Sequence Training protocol (JIST) that leverages large uncurated sets of images through a multi-task learning framework. With JIST we also introduce SeqGeM, an aggregation layer that revisits the popular GeM pooling to produce a single robust and compact embedding from a sequence of single-frame embeddings. We show that our model is able to outperform previous state of the art while being faster, using 8 times smaller descriptors, having a lighter architecture and allowing to process sequences of various lengths. Code is available at https://github.com/ga1i13o/JIST

Astronaut photography, spanning six decades of human spaceflight, presents a unique Earth observations dataset with immense value for both scientific research and disaster response. Despite its significance, accurately localizing the geographical extent of these images, crucial for effective utilization, poses substantial challenges. Current manual localization efforts are time-consuming, motivating the need for automated solutions. We propose a novel approach - leveraging image retrieval - to address this challenge efficiently. We introduce innovative training techniques, including Year-Wise Data Augmentation and a Neutral-Aware Multi-Similarity Loss, which contribute to the development of a high-performance model, EarthLoc. We develop six evaluation datasets and perform a comprehensive benchmark comparing EarthLoc to existing methods, showcasing its superior efficiency and accuracy. Our approach marks a significant advancement in automating the localization of astronaut photography, which will help bridge a critical gap in Earth observations data. Code and datasets are available at https://github.com/gmberton/EarthLoc, Comment: CVPR 2024

We present the X-ray spectral analysis of two complementary sets of intermediate Seyfert galaxies (ISs). Analyzing X-ray data, we estimate the hydrogen abundance $N_H$ and test its connection with the [O III] luminosity acquired from optical observations. The results confirm the conclusions drawn in a previous study concerning the lack of a direct correlation between the obscuration measure ($N_H$) and the intrinsic characteristics of the active nuclei ([O III] luminosity). Instead, we validate the existence of a correlation between the Seyfert type and the $N_H$ parameter, employing a separation threshold of approximately 10$^{22}$ atoms cm$^{-2}$. Simultaneously, our findings align with prior research, corroborating the relationship between X-ray luminosity and the [O III] luminosity.

The liver is the most involved organ by distant metastasis in colon-rectal cancer (CRC) patients and it comes necessary to be aware of the mutational status of the lesions to correctly design the best individual treatment. So far, efforts have been made in order to develop non-invasive and real-time methods that permit the analysis of the whole tumor, using new artificial intelligence tools to analyze the tumor's image obtained by Computed Tomography (CT) scan. In order to address the current medical workflow, that is biopsy analysis-based, we propose the first DeepLearning-based exploration, to our knowledge, of such classification approach from the patient medical imaging. We propose i) a solid pipeline for managing undersized datasets of available CT scans and ii) a baseline study for genomics mutation diagnosis support for preemptive patient follow-up. Our method is able to identify CRC RAS mutation family from CT images with 0.73 F1 score.

We consider a family of conditional nonlinear expectations defined on the space of bounded random variables and indexed by the class of all the sub-sigma-algebras of a given underlying sigma-algebra. We show that if this family satisfies a natural consistency property, then it collapses to a conditional certainty equivalent defined in terms of a state-dependent utility function. This result is obtained by embedding our problem in a decision theoretical framework and providing a new characterization of the Sure-Thing Principle. In particular we prove that this principle characterizes those preference relations which admit consistent backward conditional projections. We build our analysis on state-dependent preferences for a general state space as in Wakker and Zank (1999) and show that their numerical representation admits a continuous version of the state-dependent utility. In this way, we also answer positively to a conjecture posed in the aforementioned paper.

Variability studies of active galactic nuclei are a powerful diagnostic tool in understanding the physical processes occurring in disk-jet regions, unresolved by direct imaging with currently available techniques. Here, we report the first attempt to systematically characterize intra-night optical variability (INOV) for a sample of seven apparently radio-quiet narrow-line Seyfert 1 galaxies (RQNLSy1s) that had shown recurring flaring at 37 GHz in the radio observations at Metsahovi Radio Observatory (MRO), indicating the presence of relativistic jets in them, but no evidence for relativistic jets in the recent radio observations of Karl G. Jansky Very Large Array (JVLA) at 1.6, 5.2, and 9.0 GHz. We have conducted a total of 28 intra-night sessions, each lasting $\geq$ 3 hrs for this sample, resulting in an INOV duty cycle ($\overline{DC} ~\sim$20%) similar to that reported for $\gamma$-ray-NLSy1s (DC $\sim$25% - 30%), that display blazar-like INOV. This in turn infers the presence of relativistic jet in our sample sources. Thus, it appears that even lower-mass (M$_{BH} \sim$10$^{6}$ M$_{\odot}$) RQNLSy1 galaxies can maintain blazar-like activities. However, we note that the magnetic reconnection in the magnetosphere of the black hole can also be a viable mechanism to give rise to the INOV from these sources., Comment: 5 pages, 4 figures, accepted for publication in MNRAS Letter

We present new 1.5-8.5 GHz Very Long Baseline Array (VLBA) observations and 0.32-1.26 GHz Giant Meterwave Radio Telescope (GMRT) observations of J0354-1340, which is the only known radio-quiet (RQ) or radio-intermediate (RI) narrow-line Seyfert 1 galaxy with a 100-kpc two-sided radio jet. A pc-scale one-sided jet in the southeast direction from the core emission is found in the VLBA observations, while the kpc-scale jet observed with Karl G. Jansky Very Large Array (VLA) and GMRT is in the south-north direction. The core spectra on pc and kpc scales are presented in combination with the archival VLASS observations at 3.0 GHz and the VLA C configuration observations at 5.5 GHz. The pc-scale emission dominates the kpc-scale emission above ~ 5 GHz, and the spectrum is inverted due to synchrotron self-absorption. This indicates a compact synchrotron source with a size of ~ 0.04 pc, which is associated with either the jet base or the corona. A sub-kpc scale jet, which is unresolved on scales of ~ 3 arcsec, probably dominates the emission below ~ 5 GHz. Future radio observations can explore the jet structure between the pc and 100 kpc scales, the origin of their direction mismatch, and the pc-scale jet proper motion. It remains to be explored how common such large-scale jets are in RQ or RI AGN., Comment: Accepted for publication in ApJ

We report on the discovery of one of the most extreme cases of high-frequency radio variability ever measured in active galactic nuclei (AGN), observed on timescales of days and exhibiting variability amplitudes of three to four orders of magnitude. These sources, all radio-weak narrow-line Seyfert 1 (NLS1) galaxies, were discovered some years ago at Aalto University Mets\"ahovi Radio Observatory (MRO) based on recurring flaring at 37 GHz, strongly indicating the presence of relativistic jets. In subsequent observations with the Karl G. Jansky Very Large Array (JVLA) at 1.6, 5.2, and 9.0~GHz no signs of jets were seen. To determine the cause of their extraordinary behaviour, we observed them with the JVLA at 10, 15, 22, 33, and 45 GHz, and with the Very Long Baseline Array (VLBA) at 15 GHz. These observations were complemented with single-dish monitoring at 37 GHz at MRO, and at 15 GHz at Owens Valley Radio Observatory (OVRO). Intriguingly, all but one source either have a steep radio spectrum up to 45 GHz, or were not detected at all. Based on the 37 GHz data the timescales of the radio flares are a few days, and the derived variability brightness temperatures and variability Doppler factors comparable to those seen in blazars. We discuss alternative explanations for their extreme behaviour, but so far no definite conclusions can be made. These sources exhibit radio variability at a level rarely, if ever, seen in AGN. They might represent a new type of jetted AGN, or a new variability phenomenon, and thus deserve our continued attention., Comment: 31 pages, 37 figures, submitted to MNRAS

We present preliminary results of the analysis of optical spectra of two complementary samples of Seyfert galaxies. The first sample was extracted from a selection of the 4th Fermi Gamma-ray Large Area Telescope (4FGL) catalog, and consists of 9 $\gamma$-ray emitting jetted Seyfert galaxies. The second one was extracted from the Swift-BAT AGN Spectroscopic Survey (BASS), and is composed of 38 hard-X ray selected Active Galactic Nuclei (AGN). These two samples are complementary, with the former expected to have smaller viewing angles, while the latter may include objects with larger viewing angles. We measured emission line ratios to investigate whether the behavior of these Seyferts can be explained in terms of obscuration, as suggested by the well-known Unified Model (UM) of AGN, or if there are intrinsic differences due to the presence of jets, outflows, or the evolution. We found no indications of intrinsic differences. The UM remains the most plausible interpretation for these classes of objects even if some results can be challenging for this model., Comment: 24 pages, 10 figures, conference proceeding "14th Serbian Conference on Spectral Line Shapes in Astrophysics"

Inferring biological relationships from cellular phenotypes in high-content microscopy screens provides significant opportunity and challenge in biological research. Prior results have shown that deep vision models can capture biological signal better than hand-crafted features. This work explores how weakly supervised and self-supervised deep learning approaches scale when training larger models on larger datasets. Our results show that both CNN- and ViT-based masked autoencoders significantly outperform weakly supervised models. At the high-end of our scale, a ViT-L/8 trained on over 3.5-billion unique crops sampled from 95-million microscopy images achieves relative improvements as high as 28% over our best weakly supervised models at inferring known biological relationships curated from public databases., Comment: 4 pages, 4 figures

We studied seven nearby narrow-line Seyfert 1 (NLS1) galaxies in $J$- and $Ks$-bands with redshifts varying from 0.019 to 0.092. This is the first multi-source study targeting the hosts of southern NLS1 galaxies. Our data was obtained with the FourStar instrument of the 6.5~m Magellan Baade telescope at the Las Campanas Observatory (Chile). The aim of our study is to determine the host galaxy morphologies of these sources by using GALFIT. We were able to model six out of the seven sources reliably. Our conclusion is that all of the reliably modelled sources are disk-like galaxies, either spirals or lenticulars. None of these sources present elliptical morphology. Our findings are in agreement with the hypothesis that disk-like galaxies are the main host of jetted NLS1 galaxies. Taking advantage of observations in two bands, we also produced a $J - Ks$ colour map of each source. Five of the six colour maps show significant dust extinction near the core of the galaxy -- a feature often seen in gamma-ray-detected jetted NLS1 galaxies, and interpreted to be a consequence of a past minor merger.

Visual Place Recognition is a task that aims to predict the place of an image (called query) based solely on its visual features. This is typically done through image retrieval, where the query is matched to the most similar images from a large database of geotagged photos, using learned global descriptors. A major challenge in this task is recognizing places seen from different viewpoints. To overcome this limitation, we propose a new method, called EigenPlaces, to train our neural network on images from different point of views, which embeds viewpoint robustness into the learned global descriptors. The underlying idea is to cluster the training data so as to explicitly present the model with different views of the same points of interest. The selection of this points of interest is done without the need for extra supervision. We then present experiments on the most comprehensive set of datasets in literature, finding that EigenPlaces is able to outperform previous state of the art on the majority of datasets, while requiring 60\% less GPU memory for training and using 50\% smaller descriptors. The code and trained models for EigenPlaces are available at {\small{\url{https://github.com/gmberton/EigenPlaces}}}, while results with any other baseline can be computed with the codebase at {\small{\url{https://github.com/gmberton/auto_VPR}}}., Comment: ICCV 2023

Visual Place recognition is commonly addressed as an image retrieval problem. However, retrieval methods are impractical to scale to large datasets, densely sampled from city-wide maps, since their dimension impact negatively on the inference time. Using approximate nearest neighbour search for retrieval helps to mitigate this issue, at the cost of a performance drop. In this paper we investigate whether we can effectively approach this task as a classification problem, thus bypassing the need for a similarity search. We find that existing classification methods for coarse, planet-wide localization are not suitable for the fine-grained and city-wide setting. This is largely due to how the dataset is split into classes, because these methods are designed to handle a sparse distribution of photos and as such do not consider the visual aliasing problem across neighbouring classes that naturally arises in dense scenarios. Thus, we propose a partitioning scheme that enables a fast and accurate inference, preserving a simple learning procedure, and a novel inference pipeline based on an ensemble of novel classifiers that uses the prototypes learned via an angular margin loss. Our method, Divide&Classify (D&C), enjoys the fast inference of classification solutions and an accuracy competitive with retrieval methods on the fine-grained, city-wide setting. Moreover, we show that D&C can be paired with existing retrieval pipelines to speed up computations by over 20 times while increasing their recall, leading to new state-of-the-art results., Comment: Accepted to ICCV23