Web Image Research Articles

Oxygen Abundance Throughout the Dwarf Starburst Galaxy IC 10

Measurements of oxygen abundance throughout galaxies provide insight to the formation histories and ongoing processes. Here we present a study of the gas-phase oxygen abundance in the H ii regions and diffuse gas of the nearby starburst dwarf galaxy, IC 10. Using the Keck Cosmic Web Imager at W. M. Keck Observatory, we map the central region of IC 10 from 3500 to 5500 Å. The auroral [O III] 4363 Å line is detected with a high signal-to-noise ratio in 12 of 46 H II regions observed, allowing for direct measurement of the oxygen abundance, yielding a median and standard deviation of 12+log(O/H)=8.37±0.25 . We investigate trends between these directly measured oxygen abundances and other H II region properties, finding weak negative correlations with the radius, velocity dispersion, and luminosity. We also find weak negative correlations between the oxygen abundance and the derived quantities of turbulent pressure and ionized gas mass, and a moderate correlation with the derived dynamical mass. Strong line, R 23 abundance estimates are used in the remainder of the H II regions and on a resolved spaxel-by-spaxel basis. There is a large offset between the abundances measured with R 23 and the auroral line method. We find that the R 23 method is unable to capture the large range of abundances observed via the auroral line measurements. The extent of this variation in the measured abundances further indicates a poorly mixed interstellar medium in IC 10, which is not typical of dwarf galaxies and may be partly due to the ongoing starburst, accretion of pristine gas, or a late stage merger.

First Results from the Dragonfly Ultrawide Survey: The Largest 11 Quenched Diffuse Dwarf Galaxies in 3100 deg2 with Spectroscopic Confirmation

The Dragonfly Telephoto Array employs a unique design to detect very large and diffuse galaxies that might be missed with conventional telescopes. The Dragonfly Ultrawide Survey (DFUWS) is a new wide-field survey that will cover 10,000 deg2 of the northern sky, and it provides an ideal data set to find these large diffuse galaxies. From 3100 deg2 of DFUWS data, we identified 11 large, low surface brightness galaxies as a pilot sample for spectroscopic follow-up. These are the largest galaxies in the examined area that appear smooth and isolated, with effective radii of 12″–27″. Eight are below 24 magarcsec−2 in central g-band surface brightness. Keck Cosmic Web Imager spectra of the diffuse light show that all 11 galaxies in this sample are quiescent, and seven qualify as ultradiffuse galaxies (UDGs). Eight galaxies have distances between 15 and 30 Mpc, while the other three are in the Pegasus cluster at 50 Mpc. Their spectra show evidence of a ∼1 Gyr old stellar population in addition to an even older stellar population. The intermediate-age component is present in group and satellite galaxies but not in the Pegasus cluster UDGs. All galaxies in this sample are detected in both Dragonfly and Legacy imaging, and the sample partially overlaps with existing UDG catalogs. This pilot sample provides an excellent training set for our analysis of the upcoming full 10,000 deg2 DFUWS data, from which we may expect to discover even larger, previously unknown galaxies.

Expansion Properties of the Young Supernova Type Iax Remnant Pa 30 Revealed

The recently discovered Pa 30 nebula, the putative type Iax supernova remnant associated with the historical supernova of 1181 AD, shows puzzling characteristics that make it unique among known supernova remnants. In particular, Pa 30 exhibits a complex morphology, with a unique radial and filamentary structure, and it hosts a hot stellar remnant at its center, which displays oxygen-dominated, ultrafast winds. Because of the surviving stellar remnant and the lack of hydrogen and helium in its filaments, it has been suggested that Pa 30 is the product of a failed thermonuclear explosion in a near- or super-Chandrasekhar white dwarf, which created a subluminous transient, a rare subtype of the Ia class of supernovae called type Iax. We present here a detailed study of the 3D structure and velocities of a full radial section of the remnant. The Integral Field Unit observations, obtained with the new red channel of the Keck Cosmic Web Imager spectrograph, reveal that the ejecta are consistent with being ballistic, with velocities close to the free-expansion velocity. Additionally, we detect a large cavity inside the supernova remnant and a sharp inner edge to the filamentary structure, which coincides with the outer edge of a bright ring detected in infrared images. Finally, we detect a strong asymmetry in the amount of ejecta along the line of sight, which might hint at an asymmetric explosion. Our analysis provides strong confirmation that the explosion originated from SN 1181.

Observations of Extremely Metal-poor O Stars: Weak Winds and Constraints for Evolution Models

Metal-poor massive stars drive the evolution of low-mass galaxies, both locally and at high redshift. However, quantifying the feedback they impart to their local surroundings remains uncertain because models of stellar evolution, mass loss, and ionizing spectra are unconstrained by observations below 20% solar metallicity (Z ⊙). We present new Keck Cosmic Web Imager optical spectroscopy of three O-type stars in the nearby dwarf galaxies Leo P, Sextans A, and WLM, which have gas-phase oxygen abundances of 3%–14% Z ⊙. To characterize their fundamental stellar properties and radiation-driven winds, we fit PoWR atmosphere models to the optical spectra simultaneously with Hubble Space Telescope far-ultraviolet (FUV) spectra and multiwavelength photometry. We find that all three stars have effective temperatures consistent with their spectral types and surface gravities typical of main-sequence dwarf stars. Yet, the combination of those inferred parameters and luminosity for the two lower-Z stars is not reproduced by stellar evolution models, even those that include rotation or binary interactions. The scenario of multiple-star systems is difficult to reconcile with all available data, suggesting that these observations pose a challenge to current evolution models. We highlight the importance of validating the relationship between stellar mass, temperature, and luminosity at very low Z for accurate estimates of ionizing photon production and spectral hardness. Finally, all three stars’ FUV wind profiles reveal low mass-loss rates and terminal wind velocities in tension with expectations from widely adopted radiation-driven wind models. These results provide empirical benchmarks for future development of mass-loss and evolution models for metal-poor stellar populations.

Multi-Label and Evolvable Dataset Preparation for Web-Based Object Detection

In this paper, we focus on the emerging field of web-based object detection, which has gained considerable attention due to its ability to utilize large amounts of web data for training, thus eliminating the need for labor-intensive manual annotations. However, the noisy and ever-evolving nature of web data poses challenges in preparing high-quality datasets for web-based object detection. To address these challenges, we propose a fully automatic dataset preparation method in this paper. Our proposed method incorporates a hierarchical clustering module that assigns multiple precise labels to each image. This module is based on our observation that web image data exhibits different distributions at varying granularities. Furthermore, an evolutionary relabeling module ensures the adaptability of both the prepared dataset and trained detection models to the ever-evolving web data. Extensive experiments demonstrate that our method outperforms other web-based methods, and achieves a comparable performance to those manually labeled benchmark datasets.

The MOSDEF-KCWI Survey: Spectral Properties of Lyα Halos around z ∼ 2 Star-forming Galaxies

We present Keck Cosmic Web Imager integral field observations of extended Lyα emission in the circumgalactic medium of 27 typical star-forming galaxies at z ∼ 2, drawn from the Multi-Object Spectrometer for Infra-Red Exploration (MOSFIRE) Deep Evolution Field (MOSDEF) survey. Using composite spectra in two bins of star formation rate (SFR), star formation rate surface density (ΣSFR), and other galactic properties, we measure spatial variations in the Lyα profile across three regions in the Lyα halo. We find single-peaked, redshifted profiles are ubiquitous within a central 7 kpc radius region. Further out in the halo (7–14 and 14–21 kpc), the Lyα profile of the resonantly scattered emission exhibits more diversity, either transitioning to a double-peaked profile or remaining single peaked across the halo. We find a shorter scale length of the Lyα halo surface brightness profile for composite halos with faster winds. The composites have a similar average inclination, suggesting those with faster winds clear channels in the interstellar medium (ISM), reducing the fraction of Lyα photons resonantly scattered to large radii. A uniform expanding shell radiative transfer model reproduces the shape but not the normalization of the observed double-peaked Lyα profiles. Models that adopt a more realistic, clumpy ISM are likely needed to reproduce both the shape and normalization of the Lyα profiles.

Noise-Tolerant Hybrid Prototypical Learning with Noisy Web Data

We focus on the challenging problem of learning an unbiased classifier from a large number of potentially relevant but noisily labeled web images given only a few clean labeled images. This problem is particularly practical, because it reduces the expensive annotation costs by utilizing freely accessible web images with noisy labels. Typically, prototypes are representative images or features used to classify or identify other images. However, in the few clean and many noisy scenarios, the class prototype can be severely biased due to the presence of irrelevant noisy images. The resulting prototypes are less compact and discriminative, as previous methods do not take into account the diverse range of images in the noisy web image collections. On the other hand, the relation modeling between noisy and clean images is not learned for the class prototype generation in an end-to-end manner, which results in a suboptimal class prototype. In this paper, we introduce a similarity maximization loss named the SimNoiPro. Our SimNoiPro first generates noise-tolerant hybrid prototypes composed of clean and noise-tolerant prototypes, and then pulls them closer to each other. Our approach considers the diversity of noisy images by explicit division and overcomes the optimization discrepancy issue. This enables better relation modeling between clean and noisy images and helps extract judicious information from the noisy image set. The evaluation results on two extended few-shot classification benchmarks confirm that our SimNoiPro outperforms prior methods in measuring image relations and cleaning noisy data.

External validation of an artificial intelligence-based method for the detection and classification of molar incisor hypomineralisation in dental photographs

ObjectivesThis ex vivo diagnostic study aimed to externally validate an open-access artificial intelligence (AI)-based model for the detection, classification, localisation and segmentation of enamel/molar incisor hypomineralisation (EH/MIH). MethodsAn independent sample of web images showing teeth with (n = 277) and without (n = 178) EH/MIH was evaluated by a workgroup of dentists whose consensus served as the reference standard. Then, an AI-based model was used for the detection of EH/MIH, followed by automated classification and segmentation of the findings (test method). The accuracy (ACC), sensitivity (SE), specificity (SP) and area under the curve (AUC) were determined. Furthermore, the correctness of EH/MIH lesion localisation and segmentation was evaluated. ResultsAn overall ACC of 94.3 % was achieved for image-based detection of EH/MIH. Cross-classification of the AI-based class prediction and the reference standard resulted in an agreement of 89.2 % for all diagnostic decisions (n = 594), with an ACC between 91.4 % and 97.8 %. The corresponding SE and SP values ranged from 81.7 % to 92.8 % and 91.9 % to 98.7 %, respectively. The AUC varied between 0.894 and 0.945. Image size had only a limited impact on diagnostic performance. The AI-based model correctly predicted EH/MIH localisation in 97.3 % of cases. For the detected lesions, segmentation was fully correct in 63.4 % of all cases and partially correct in 33.9 %. ConclusionsThis study documented the promising diagnostic performance of an open-access AI tool in the detection and classification of EH/MIH in external images. Clinical significanceExternally validated AI-based diagnostic methods could facilitate the detection of EH/MIH lesions in dental photographs.

Ly α Halo Properties and Dust in the Circumgalactic Medium of z ∼ 2 Star-forming Galaxies

We present Keck Cosmic Web Imager integral-field unit observations around extended Lyα halos of 27 typical star-forming galaxies with redshifts 2.0 < z < 3.2 drawn from the MOSFIRE Deep Evolution Field survey. We examine the average Lyα surface brightness profiles in bins of star formation rate (SFR), stellar mass (M *), age, stellar continuum reddening, SFR surface density (ΣSFR), and ΣSFR normalized by stellar mass (ΣsSFR). The scale lengths of the halos correlate with stellar mass, age, and stellar continuum reddening and anticorrelate with SFR, ΣSFR, and ΣsSFR. These results are consistent with a scenario in which the down-the-barrel fraction of Lyα emission is modulated by the low-column-density channels in the interstellar medium, and in which the neutral gas covering fraction is related to the physical properties of the galaxies. Specifically, we find that this covering fraction increases with stellar mass, age, and E(B − V) and decreases with SFR, ΣSFR, and ΣsSFR. We also find that the resonantly scattered Lyα emission suffers greater attenuation than the (nonresonant) stellar continuum emission, and that the difference in attenuation increases with stellar mass, age, and stellar continuum reddening, and decreases with ΣsSFR. These results imply that more reddened galaxies have more dust in their circumgalactic medium.

Linking Mg II and [O II] spatial distribution to ionizing photon escape in confirmed LyC leakers and non-leakers

The geometry of the neutral gas in and around galaxies is a key regulator of the escape of ionizing photons. We present the first statistical study aimed at linking the neutral and ionized gas distributions to the Lyman continuum (LyC) escape fraction (fescLyC) in a sample of 22 confirmed LyC leakers and non-leakers atz ≈ 0.35 using the Keck Cosmic Web Imager (Keck/KCWI) and the Low Resolution Spectrograph 2 (HET/LRS2). Our integral field unit data enable the detection of neutral and low-ionization gas, as traced by Mg II, and ionized gas, as traced by [O II], extending beyond the stellar continuum for seven and ten objects, respectively. All but one object with extended Mg IIemission also show extended [O II] emission; in this case, Mg IIemission is always more extended than [O II] by a factor 1.2 on average. Most of the galaxies with extended emission are non or weak LyC leakers (fescLyC &lt; 5%), but we find a large diversity of neutral and low-ionization gas configurations around these weakly LyC-emitting galaxies. Conversely, the strongest leakers (fescLyC &gt; 5%) appear uniformly compact in both Mg IIand [O II] with exponential scale lengths ≲1 kpc. Most are unresolved at the resolution of our data. We also find a trend betweenfescLyCand the spatial offsets of the nebular gas and the stellar continuum emission. Moreover, we find significant anticorrelations between the spatial extent of the neutral and/or low-ionization gas and the [O III]/[O II] ratio, and Hβequivalent width, as well as positive correlations with metallicity and UV size, suggesting that galaxies with more compact neutral and/or low-ionization gas sizes are more highly ionized. The observations suggest that strong LyC emitters do not have extended neutral and/or low-ionization gas halos and ionizing photons may be emitted in many directions. Combined with high ionization diagnostics, we propose that the Mg II, and potentially [O II], spatial compactness are indirect indicators of LyC emitting galaxies at high redshift.

H i Lyα Emission from a Metal-poor Cool Stream Fueling an Early Dusty Starburst

The GAMA J0913−0107 system is a rare conjunction of a submillimeter galaxy (SMG) at z ≈ 2.7 and two background QSOs with projected separations <200 kpc. Previous high-resolution QSO absorption-line spectroscopy has revealed high H i column density, extremely metal-poor (∼1% solar) gas streams in the circumgalactic medium of the SMG. Here we present deep optical integral-field spectroscopy of the system with the Keck Cosmic Web Imager (KCWI). Reaching a 2σ surface brightness limit ≈10−19 erg s−1 cm−2 arcsec−2 with ∼2 hr of integration time, we detect a filamentary Lyα nebula stretching ∼180 kpc from the SMG intercepting both QSO sightlines. This Lyα filament may correspond to the same cool gas stream penetrating through the hot halo seen in the absorption. In contrast to Lyα nebulae around QSOs, there is no obvious local source for photoionization due to the massive dust content. While uncertain, we consider the possibility that the nebula is ionized by shocks induced by the infall, obscured star formation, and/or a boosted UV background. The SMG–QSOs conjunction multiplied the efficiency of the KCWI observations, allowing a direct comparison of Lyα nebulae in two distinct environments. We find that the nebulae around the QSOs are much brighter and show steeper surface brightness profiles than the SMG nebula. This is consistent with the additional photoionization and Lyα scattering provided by the QSOs. While illustrating the challenges of detecting Lyα nebulae around SMGs, our work also demonstrates that important insights can be gained from comparative studies of high-z Lyα nebulae.

Unveiling the complex circumgalactic medium: a comparative study of merging and non-interacting galaxy groups

ABSTRACT While most galaxies live in group environments where they undergo an accelerated evolution, the characteristics of their circumgalactic medium (CGM) remain uncertain. We present an analysis of the CGM of two galaxy groups in different stages of interaction: (G1) a close pair of galaxies (z = 0.043) separated by 87 kpc that do not show signs of interactions and (G2) four merging galaxies (z = 0.098) separated by 10 kpc. We present spatially resolved Keck/Keck Cosmic Web Imager galaxy observations and Hubble Space Telescope (HST)/COS quasar spectra (G1 at 48 kpc and G2 at 100 kpc away) to quantify both the resolved galaxy and CGM properties in these two different group environments. G1 contains two typical star-forming galaxies with no evidence of strong outflows. G2 contains two star-forming, one post-starburst and one quiescent galaxy. Both groups have a range of CGM-detected metal lines (H i, C ii, Si ii, Si iii, N v, and O vi). Despite G2 being twice as far from the quasar, G2 has $\log (N({{{\rm H}{\small I}}})/{\rm cm}^{-2})=17.33$, compared to $\log (N({{{\rm H}{\small I}}})/{\rm cm}^{-2})=16.43$ for G1. We find that the CGM of the merging galaxies (G2) is more kinematically complex, is in a higher ionization state, spans a wider range of metallicities and column densities, has smaller cloud sizes, and is inconsistent with the simple superposition model that seems to match well with G1. We conclude that the complexity of the CGM in merging galaxies surpasses that of not strongly interacting galaxies, suggesting that mergers play a significant role in shaping the intricate structure of the CGM.

Analysis of galaxies at the extremes: a kinematic analysis of the Virgo cluster dwarfs VCC 9 and VCC 1448 using the Keck cosmic web imager

ABSTRACT We present spatially resolved Keck Cosmic Web Imager stellar spectroscopy of the Virgo cluster dwarf galaxies VCC 9 and VCC 1448. These galaxies have similar stellar masses and large half-light radii but very different globular cluster (GC) system richness (∼25 versus ∼99 GCs). Using the KCWI data, we spectroscopically confirm 10 GCs associated with VCC 1448 and one GC associated with VCC 9. We make two measurements of dynamical mass for VCC 1448 based on the stellar and GC velocities, respectively. VCC 1448’s mass measurements suggest that it resides in a halo in better agreement with the expectation of the stellar mass–halo mass relationship than the expectation from its large GC counts. For VCC 9, the dynamical mass we measure agrees with the expected halo mass from both relationships. We compare VCC 1448 and VCC 9 to the GC-rich galaxy Dragonfly 44 (∼74 GCs), which is similar in size but has ∼1 dex less stellar mass than either Virgo galaxy. In dynamical mass – GC number space, Dragonfly 44 and VCC 1448 exhibit richer GC systems given their dynamical mass than that of VCC 9 and other ‘normal’ galaxies. We also place the galaxies in kinematics–ellipticity space finding evidence of an anticorrelation between rotational support and the fraction of a galaxy’s stellar mass in its GC system, that is, VCC 9 is more rotationally supported than VCC 1448, which is more rotationally supported than Dragonfly 44. This trend may be expected if a galaxy’s GC content depends on its natal gas properties at formation.

Resolving the Mechanical and Radiative Feedback in J1044+0353 with Keck Cosmic Web Imager Spectral Mapping

We present integral field spectroscopy toward and around J1044+0353, a rapidly growing, low-metallicity galaxy that produces extreme [O iii] line emission. A new map of the O32 flux ratio reveals a density-bounded ionization cone emerging from the starburst. The interaction of the hydrogen-ionizing radiation, produced by the very young starburst, with a cavity previously carved out by a galactic outflow, whose apex lies well outside the starburst region, determines the pathway for global Lyman continuum (LyC) escape. In the region within a few hundred parsecs of the young starburst, we demonstrate that superbubble breakthrough and blowout contribute distinct components to the [O iii] line profile: broad and very broad emission line wings, respectively. We draw attention to the large [O iii] luminosity of the broad component and argue that this emission comes from photoionized, superbubble shells rather than a galactic wind as is often assumed. The spatially resolved He ii λ4686 nebula appears to be photoionized by young star clusters. Stellar wind emission from these stars is likely the source of line wings detected on the He ii line profile. This broader He ii component indicates slow stellar winds, consistent with an increase in stellar rotation (and a decrease in effective escape speed) at the metallicity of J1044+0353. At least in J1044+0353, the recent star formation history plays a critical role in generating a global pathway for LyC escape, and the anisotropic escape would likely be missed by direct observations of the LyC.

Investigating the Drivers of Electron Temperature Variations in H ii Regions with Keck-KCWI and VLT-MUSE

H ii region electron temperatures are a critical ingredient in metallicity determinations, and recent observations have revealed systematic variations in the temperatures measured using different ions. We present electron temperatures (T e ) measured using the optical auroral lines ([N ii]λ5756, [O ii]λ λ7320, 7330, [S ii]λ λ4069, 4076, [O iii]λ4363, and [S iii]λ6312) for a sample of H ii regions in seven nearby galaxies. We use observations from the Physics at High Angular resolution in Nearby Galaxies survey (PHANGS) obtained with integral field spectrographs on Keck (Keck Cosmic Web Imager) and the Very Large Telescope (Multi-Unit Spectroscopic Explorer). We compare the different T e measurements with H ii region and ISM environmental properties such as electron density, ionization parameter, molecular gas velocity dispersion, and stellar association/cluster mass and age obtained from PHANGS. We find that the temperatures from [O ii] and [S ii] are likely overestimated due to the presence of electron density inhomogeneities in H ii regions. We measure high [O iii] temperatures in a subset of regions with high molecular gas velocity dispersion and low ionization parameter, which may be explained by the presence of low-velocity shocks. In agreement with previous studies, the T e–T e between [N ii] and [S iii] temperatures have the lowest observed scatter and follow predictions from photoionization modeling, which suggests that these tracers reflect H ii region temperatures across the various ionization zones better than [O ii], [S ii], and [O iii].

DUVET: Resolved direct metallicity measurements in the outflow of starburst galaxy NGC 1569

Abstract We present the results of direct-method metallicity measurements in the disk and outflow of the low-metallicity starburst galaxy NGC 1569. We use Keck Cosmic Web Imager observations to map the galaxy across 54″ (800 pc) along the major axis and 48″ (700 pc) along the minor axis with a spatial resolution of 1″ (∼15 pc). We detect common strong emission lines ([O iii] λ5007, Hβ, [O ii] λ3727) and the fainter [O iii] λ4363 auroral line, which allows us to measure electron temperature (Te) and metallicity. Theory suggests that outflows drive metals out of the disk driving observed trends between stellar mass and gas-phase metallicity. Our main result is that the metallicity in the outflow is similar to that of the disk, Zout/ZISM ≈ 1. This is consistent with previous absorption line studies in higher mass galaxies. Assumption of a mass-loading factor of $\dot{M}_{\rm out}/{\rm SFR}\sim 3$ makes the metal-loading of NGC 1569 consistent with expectations derived from the mass-metallicity relationship. Our high spatial resolution metallicity maps reveal a region around a supermassive star cluster (SSC-B) with distinctly higher metallicity and higher electron density, compared to the disk. Given the known properties of SSC-B the higher metallicity and density of this region are likely the result of star formation-driven feedback acting on the local scale. Overall, our results are consistent with the picture in which metal-enriched winds pollute the circumgalactic medium surrounding galaxies, and thus connect the small-scale feedback processes to large-scale properties of galaxy halos.

Decolonizing gender through ICT: a semiotic analysis of web images from Two-Spirit people websites

ABSTRACT Freedom of self-expression is essential for human development. Recent studies have investigated ICT’s role in bringing visibility to activist movements working to improve the lives of gender and sexual minorities. However, this literature often characterizes visibility based on gender or sexuality alone, with little regard for ethnic identity and gender traditions of Indigenous cultures. This paper conducts a semiotic analysis of images from Two-Spirit websites to examine how websites bring visibility to Two-Spirit people. Two-Spirit is a contemporary term used by Indigenous Peoples of North America that denotes the fluidity of gender identity through the coexistence of feminine and masculine qualities. The analysis shows how values of empowerment, supportiveness, and restoration in web images bring visibility to Two-Spirit people through cultural emancipation. This study contributes to the ICT for development literature by broadening our understanding of ICT’s role in bringing visibility to efforts for self-expression free of racism and homophobia.

Robust fine‐grained visual recognition with images based on internet of things

AbstractLabeling fine‐grained objects manually is extremely challenging, as it is not only label‐intensive but also requires professional knowledge. Accordingly, robust learning methods for fine‐grained recognition with web images collected from Internet of Things have drawn significant attention. However, training deep fine‐grained models directly using untrusted web images is confronted by two primary obstacles: (1) label noise in web images and (2) domain variance between the online sources and test datasets. To this end, in this study, we mainly focus on addressing these two pivotal problems associated with untrusted web images. To be specific, we introduce an end‐to‐end network that collaboratively addresses these concerns in the process of separating trusted data from untrusted web images. To validate the efficacy of our proposed model, untrusted web images are first collected by utilizing the text category labels found within fine‐grained datasets. Subsequently, we employ the designed deep model to eliminate label noise and ameliorate domain mismatch. And the chosen trusted web data are utilized for model training. Comprehensive experiments and ablation studies validate that our method consistently surpasses other state‐of‐the‐art approaches for fine‐grained recognition tasks in real‐world scenarios, demonstrating a significant improvement margin (2.51% on CUB200‐2011 and 2.92% on Stanford Dogs). The source code and models can be accessed at: https://github.com/Codeczh/FGVC‐IoT.

Group benefits instance for data purification

Manually annotating datasets for training deep models is very labor-intensive and time-consuming. To overcome such inferiority, directly leveraging web images to conduct training data becomes a natural choice. Nevertheless, the presence of label noise in web data usually degrades the model performance. Existing methods for combating label noise are typically designed and tested on synthetic noisy datasets. However, they tend to fail to achieve satisfying results on real-world noisy datasets. To this end, we propose a method named GRIP to alleviate the noisy label problem for both synthetic and real-world datasets. Specifically, GRIP utilizes a group regularization strategy that estimates class soft labels to improve noise-robustness. Soft label supervision reduces overfitting on noisy labels and learns inter-class similarities to benefit classification. Furthermore, an instance purification operation globally identifies noisy labels by measuring the difference between each training sample and its class soft label. Through operations at both group and instance levels, our approach integrates the advantages of noise-robust and noise-cleaning methods and remarkably alleviates the performance degradation caused by noisy labels. Comprehensive experimental results on synthetic and real-world datasets demonstrate the superiority of GRIP over the existing state-of-the-art methods. The data and source code of this work have been made available at: https://github.com/NUST-Machine-Intelligence-Laboratory/GRIP.

Circumgalactic Environments Around Distant Quasars 3C 9 and 4C 05.84

We present results from the Quasar hosts Unveiled by high Angular Resolution Techniques survey studying the circumgalactic medium (CGM) by observing rest-frame UV emission lines Lyα, C iv, and He ii around two radio-loud quasars, 3C 9 (z = 2.02) and 4C 05.84 (z = 2.32), using the Keck Cosmic Web Imager. We detect large-scale Lyα nebulae around both quasars with projected diameters ∼100 kpc, with spatially resolved, embedded 15–30 kpc He ii and C iv nebulae around both quasars as well as kinematically distinct He ii and C iv nebulae at a physical separation of ∼15 kpc from both quasars. Observations of Hα, Hβ, and [O iii] emission using Keck MOSFIRE spectroscopically confirm that the Lyα nebulae extend to companion galaxies and that these quasars are in a protogroup/protocluster environment. We confirm that the He ii and C iv emission is kinematically and spatially coincident with the companion galaxies. We estimate the virial masses of the companion galaxies, their metallicities, and star formation rates, and investigate the sources of ionization. We measure the dynamical mass of the host dark matter halos and estimate that the dark matter halos of these systems will grow to a mass of 2 × 1014 M ⊙ (3C 9) and 2 × 1013 M ⊙ (4C 05.84) by z = 0. The combined CGM and companion galaxies observations indicate Lyα substructure can indicate the presence of companion galaxies in the CGM.