arXiv subjects

Christian Wolf

Publications and source records attributed to Christian Wolf.

At least 37 records · Page 2Linked to original sources

arXiv2025

An Extremely-High Velocity Outflow in SMSS J2157-3602, the most luminous quasar in the first 1.3 Gyr

We report the discovery of an extremely high-velocity outflow (EHVO) in the most luminous ($L\rm_{Bol}$ $\sim$ 2.29 $\times$ 10$^{48}$ erg/s) QSO, SMSS J2157-3602, at z=4.692. Combined XSHOOTER and NIRES observations reveal that the EHVO reaches a maximum velocity of v$_\mathrm{max} \sim 0.13c$ and persists over rest-frame timescales of a few months to one year. SMSS J2157-3602 also exhibits one of the highest balnicity index discovered in an EHVO so far. In addition, the blueshifted CIV emission traces a high-velocity (v$\rm_{CIV}^{50}\sim$ 4660 km/s) outflow from the broad-line region. Thanks to an XMM-Newton observation, we also discover the X-ray weak nature of this QSO, which likely prevents the overionization of the innermost disk atmosphere and facilitates the efficient launch of the detected EHVO and BLR winds. The extraordinary luminosity of SMSS J2157-3602 and the extreme velocity of the EHVO make it a unique laboratory for testing AGN driven feedback under extreme conditions. Current uncertainties on the outflow's location and column density strengthen the case for dedicated follow-up, which will be essential to assess the full feedback potential of this remarkable quasar.

astro-ph.GA↗

arXiv2025

Computability for Axiom A Polynomial Skew Products of $\mathbb{C}^2$

The computability of Julia sets of rational maps on the Riemann sphere has been intensively studied in recent years (see, e.g. https://doi.org/10.17323/1609-4514-2008-8-2-185-231, https://doi.org/10.1090/conm/797/15936) for an overview. For example, by Braverman's results (https://doi.org/10.1016/j.entcs.2004.06.031, https://doi.org/10.1088/0951-7715/19/6/009), hyperbolic and parabolic Julia sets are computable in polynomial time. In this paper, we present the first work on computability related to maps of more than one complex dimension. We examine a family of polynomial endomorphisms of $\mathbb{C}^2$, the polynomial skew products; i.e., maps of the form $f(z,w) = (p(z), q(z,w)),$ where $p$ and $q$ are complex polynomials of the same degree $d\geq 2$. We show that if a polynomial skew product is Axiom A, then its chain recurrent set, which is equal to its non-wandering set and also equal to the closure of the periodic orbits, is computable. Our algorithm also identifies the various hyperbolic sets of different types, i.e., expanding, attracting, and hyperbolic sets of saddle-type. One consequence of our results is that Axiom A is a semi-decidable property on the closure of the Axiom A polynomial skew product locus. Finally, we introduce an algorithm that establishes the lower semi-computability of the hyperbolicity locus of polynomial skew products of a fixed degree.

math.DS↗

arXiv2025

The CatSouth Quasar Candidate Catalog for the Southern Sky and a Unified All-Sky Catalog Based on Gaia DR3

The Gaia DR3 has provided a large sample of more than 6.6 million quasar candidates with high completeness but low purity. Previous work on the CatNorth quasar candidate catalog has shown that including external multiband data and applying machine-learning methods can efficiently purify the original Gaia DR3 quasar candidate catalog and improve the redshift estimates. In this paper, we extend the Gaia DR3 quasar candidate selection to the southern hemisphere using data from SkyMappper, CatWISE, and VISTA surveys. We train an XGBoost classifier on a unified set of high-confidence stars and spectroscopically confirmed quasars and galaxies. For sources with available Gaia BP/RP spectra, spectroscopic redshifts are derived using a pre-trained convolutional neural network (RegNet). We also train an ensemble photometric redshift estimation model based on XGBoost, TabNet, and FT-Transformer, achieving an RMSE of 0.2256 and a normalized median absolute deviation of 0.0187 on the validation set. By merging CatSouth with the previously published CatNorth catalog, we construct the unified all-sky CatGlobe catalog with nearly 1.9 million sources at $G<21$, providing a comprehensive and high-purity quasar candidate sample for future spectroscopic and cosmological investigations.

astro-ph.GA↗

arXiv2025

RANa: Retrieval-Augmented Navigation

Methods for navigation based on large-scale learning typically treat each episode as a new problem, where the agent is spawned with a clean memory in an unknown environment. While these generalization capabilities to an unknown environment are extremely important, we claim that, in a realistic setting, an agent should have the capacity of exploiting information collected during earlier robot operations. We address this by introducing a new retrieval-augmented agent, trained with RL, capable of querying a database collected from previous episodes in the same environment and learning how to integrate this additional context information. We introduce a unique agent architecture for the general navigation task, evaluated on ImageNav, Instance-ImageNav and ObjectNav. Our retrieval and context encoding methods are data-driven and employ vision foundation models (FM) for both semantic and geometric understanding. We propose new benchmarks for these settings and we show that retrieval allows zero-shot transfer across tasks and environments while significantly improving performance.

cs.CV↗

arXiv2025

Reasoning in visual navigation of end-to-end trained agents: a dynamical systems approach

Progress in Embodied AI has made it possible for end-to-end-trained agents to navigate in photo-realistic environments with high-level reasoning and zero-shot or language-conditioned behavior, but benchmarks are still dominated by simulation. In this work, we focus on the fine-grained behavior of fast-moving real robots and present a large-scale experimental study involving \numepisodes{} navigation episodes in a real environment with a physical robot, where we analyze the type of reasoning emerging from end-to-end training. In particular, we study the presence of realistic dynamics which the agent learned for open-loop forecasting, and their interplay with sensing. We analyze the way the agent uses latent memory to hold elements of the scene structure and information gathered during exploration. We probe the planning capabilities of the agent, and find in its memory evidence for somewhat precise plans over a limited horizon. Furthermore, we show in a post-hoc analysis that the value function learned by the agent relates to long-term planning. Put together, our experiments paint a new picture on how using tools from computer vision and sequential decision making have led to new capabilities in robotics and control. An interactive tool is available at europe.naverlabs.com/research/publications/reasoning-in-visual-navigation-of-end-to-end-trained-agents.

cs.RO↗

arXiv2025

Detection of Ubiquitous Circumbinary Matter in Hot Subdwarfs Formed from Common-Envelope Ejections

The formation of compact binary systems is largely driven by their evolution through a common envelope (CE) phase, crucial for understanding phenomena such as type Ia supernovae and black hole mergers. Despite their importance, direct observational evidence for CE material has been elusive due to the transient nature of these envelopes. Numerical simulations suggest that some envelope material may persist post-ejection. In this study, we investigate circumstellar material (CSM) surrounding hot subdwarf (sdB) stars, focusing on material ejected during the CE phase of binary evolution. We analyze Ca II K absorption lines in 727 sdB candidates from the LAMOST-LRS survey, selecting 145 stars with strong absorption features, indicating the presence of CSM. We compare the velocities of the Ca II K lines with the systemic velocities of sdB binaries, confirming that the material originates from ejected common-envelope material. The results show that the CSM persists long after the CE event, suggesting the formation of a stable, long-lived circumstellar environment around sdB stars. This study enhances our understanding of the role of CSM in post-CE evolution and provides new insights into the physical processes shaping the evolution of sdB binaries.

astro-ph.SR↗

arXiv2024

The Variability Structure Function of the Highest-Luminosity Quasars on Short Timescales

The stochastic photometric variability of quasars is known to follow a random-walk phenomenology on emission timescales of months to years. Some high-cadence restframe optical monitoring in the past has hinted at a suppression of variability amplitudes on shorter timescales of a few days or weeks, opening the question of what drives the suppression and how it might scale with quasar properties. Here, we study a few thousand of the highest-luminosity quasars in the sky, mostly in the luminosity range of $L_{\rm bol}=[46.4, 47.3]$ and redshift range of $z=[0.7, 2.4]$. We use a dataset from the NASA/ATLAS facility with nightly cadence, weather permitting, which has been used before to quantify strong regularity in longer-term restframe-UV variability. As we focus on a careful treatment of short timescales across the sample, we find that a linear function is sufficient to describe the UV variability structure function. Although the result can not rule out the existence of breaks in some groups completely, a simpler model is usually favoured under this circumstance. In conclusion, the data is consistent with a single-slope random walk across restframe timescales of $Δt=[10, 250]$ days.

astro-ph.GA↗

arXiv2024

Discovering Changing-Look AGN in the 6dF Galaxy Survey using ATLAS Light curves

Changing-Look Active Galactic Nuclei (CLAGN) are characterised by extreme variations in line emission over short timescales, mostly affecting broad H$β$ lines. While a few hundred CLAGN are known, a complete sample of turn-on CLAGN is still elusive. Here, we present a search for turn-on CLAGN in a complete sample of galaxies, using archival spectra and recent light curves. We obtained light curves from the Asteroid Terrestrial Impact Last Alert System (ATLAS) for 16,232 emission line galaxies, including both star-forming and active galaxies, at $z<0.1$ with spectra from the Six-degree Field Galaxy Survey (6dFGS). We first establish typical variability behaviour for different AGN types, as recorded between 2001 and 2009, and then select outliers from the bulk behaviour as CLAGN candidates. We obtain new spectra for the candidates and identify 12 new turn-on CLAGN (appearing broad H$β$ line) and 19 new turn-off CLAGN (disappearing broad H$β$ line). We may have missed AGN that changed and reverted their state over the 15$-$20 years since 6dFGS spectra were taken, and thus our CLAGN rates of 1.7$\%$ for turn-on and 9.6$\%$ for turn-off are lower limits. The turn-on rate is naturally much lower as the type 1.9/2 sample is dominated by obscured AGN due to orientation, which are not expected to change. However, the number of turn-on (27) and turn-off (24) CLAGN we find are similar, suggesting that our parent AGN sample is reasonably complete in our search volume at $z<0.1$.

astro-ph.GA↗

arXiv2024

Broad-line Active Galactic Nuclei in the 6dF Galaxy Survey

The Six-degree Field Galaxy Survey (6dFGS) is a spectroscopic redshift survey of the Southern hemisphere completed in 2006. While it provides 136,304 spectra of mostly low-redshift galaxies, a large and reliable catalogue of Active Galactic Nuclei (AGN) that are selected based on spectral signatures is still lacking. In this work, we present an extensive list of verified broad emission-line AGN in the 6dFGS sample. We visually confirm the AGN nature of all spectra, and disentangle fibre cross-talk to remove bogus AGN. The final catalogue contains 2,515 unique broad-line AGN with a median redshift of 0.207, of which 891 are identified for the first time. A flux-limited subsample contains 665 AGN to a K-band magnitude of 13. This new sample adds to the list of known low-luminosity AGN in the Southern hemisphere and thus provides a basis for investigations of low-redshift AGN with the upcoming Vera C. Rubin Observatory.

astro-ph.GA↗

arXiv2024

Ergodic theory on coded shift spaces

We study ergodic-theoretic properties of coded shift spaces. A coded shift space is defined as a closure of all bi-infinite concatenations of words from a fixed countable generating set. We derive sufficient conditions for the uniqueness of measures of maximal entropy and equilibrium states of Hoelder continuous potentials based on the partition of the coded shift into its concatenation set (sequences that are concatenations of generating words) and its residual set (sequences added under the closure). In this case we provide a simple explicit description of the measure of maximal entropy. We also obtain flexibility results for the entropy on the concatenation and residual sets. Finally, we prove a local structure theorem for intrinsically ergodic coded shift spaces which shows that our results apply to a larger class of coded shift spaces compared to previous works by Climenhaga, Climenhaga and Thompson, and Pavlov.

math.DS↗

arXiv2024

Supermassive black holes are growing slowly by $z\sim5$

We investigate the black hole mass function at $z\sim5$ using XQz5, our recent sample of the most luminous quasars between the redshifts $4.5 < z < 5.3$. We include 72 quasars with black hole masses estimated from velocity-broadened emission-line measurements and single-epoch virial prescriptions in the footprint of a highly complete parent survey. The sample mean Eddington ratio and standard deviation is $\logλ\approx -0.20\pm0.24$. The completeness-corrected mass function is modelled as a double power-law, and we constrain its evolution across redshift assuming accretion-dominated mass growth. We estimate the evolution of the mass function from $z=5-4$, presenting joint constraints on accretion properties through a measured dimensionless e-folding parameter, $k_{\rm{ef}} \equiv \langleλ\rangle U (1-ε)/ε= 1.79\pm0.06$, where $\langleλ\rangle$ is the mean Eddington ratio, $U$ is the duty cycle, and $ε$ is the radiative efficiency. If these supermassive black holes were to form from seeds smaller than $10^8\,M_{\odot}$, the growth rate must have been considerably faster at $z\gg5$ than observed from $z=5-4$. A growth rate exceeding $3\times$ the observed rate would reduce the initial heavy seed mass to $10^{5-6}\,M_{\odot}$, aligning with supermassive star and/or direct collapse seed masses. Stellar mass ($10^2\,M_{\odot}$) black hole seeds would require $\gtrsim4.5\times$ the observed growth rate at $z\gg5$ to reproduce the measured active black hole mass function. A possible pathway to produce the most extreme quasars is radiatively inefficient accretion flow, suggesting black holes with low angular momentum or photon trapping in supercritically accreting thick discs.

astro-ph.GA↗

arXiv2024

Task-conditioned adaptation of visual features in multi-task policy learning

Successfully addressing a wide variety of tasks is a core ability of autonomous agents, requiring flexibly adapting the underlying decision-making strategies and, as we argue in this work, also adapting the perception modules. An analogical argument would be the human visual system, which uses top-down signals to focus attention determined by the current task. Similarly, we adapt pre-trained large vision models conditioned on specific downstream tasks in the context of multi-task policy learning. We introduce task-conditioned adapters that do not require finetuning any pre-trained weights, combined with a single policy trained with behavior cloning and capable of addressing multiple tasks. We condition the visual adapters on task embeddings, which can be selected at inference if the task is known, or alternatively inferred from a set of example demonstrations. To this end, we propose a new optimization-based estimator. We evaluate the method on a wide variety of tasks from the CortexBench benchmark and show that, compared to existing work, it can be addressed with a single policy. In particular, we demonstrate that adapting visual features is a key design choice and that the method generalizes to unseen tasks given a few demonstrations.

cs.CV↗

arXiv2024

Zero-BEV: Zero-shot Projection of Any First-Person Modality to BEV Maps

Bird's-eye view (BEV) maps are an important geometrically structured representation widely used in robotics, in particular self-driving vehicles and terrestrial robots. Existing algorithms either require depth information for the geometric projection, which is not always reliably available, or are trained end-to-end in a fully supervised way to map visual first-person observations to BEV representation, and are therefore restricted to the output modality they have been trained for. In contrast, we propose a new model capable of performing zero-shot projections of any modality available in a first person view to the corresponding BEV map. This is achieved by disentangling the geometric inverse perspective projection from the modality transformation, eg. RGB to occupancy. The method is general and we showcase experiments projecting to BEV three different modalities: semantic segmentation, motion vectors and object bounding boxes detected in first person. We experimentally show that the model outperforms competing methods, in particular the widely used baseline resorting to monocular depth estimation.

cs.CV↗

arXiv2024

Discovery of A New Blue Large-Amplitude Pulsator in the SkyMapper DR2: SMSS J184506.82-300804.7

We report the discovery of a new Blue Large-Amplitude Pulsator, SMSS J184506-300804 (SMSS-BLAP-1) in Data Release 2 of the SkyMapper Southern Sky Survey. We conduct high-cadence photometric observations in the $u$ band to confirm a periodic modulation of the lightcurve. SMSS-BLAP-1 has a ~19-min pulsation period with an amplitude of 0.2 mag in $u$ band, and is similar to the classical BLAPs found by OGLE. From spectroscopic observations with the Wide-Field Spectrograph on the ANU 2.3m telescope, we confirm it as a low-gravity BLAP: best-fit parameters from the non-LTE Tlusty model are estimated as $T_\mathrm{eff}$ = 29,020$^{+193}_{-34}$ K, $\log g$ = 4.661$^{+0.008}_{-0.143}$ (cm s$^{-2}$), and $\log$ n(He)/n(H) = -2.722$^{+0.057}_{-0.074}$ dex. However, our BLAP exhibits a very He-deficient atmosphere compared to both low- and high-gravity BLAPs, which have $\log$ n(He)/n(H) in the range -0.41~-2.4.

astro-ph.SR↗

arXiv2024

The accretion of a solar mass per day by a 17-billion solar mass black hole

Around a million quasars have been catalogued in the Universe by probing deeper and using new methods for discovery. However, the hardest ones to find seem to be the rarest and brightest specimen. In this work, we study the properties of the most luminous of all quasars found so far. It has been overlooked until recently, which demonstrates that modern all-sky surveys have much to reveal. The black hole in this quasar accretes around one solar mass per day onto an existing mass of $\sim$17 billion solar masses. In this process its accretion disc alone releases a radiative energy of $2\times 10^{41}$ Watts. If the quasar is not strongly gravitationally lensed, then its broad line region (BLR) is expected to have the largest physical and angular diameter occurring in the Universe, and will allow the Very Large Telescope Interferometer to image its rotation and measure its black hole mass directly. This will be an important test for BLR size-luminosity relations, whose extrapolation has underpinned common black-hole mass estimates at high redshift.

astro-ph.CO↗

arXiv2024

CatNorth: An Improved Gaia DR3 Quasar Candidate Catalog with Pan-STARRS1 and CatWISE

A complete and pure sample of quasars with accurate redshifts is crucial for quasar studies and cosmology. In this paper, we present CatNorth, an improved Gaia DR3 quasar candidate catalog with more than 1.5 million sources in the 3$π$ sky built with data from Gaia, Pan-STARRS1, and CatWISE2020. The XGBoost algorithm is used to reclassify the original Gaia DR3 quasar candidates as stars, galaxies, and quasars. To construct training/validation datasets for the classification, we carefully built two different master stellar samples in addition to the spectroscopic galaxy and quasar samples. An ensemble classification model is obtained by averaging two XGBoost classifiers trained with different master stellar samples. Using a probability threshold of $p_{\mathrm{QSO\_mean}}>0.95$ in our ensemble classification model and an additional cut on the logarithmic probability density of zero proper motion, we retrieved 1,545,514 reliable quasar candidates from the parent Gaia DR3 quasar candidate catalog. We provide photometric redshifts for all candidates with an ensemble regression model. For a subset of 89,100 candidates, accurate spectroscopic redshifts are estimated with the Convolutional Neural Network from the Gaia BP/RP spectra. The CatNorth catalog has a high purity of ~ 90% while maintaining high completeness, which is an ideal sample to understand the quasar population and its statistical properties. The CatNorth catalog is used as the main source of input catalog for the LAMOST phase III quasar survey, which is expected to build a highly complete sample of bright quasars with $i < 19.5$.

astro-ph.GA↗

arXiv2024

SkyMapper Southern Survey: Data Release 4

We present the fourth data release (DR4) of the SkyMapper Southern Survey (SMSS), the last major step in our hemispheric survey with six optical filters: u, v, g, r, i, z. SMSS DR4 covers 26,000 sq.deg from over 400,000 images acquired by the 1.3m SkyMapper telescope between 2014-03 and 2021-09. The 6-band sky coverage extends from the South Celestial Pole to Dec = +16deg, with some images reaching Dec ~ +28deg. In contrast to previous DRs, we include all good-quality images from the facility taken during that time span, not only those explicitly taken for the public Survey. From the image dataset, we produce a catalogue of nearly 13 billion detections made from ~700 million unique astrophysical objects. The typical 10sigma depths for each field range between 18.5 and 20.5 mag, depending on the filter, but certain sky regions include longer exposures that reach as deep as 22 mag in some filters. As with previous SMSS catalogues, we have cross-matched with a host of other imaging and spectroscopic datasets to facilitate additional science outcomes. SMSS DR4 is now available to the worldwide astronomical community.

astro-ph.CO↗

arXiv2024

Learning to navigate efficiently and precisely in real environments

In the context of autonomous navigation of terrestrial robots, the creation of realistic models for agent dynamics and sensing is a widespread habit in the robotics literature and in commercial applications, where they are used for model based control and/or for localization and mapping. The more recent Embodied AI literature, on the other hand, focuses on modular or end-to-end agents trained in simulators like Habitat or AI-Thor, where the emphasis is put on photo-realistic rendering and scene diversity, but high-fidelity robot motion is assigned a less privileged role. The resulting sim2real gap significantly impacts transfer of the trained models to real robotic platforms. In this work we explore end-to-end training of agents in simulation in settings which minimize the sim2real gap both, in sensing and in actuation. Our agent directly predicts (discretized) velocity commands, which are maintained through closed-loop control in the real robot. The behavior of the real robot (including the underlying low-level controller) is identified and simulated in a modified Habitat simulator. Noise models for odometry and localization further contribute in lowering the sim2real gap. We evaluate on real navigation scenarios, explore different localization and point goal calculation methods and report significant gains in performance and robustness compared to prior work.

cs.RO↗