Search

This paper illustrates the results obtained by using pre-trained semantic segmentation deep learning models for the detection of archaeological sites within the Mesopotamian floodplains environment. The models were fine-tuned using openly available satellite imagery and vector shapes coming from a large corpus of annotations (i.e., surveyed sites). A randomized test showed that the best model reaches a detection accuracy in the neighborhood of 80%. Integrating domain expertise was crucial to define how to build the dataset and how to evaluate the predictions, since defining if a proposed mask counts as a prediction is very subjective. Furthermore, even an inaccurate prediction can be useful when put into context and interpreted by a trained archaeologist. Coming from these considerations we close the paper with a vision for a Human-AI collaboration workflow. Starting with an annotated dataset that is refined by the human expert we obtain a model whose predictions can either be combined to create a heatmap, to be overlaid on satellite and/or aerial imagery, or alternatively can be vectorized to make further analysis in a GIS software easier and automatic. In turn, the archaeologists can analyze the predictions, organize their onsite surveys, and refine the dataset with new, corrected, annotation, Comment: 15 pages, 5 figures, 2 tables

We introduce a novel computational unit for neural networks that features multiple biases, challenging the traditional perceptron structure. This unit emphasizes the importance of preserving uncorrupted information as it is passed from one unit to the next, applying activation functions later in the process with specialized biases for each unit. Through both empirical and theoretical analyses, we show that by focusing on increasing biases rather than weights, there is potential for significant enhancement in a neural network model's performance. This approach offers an alternative perspective on optimizing information flow within neural networks. See source code at https://github.com/CuriosAI/dac-dev., Comment: Major rewriting. Supersedes v1 and v2. Focusing on the fact that not all parameters are born equal: biases can be more important than weights. Accordingly, new title and new abstract, and many more experiments on fully connected architectures. This is the extended version of the paper published at WACV 2024

Orientational order, encoded in anisotropic fields, plays an important role during the development of an organism. A striking example of this is the freshwater polyp Hydra, where topological defects in the muscle fiber orientation have been shown to localize to key features of the body plan. This body plan is organized by morphogen concentration gradients, raising the question how muscle fiber orientation, morphogen gradients and body shape interact. Here, we introduce a minimal model that couples nematic orientational order to the gradient of a morphogen field. We show that on a planar surface alignment to a radial concentration gradient can induce unbinding of topological defects, as observed during budding and tentacle formation in Hydra, and stabilize aster/vortex-like defects, as observed at a Hydra's mouth. On curved surfaces mimicking the morphologies of Hydra in various stages of development -- from spheroid to adult -- our model reproduces the experimentally observed reorganization of orientational order. Our results suggest how gradient alignment and curvature effects may work together to control orientational order during development and lays the foundations for future modeling efforts that will include the tissue mechanics that drive shape deformations., Comment: v2: Published version with minor revisions. v1: 12 pages, 7 figures (4 main, 3 suppl.). Movies: https://youtube.com/playlist?list=PLdejNN8-Lt6HJ52YfhyegpI0PGoRq8pSb

The polynomial kernels are widely used in machine learning and they are one of the default choices to develop kernel-based classification and regression models. However, they are rarely used and considered in numerical analysis due to their lack of strict positive definiteness. In particular they do not enjoy the usual property of unisolvency for arbitrary point sets, which is one of the key properties used to build kernel-based interpolation methods. This paper is devoted to establish some initial results for the study of these kernels, and their related interpolation algorithms, in the context of approximation theory. We will first prove necessary and sufficient conditions on point sets which guarantee the existence and uniqueness of an interpolant. We will then study the Reproducing Kernel Hilbert Spaces (or native spaces) of these kernels and their norms, and provide inclusion relations between spaces corresponding to different kernel parameters. With these spaces at hand, it will be further possible to derive generic error estimates which apply to sufficiently smooth functions, thus escaping the native space. Finally, we will show how to employ an efficient stable algorithm to these kernels to obtain accurate interpolants, and we will test them in some numerical experiment. After this analysis several computational and theoretical aspects remain open, and we will outline possible further research directions in a concluding section. This work builds some bridges between kernel and polynomial interpolation, two topics to which the authors, to different extents, have been introduced under the supervision or through the work of Stefano De Marchi. For this reason, they wish to dedicate this work to him in the occasion of his 60th birthday.

We study systematically the role of temperature in the optical response of doped two-dimensional semiconductors. By making use of a finite-temperature Fermi-polaron theory, we reveal a crossover from a quantum-degenerate regime with well-defined polaron quasiparticles to an incoherent regime at high temperature or low doping where the lowest energy "attractive" polaron quasiparticle is destroyed, becoming subsumed into a broad trion-hole continuum. We demonstrate that the crossover is accompanied by significant qualitative changes in both absorption and photoluminescence. In particular, with increasing temperature (or decreasing doping), the emission profile of the attractive branch evolves from a symmetric Lorentzian to an asymmetric peak with an exponential tail involving trions and recoil electrons at finite momentum. We discuss the effect of temperature on the coupling to light for structures embedded into a microcavity, and we show that there can exist well-defined polariton quasiparticles even when the exciton-polaron quasiparticle has been destroyed, where the transition from weak to strong light-matter coupling can be explained in terms of the polaron linewidths and spectral weights., Comment: 19 pages, 9 figures

We present a quantum virial expansion for the optical response of a doped two-dimensional semiconductor. As we show, this constitutes a perturbatively exact theory in the high-temperature or low-doping regime, where the electrons' thermal wavelength is smaller than their interparticle spacing. The virial expansion predicts new features of the photoluminescence, such as a non-trivial shape of the attractive branch related to universal low-energy exciton-electron scattering and an associated shift of the attractive peak from the trion energy. Our results are in excellent agreement with recent experiments on doped monolayer MoSe$_2$ [Zipfel et al., Phys. Rev. B 105, 075311 (2022)] and they imply that the trion binding energy is likely to have been overestimated in previous measurements. Our theory furthermore allows us to formally unify two distinct theoretical pictures that have been applied to this system, with the conventional trion picture results emerging as a high-temperature and weak-interaction limit of Fermi polaron theory., Comment: 7 pages, 2 figures (published version)

In this article, we propose a space spreading-assisted framework that leverages either time or frequency diversity or both to reduce interference and signal loss owing to channel impairments and facilitate the efficient operation of large-scale dense Internet-of-Things (IoT). Our approach employs dispersion of data-streams transmitted from individual IoT devices over indexed space-time (ST), space-frequency (SF) or space-time-frequency (STF) blocks. As a result, no two devices transmit on the same block; only one is activated while the rest of the devices in the network is silent, thereby minimizing possibility of interference on the transmit side. On the receive side, multiple-antenna array ameliorates performance in presence of channel impairments while exploiting array-processing gain. As interference due to superposition of multiple data-streams is killed at its root, no extra energy is wasted in fighting interference and other impairments, thereby enabling energy-efficient transmission from multiple devices over multiple access channel (MAC). To validate the proposed concept, we simulate the performance of the framework against dense IoT networks deployed in generalized indoor and outdoor scenarios in terms of probability of signal outage. Results demonstrate that our conceptualized framework benefits from interference-free transmission as well as enhancement in overall system performance.

Topological defects play a central role in the physics of many materials, including magnets, superconductors and liquid crystals. In active fluids, defects become autonomous particles that spontaneously propel from internal active stresses and drive chaotic flows stirring the fluid. The intimate connection between defect textures and active flow suggests that properties of active materials can be engineered by controlling defects, but design principles for their spatiotemporal control remain elusive. Here we propose a symmetry-based additive strategy for using elementary activity patterns, as active topological tweezers, to create, move and braid such defects. By combining theory and simulations, we demonstrate how, at the collective level, spatial activity gradients act like electric fields which, when strong enough, induce an inverted topological polarization of defects, akin to a negative susceptibility dielectric. We harness this feature in a dynamic setting to collectively pattern and transport interacting active defects. Our work establishes an additive framework to sculpt flows and manipulate active defects in both space and time, paving the way to design programmable active and living materials for transport, memory and logic., Comment: 11 pages (including Methods), 5 figures. Changed title and format, final version

The vertex model of epithelia describes the apical surface of a tissue as a tiling of polygonal cells, with a mechanical energy governed by deviations in cell shape from preferred, or target, area, $A_0$, and perimeter, $P_0$. The model exhibits a rigidity transition driven by geometric incompatibility as tuned by the target shape index, $p_0 = P_0 / \sqrt{A_0}$. For $p_0 > p_*(6) = \sqrt{8 \sqrt{3}} \approx 3.72$, with $p_*(6)$ the perimeter of a regular hexagon of unit area, a cell can simultaneously attain both the preferred area and preferred perimeter. As a result, the tissue is in a mechanically soft compatible state, with zero shear and Young's moduli. For $p_0 < p_*(6)$, it is geometrically impossible for any cell to realize the preferred area and perimeter simultaneously, and the tissue is in an incompatible rigid solid state. Using a mean-field approach, we present a complete analytical calculation of the linear elastic moduli of an ordered vertex model. We analyze a relaxation step that includes non-affine deformations, leading to a softer response than previously reported. The origin of the vanishing shear and Young's moduli in the compatible state is the presence of zero-energy deformations of cell shape. The bulk modulus exhibits a jump discontinuity at the transition and can be lower in the rigid state than in the fluid-like state. The Poisson's ratio can become negative which lowers the bulk and Young's moduli. Our work provides a unified treatment of linear elasticity for the vertex model and demonstrates that this linear response is protocol-dependent.

Controlling the continuum limit and extracting effective gravitational physics are shared challenges for quantum gravity approaches based on quantum discrete structures. The description of quantum gravity in terms of tensorial group field theory (TGFT) has recently led to much progress in its application to phenomenology, in particular cosmology. This application relies on the assumption of a phase transition to a nontrivial vacuum (condensate) state describable by mean-field theory, an assumption that is difficult to corroborate by a full renormalization group flow analysis due to the complexity of the relevant TGFT models. Here we demonstrate that this assumption is justified due to the specific ingredients of realistic quantum geometric TGFT models: combinatorially non-local interactions, matter degrees of freedom and Lorentz group data together with the encoding of micro-causality. This greatly strengthens the evidence for the existence of a meaningful continuum gravitational regime in group-field and spin-foam quantum gravity, the phenomenology of which is amenable to explicit computations in a mean-field approximation., Comment: Revised version in line with the publication in PRL

The study of post asymptotic giant branch (post-AGB) stars is a valuable tool to study still poorly known aspects of the evolution of the stars through the AGB. This is due to the accurate determination of their surface chemical composition and to the peculiar shape of the SED: the emission from the central star can be easily disentangled from the contribution from the dusty shell, which can then be characterized. The goal of the present study is to reconstruct the dust formation process and more generally the late phases of the evolution of O-rich stars across the AGB phase. This is performed by studying O-rich, post-AGB stars, which are analyzed in terms of their luminosity, Teff and infrared excess. We study sources classified as single, O-rich, post-AGB stars in the Galaxy, which exhibit a double-peaked (shell-type) SED. We use results from stellar evolution modelling combined with dust formation and radiative transfer modelling to reconstruct the late AGB phases and the initial contraction to the post-AGB phase. We also determine the mass-loss and dust formation rates for stars of different mass and chemical composition. The analysis of the IR excess of the post-AGB, O-rich stars examined in this study outlines an interesting complexity, in terms of the correlation between the dust in the surroundings of the stars, the evolutionary status and the progenitor's mass. The sources descending from massive AGBs (>3Msun depending on metallicity) are generally characterized by higher IR excess than the lower mass counterparts, owing to the more intense dust formation taking place during the final AGB phases. From the determination of the location of the dusty regions we deduce that the expanding velocities of the outflow change significantly from star to star. The possibility that radiation pressure is unable to accelerate the wind in the faintest objects is also commented., Comment: 11 pages, 6 figures and 1 table. Accepted for publication in A&A

We present deep 1.4 GHz source counts from $\sim$5 deg$^2$ of the continuum Early Science data release of the MeerKAT International Gigahertz Tiered Extragalactic Exploration (MIGHTEE) survey down to $S_{1.4\textrm{GHz}}\sim$15 $\mu$Jy. Using observations over two extragalactic fields (COSMOS and XMM-LSS), we provide a comprehensive investigation into correcting the incompleteness of the raw source counts within the survey to understand the true underlying source count population. We use a variety of simulations that account for: errors in source detection and characterisation, clustering, and variations in the assumed source model used to simulate sources within the field and characterise source count incompleteness. We present these deep source count distributions and use them to investigate the contribution of extragalactic sources to the sky background temperature at 1.4 GHz using a relatively large sky area. We then use the wealth of ancillary data covering{a subset of the COSMOS field to investigate the specific contributions from both active galactic nuclei (AGN) and star forming galaxies (SFGs) to the source counts and sky background temperature. We find, similar to previous deep studies, that we are unable to reconcile the sky temperature observed by the ARCADE 2 experiment. We show that AGN provide the majority contribution to the sky temperature contribution from radio sources, but the relative contribution of SFGs rises sharply below 1 mJy, reaching an approximate 15-25% contribution to the total sky background temperature ($T_b\sim$100 mK) at $\sim$15 $\mu$Jy., Comment: 24 pages, 12 figures; Accepted for publication in MNRAS

The ability of cells to reorganize in response to external stimuli is important in areas ranging from morphogenesis to tissue engineering. Elongated cells can co-align due to steric effects, forming states with local order. We show that molecular-scale substrate anisotropy can direct cell organization, resulting in the emergence of nematic order on tissue scales. To quantitatively examine the disorder-order transition, we developed a high-throughput imaging platform to analyze velocity and orientational correlations for several thousand cells over days. The establishment of global, seemingly long-ranged order is facilitated by enhanced cell division along the substrate's nematic axis, and associated extensile stresses that restructure the cells' actomyosin networks. Our work, which connects to a class of systems known as active dry nematics, provides a new understanding of the dynamics of cellular remodeling and organization in weakly interacting cell collectives. This enables data-driven discovery of cell-cell interactions and points to strategies for tissue engineering., Comment: 29 pages, 7 figures

Radio-frequency (RF) signal generators are standard laboratory equipment and a wide-range of open-source and commercial devices exists to address their many applications. Nonetheless, only few expensive and proprietary solutions can be re-configured within a wide frequency band and triggered on a micro-second timescale. Such specifications are required for applications that use variable radio-frequencies to generate programmed mixed signals, to control processes or states and to precisely steer laser beams using acousto-optical devices, tasks often needed in industrial manufacturing, atomic and molecular physics or microscopy. Here we present an open-source low-cost Arduino-based control system that can store up to millions of commands received from a computer and then perform reliable high-speed programming of an arbitrary device under its control (DUC) via a single- or quad-wire Serial Peripheral Interface. The software architecture operates as a real-time state machine, making it easily extensible and adaptable to any DUC. Each configuration change can be triggered either externally or internally, reaching ~1 MHz rates when using a Teensy 4.1 Arduino-compatible board. Leveraging this flexible system, we developed a programmable four-channel RF signal generator, based on an Analog Devices 9959 Evaluation board, and we demonstrated its capability and validated its performance.

We introduce a non-parametric density estimator deemed Radial Voronoi Density Estimator (RVDE). RVDE is grounded in the geometry of Voronoi tessellations and as such benefits from local geometric adaptiveness and broad convergence properties. Due to its radial definition RVDE is continuous and computable in linear time with respect to the dataset size. This amends for the main shortcomings of previously studied VDEs, which are highly discontinuous and computationally expensive. We provide a theoretical study of the modes of RVDE as well as an empirical investigation of its performance on high-dimensional data. Results show that RVDE outperforms other non-parametric density estimators, including recently introduced VDEs., Comment: 13 pages

The rheology of biological tissue is key to processes such as embryo development, wound healing and cancer metastasis. Vertex models of confluent tissue monolayers have uncovered a spontaneous liquid-solid transition tuned by cell shape; and a shear-induced solidification transition of an initially liquid-like tissue. Alongside this jamming/unjamming behaviour, biological tissue also displays an inherent viscoelasticity, with a slow time and rate dependent mechanics. With this motivation, we combine simulations and continuum theory to examine the rheology of the vertex model in nonlinear shear across a full range of shear rates from quastistatic to fast, elucidating its nonlinear stress-strain curves after the inception of shear of finite rate, and its steady state flow curves of stress as a function of strain rate. We formulate a rheological constitutive model that couples cell shape to flow and captures both the tissue solid-liquid transition and its rich linear and nonlinear rheology., Comment: 5 pages, 3 figures; plus supplemental material

Parkinson's disease (PD) is the second most common neurodegenerative disorder worldwide, yet there is no disease-modifying therapy up to this date. The biological complexity underlying PD hampers the investigation of the principal contributors to its pathogenesis. In this context, mechanistic models grounded in molecular-level knowledge provide virtual labs to uncover the primary events triggering PD onset and progression and suggest promising therapeutic targets. Multiple modeling efforts in PD research have focused on the pathological role of $\alpha$-synuclein ($\alpha$syn), a presynaptic protein that emerges from the intricate molecular network as a crucial driver of neurodegeneration. Here, we collect the advances in mathematical modeling of $\alpha$syn homeostasis, focusing on aggregation and degradation pathways, and discussing potential modeling improvements and possible implications in PD therapeutic strategy design.

Operational flare forecasting aims at providing predictions that can be used to make decisions, typically at a daily scale, about the space weather impacts of flare occurrence. This study shows that video-based deep learning can be used for operational purposes when the training and validation sets used for the network optimization are generated while accounting for the periodicity of the solar cycle. Specifically, the paper describes an algorithm that can be applied to build up sets of active regions that are balanced according to the flare class rates associated to a specific cycle phase. These sets are used to train and validate a Long-term Recurrent Convolutional Network made of a combination of a convolutional neural network and a Long-Short Memory network. The reliability of this approach is assessed in the case of two prediction windows containing the solar storm of March 2015 and September 2017, respectively.

In the tensorial group field theory (TGFT) approach to quantum gravity, the basic quanta of the theory correspond to discrete building blocks of geometry. It is expected that their collective dynamics gives rise to continuum spacetime at a coarse grained level, via a process involving a phase transition. In this work we show for the first time how phase transitions for realistic TGFT models can be realized using Landau-Ginzburg mean-field theory. More precisely, we consider models generating 4-dimensional Lorentzian triangulations formed by spacelike tetrahedra whose quantum geometry is encoded in non-local degrees of freedom on the non-compact group $\mathrm{SL}(2,\mathbb{C})$ and subject to gauge and simplicity constraints. Further we include $\mathbb{R}$-valued variables which may be interpreted as discretized scalar fields typically employed as a matter reference frame. We apply the Ginzburg criterion finding that fluctuations around the non-vanishing mean-field vacuum remain small at large correlation lengths regardless of the combinatorics of the non-local interaction validating the mean-field theory description of the phase transition. This work represents a first crucial step to understand phase transitions in compelling TGFT models for quantum gravity and paves the way for a more complete analysis via functional renormalization group techniques. Moreover, it supports the recent extraction of effective cosmological dynamics from TGFTs in the context of a mean-field approximation., Comment: 41 + 20 pages, 2 figures; matches published version

We study the dynamics of topological defects in active nematic films with spatially-varying activity and consider two setups: i) a constant activity gradient, and ii) a sharp jump in activity. A constant gradient of extensile (contractile) activity endows the comet-like $+1/2$ defect with a finite vorticity that drives the defect to align its nose in the direction of decreasing (increasing) gradient. A constant gradient does not, however, affect the known self-propulsion of the $+1/2$ defect and has no effect on the $-1/2$ that remains a non-motile particle. A sharp jump in activity acts like a wall that traps the defects, affecting the translational and rotational motion of both charges. The $+1/2$ defect slows down as it approaches the interface and the net vorticity tends to reorient the defect polarization so that it becomes perpendicular to the interface. The $-1/2$ defect acquires a self-propulsion towards the activity interface, while the vorticity-induced active torque tends to align the defect to a preferred orientation. This effective attraction of the negative defects to the wall is consistent with the observation of an accumulation of negative topological charge at both active/passive interfaces and physical boundaries., Comment: 14 pages, 8 figures

Demixing of binary liquids is a ubiquitous transition, which is explained using a well-established thermodynamic formalism that requires equality of intensive thermodynamics parameters across the phase boundaries. Demixing transitions also occur when binary fluid mixtures are driven away from equilibrium, for example, by external shear flow. Predicting demixing transition under non-equilibrium non-potential conditions remains, however, a challenge. We drive liquid-liquid phase separation (LLPS) of attractive DNA nanostar molecules away from equilibrium using an internally driven microtubule-based active fluid. Activity lowers the critical temperature and narrows the coexistence concentrations, but only when there are mechanical bonds between the liquid droplets and the reconfiguring active fluid. Similar behaviors are observed in numerical simulations, suggesting that activity suppression of liquid-liquid phase separation is a generic feature of active LLPS. Our work describes a platform for building soft active materials with feedback control while also providing insight into cell biology, where phase separation emerged as a ubiquitous self-organizational principle., Comment: 9 pages, 6 figures, supplementary information

Controlling interfaces of phase separating fluid mixtures is key to creating diverse functional soft materials. Traditionally, this is accomplished with surface-modifying chemical agents. Using experiment and theory, we study how mechanical activity shapes soft interfaces that separate an active and a passive fluid. Chaotic flows in the active fluid give rise to giant interfacial fluctuations and non-inertial propagating active waves. At high activities, stresses disrupt interface continuity and drive droplet generation, producing an emulsion-like active state comprised of finite-sized droplets. When in contact with a solid boundary, active interfaces exhibit non-equilibrium wetting transitions, wherein the fluid climbs the wall against gravity. These results demonstrate the promise of mechanically driven interfaces for creating a new class of soft active matter., Comment: 62 pages, 5 figures, 13 supplementary figures

The evolution of fifth generation (5G) networks needs to support the latest use cases, which demand robust network connectivity for the collaborative performance of the network agents, like multi-robot systems and vehicle to anything (V2X) communication. Unfortunately, the user device's limited communication range and battery constraint confirm the unfitness of known robustness metrics suggested for fixed networks, when applied to time-switching communication graphs. Furthermore, the calculation of most of the existing robustness metrics involves non-deterministic polynomial-time complexity, and hence are best-fitted only for small networks. Despite a large volume of works, the complete analysis of a $\textit{low-complexity}$ temporal robustness metric for a communication network is absent in the literature, and the present work aims to fill this gap. More in detail, our work provides a stochastic analysis of network robustness for a massive machine type communication (mMTC) network. The numerical investigation corroborates the exactness of the proposed analytical framework for temporal robustness metric. Along with studying the impact on network robustness of various system parameters, such as cluster head (CH) probability, power threshold value, network size, and node failure probability, we justify the observed trend of numerical results probabilistically.

The identification of AGN in large surveys has been hampered by seemingly discordant classifications arising from differing diagnostic methods, usually tracing distinct processes specific to a particular wavelength regime. However, as shown in Yao et al. (2020), the combination of optical emission line measurements and mid-infrared photometry can be used to optimise the discrimination capability between AGN and star formation activity. In this paper we test our new classification scheme by combining the existing GAMA-WISE data with high-quality MeerKAT radio continuum data covering 8 deg$^2$ of the GAMA G23 region. Using this sample of 1 841 galaxies (z < 0.25), we investigate the total infrared (derived from 12$\mu$m) to radio luminosity ratio, q(TIR), and its relationship to optical-infrared AGN and star-forming (SF) classifications. We find that while q(TIR) is efficient at detecting AGN activity in massive galaxies generally appearing quiescent in the infrared, it becomes less reliable for cases where the emission from star formation in the host galaxy is dominant. However, we find that the q(TIR) can identify up to 70 % more AGNs not discernible at optical and/or infrared wavelengths. The median q(TIR) of our SF sample is 2.57 $\pm$ 0.23 consistent with previous local universe estimates.

We study the nature of the faint radio source population detected in the MeerKAT International GHz Tiered Extragalactic Exploration (MIGHTEE) Early Science data in the COSMOS field, focusing on the properties of the radio-loud active galactic nuclei (AGN). Using the extensive multi-wavelength data available in the field, we are able to classify 88 per cent of the 5223 radio sources in the field with host galaxy identifications as AGN (35 per cent) or star-forming galaxies (54 per cent). We select a sample of radio-loud AGN with redshifts out to $z \sim 6$ and radio luminosities $10^{20} < \textrm{L}_{1.4~\textrm{GHz}} / \textrm{W Hz}^{-1} < 10^{27}$ and classify them as high-excitation and low-excitation radio galaxies (HERGs and LERGs). The classification catalogue is released with this work. We find no significant difference in the host galaxy properties of the HERGs and LERGs in our sample. In contrast to previous work, we find that the HERGs and LERGs have very similar Eddington-scaled accretion rates; in particular we identify a population of very slowly accreting AGN that are formally classified as HERGs at these low radio luminosities, where separating into HERGs and LERGs possibly becomes redundant. We investigate how black hole mass affects jet power, and find that a black hole mass $\gtrsim 10^{7.8}~\textrm{M}_\odot$ is required to power a jet with mechanical power greater than the radiative luminosity of the AGN ($L_\textrm{mech}/L_\textrm{bol} > 1$). We discuss that both a high black hole mass and black hole spin may be necessary to launch and sustain a dominant radio jet., Comment: Accepted for publication in MNRAS. The multi-wavelength source classification catalogue is released publicly with this work. 20 pages, 15 figures

Learning from previously collected datasets of expert data offers the promise of acquiring robotic policies without unsafe and costly online explorations. However, a major challenge is a distributional shift between the states in the training dataset and the ones visited by the learned policy at the test time. While prior works mainly studied the distribution shift caused by the policy during the offline training, the problem of recovering from out-of-distribution states at the deployment time is not very well studied yet. We alleviate the distributional shift at the deployment time by introducing a recovery policy that brings the agent back to the training manifold whenever it steps out of the in-distribution states, e.g., due to an external perturbation. The recovery policy relies on an approximation of the training data density and a learned equivariant mapping that maps visual observations into a latent space in which translations correspond to the robot actions. We demonstrate the effectiveness of the proposed method through several manipulation experiments on a real robotic platform. Our results show that the recovery policy enables the agent to complete tasks while the behavioral cloning alone fails because of the distributional shift problem.

Minimalist game design was introduced a decade ago as a general design principle with a list of key properties for minimalist games: basic controls, simple but aesthetically pleasing visuals, interesting player choices with vast possibility spaces, and sounds that resonate with the design. In this paper, we present an experiment we did to explore minimalism in games using a bottom-up approach. We invited a small group of professional game designers and a larger group of game design students to participate in a seminal experiment on minimalism in game design. We started from the most basic game elements: one pixel and one key which provide the least amount of information we can display and reasonably the most elementary action players can perform. We designed a game that starts with a black pixel and asks players to press a key when the pixel turns white. This minimal game, almost a Skinner box, captures the essential elements of the mechanics of games like "The Impossible Game," which asks players to do nothing more than press a key at the right moment. We presented this game concept to the professional game designers and challenged them to create other games with the least amount of player interaction and displayed information. We did not specify any constraints (as usually done in other contexts) and left them free to express their view of minimalistic game design. We repeated the experiment with 100+ students attending a master-level course on video game design and development at our institution. We then analyzed the creations of the two groups, discussing the idea of minimalistic design that emerges from the submitted game concepts.

We introduce a general method for learning representations that are equivariant to symmetries of data. Our central idea is to decompose the latent space into an invariant factor and the symmetry group itself. The components semantically correspond to intrinsic data classes and poses respectively. The learner is trained on a loss encouraging equivariance based on supervision from relative symmetry information. The approach is motivated by theoretical results from group theory and guarantees representations that are lossless, interpretable and disentangled. We provide an empirical investigation via experiments involving datasets with a variety of symmetries. Results show that our representations capture the geometry of data and outperform other equivariant representation learning frameworks., Comment: 12 pages

We present photometric metallicity measurements for a sample of 2.6 million bulge red clump stars extracted from the Blanco DECam Bulge Survey (BDBS). Similar to previous studies, we find that the bulge exhibits a strong vertical metallicity gradient, and that at least two peaks in the metallicity distribution functions appear at b < -5. We can discern a metal-poor ([Fe/H] ~ -0.3) and metal-rich ([Fe/H] ~ +0.2) abundance distribution that each show clear systematic trends with latitude, and may be best understood by changes in the bulge's star formation/enrichment processes. Both groups exhibit asymmetric tails, and as a result we argue that the proximity of a star to either peak in [Fe/H] space is not necessarily an affirmation of group membership. The metal-poor peak shifts to lower [Fe/H] values at larger distances from the plane while the metal-rich tail truncates. Close to the plane, the metal-rich tail appears broader along the minor axis than in off-axis fields. We also posit that the bulge has two metal-poor populations -- one that belongs to the metal-poor tail of the low latitude and predominantly metal-rich group, and another belonging to the metal-poor group that dominates in the outer bulge. We detect the X-shape structure in fields with |Z| > 0.7 kpc and for stars with [Fe/H] > -0.5. Stars with [Fe/H] < -0.5 may form a spheroidal or "thick bar" distribution while those with [Fe/H] > -0.1 are strongly concentrated near the plane., Comment: 26 pages, 22 figures, accepted for publication in MNRAS; the full data table is very long so only a stub table has been provided here; the full electronic table will be provided through MNRAS upon publication, but early access to the full table will be granted upon request to the authors

We study the dynamics of active nematic films on a substrate driven by active flows with or without the incompressible constraint.Through simulations and theoretical analysis, we show that arch patterns are stable in the compressible case, whilst they become unstable under the incompressibility constraint. For compressible flows at high enough activity, stable arches organize themselves into a smectic-like pattern, which induce an associated global polar ordering of $+1/2$ nematic defects. By contrast, divergence-free flows give rise to a local nematic order of the $+1/2$ defects, consisting of anti-aligned pairs of neighboring defects, as established in previous studies., Comment: 10 pages, 4 figures

The Blanco DECam Bulge Survey (BDBS) has imaged more than 200 square degrees of the southern Galactic bulge, providing photometry in the ugrizy filters for $\sim 250$ million unique stars. The presence of a strong foreground disk population, along with complex reddening and extreme image crowding, has made it difficult to constrain the presence of young and intermediate age stars in the bulge population. We employed an accurate cross-match of BDBS with the latest data release (EDR3) from the Gaia mission, matching more than 140 million sources with BDBS photometry and Gaia EDR3 photometry and astrometry. We relied on Gaia EDR3 astrometry, without any photometric selection, to produce clean BDBS bulge colour-magnitude diagrams (CMDs). Gaia parallaxes were used to filter out bright foreground sources, and a Gaussian mixture model fit to Galactic proper motions could identify stars kinematically consistent with bulge membership. We applied this method to 127 different bulge fields of $1$ deg$^2$ each, with $|\ell| \leq 9.5^\circ$ and $-9.5^\circ \leq b \leq -2.5^\circ$. The astrometric cleaning procedure removes the majority of blue stars in each field, especially near the Galactic plane, where the ratio of blue to red stars is $\lesssim 10\%$, increasing to values $\sim 20\%$ at higher Galactic latitudes. We rule out the presence of a widespread population of stars younger than 2 Gyr. The vast majority of blue stars brighter than the turnoff belong to the foreground population, according to their measured astrometry. We introduce the distance between the observed red giant branch bump and the red clump as a simple age proxy for the dominant population in the field, and we confirm the picture of a predominantly old bulge. Further work is needed to apply the method to estimate ages to fields at higher latitudes, and to model the complex morphology of the Galactic bulge., Comment: 21 pages, 22 figures, accepted for publication in A&A

We use a continuum model to examine the effect of activity on a phase separating mixture of an extensile active nematic and a passive fluid. We highlight the distinct role of previously considered interfacial active stresses and bulk active stresses that couple to liquid crystalline degrees of freedom. Interfacial active stresses can arrest phase separation, as previously demonstrated. Bulk extensile active stresses can additionally strongly suppress phase separation by sustained self-stirring of the fluid, substantially reducing the size of the coexistence region in the temperature/concentration plane relative to that of the passive system. The phase separated state is a dynamical emulsion-like steady state of continuously splitting and merging droplets, as suggested by recent experiments. Using scaling analysis and simulations, we identify various regimes for the dependence of droplet size on activity. These results can provide a criterion for identifying the mechanisms responsible for arresting phase separation in experiments.

We introduce an algorithm for active function approximation based on nearest neighbor regression. Our Active Nearest Neighbor Regressor (ANNR) relies on the Voronoi-Delaunay framework from computational geometry to subdivide the space into cells with constant estimated function value and select novel query points in a way that takes the geometry of the function graph into account. We consider the recent state-of-the-art active function approximator called DEFER, which is based on incremental rectangular partitioning of the space, as the main baseline. The ANNR addresses a number of limitations that arise from the space subdivision strategy used in DEFER. We provide a computationally efficient implementation of our method, as well as theoretical halting guarantees. Empirical results show that ANNR outperforms the baseline for both closed-form functions and real-world examples, such as gravitational wave parameter inference and exploration of the latent space of a generative model., Comment: Accepted at the International Conference on Machine Learning (ICML) 2022

The Voronoi Density Estimator (VDE) is an established density estimation technique that adapts to the local geometry of data. However, its applicability has been so far limited to problems in two and three dimensions. This is because Voronoi cells rapidly increase in complexity as dimensions grow, making the necessary explicit computations infeasible. We define a variant of the VDE deemed Compactified Voronoi Density Estimator (CVDE), suitable for higher dimensions. We propose computationally efficient algorithms for numerical approximation of the CVDE and formally prove convergence of the estimated density to the original one. We implement and empirically validate the CVDE through a comparison with the Kernel Density Estimator (KDE). Our results indicate that the CVDE outperforms the KDE on sound and image data., Comment: Accepted at the Conference on Uncertainty in Artificial Intelligence (UAI) 2022. This version contains erratas of the published material

The third data release (DR3) of the European Space Agency satellite Gaia provides coordinates, parallaxes, proper motions, and radial velocities for a sample of $\sim 34$ million stars. We use the combined 6-dimensional phase space information to search for hypervelocity stars (HVSs), unbound stars accelerated by dynamical processes happening in the Galactic Centre. By looking at the kinematics of Gaia DR3 stars in Galactocentric coordinates and by integrating their orbits in the Galactic potential, we do not identify any HVS candidates with a velocity higher than $700$ km s$^{-1}$ and robustly observed kinematics. Assuming a scenario wherein the interaction between a stellar binary and the massive black hole Sgr A$^*$ is responsible for HVS ejections from the Galactic Centre, we derive degenerate limits on the ejection rate of HVSs and the slope of the initial mass function of the primary star among binaries in the Galactic Centre. Our results indicate that the HVS ejection rate is $\lesssim 8\times10^{-5}$ yr$^{-1}$ assuming a Salpeter mass function, and this upper limit becomes progressively smaller for an increasingly top-heavy mass distribution. A fiducial HVS ejection rate of $10^{-4}$ yr$^{-1}$ prefers a mass function slope $\lesssim -2.35$, disfavouring previously claimed top-heavy initial mass functions among stars in the Galactic Centre., Comment: 8 pages, 3 figures, accepted for publication in MNRAS

We consider couplings of electrically and magnetically charged sources to the maximally symmetric non-linear extension of Maxwell's theory called ModMax. The aim is to reveal physical effects which distinguish ModMax from Maxwell's electrodynamics. We find that, in contrast to generic models of non-linear electrodynamics, Lienard-Wiechert fields induced by a moving electric or magnetic particle, or a dyon are exact solutions of the ModMax equations of motion. We then study whether and how ModMax non-linearity affects properties of electromagnetic interactions of charged objects, in particular the Lorentz force, the Coulomb law, the Lienard-Wiechert fields, Dirac's and Schwinger's quantization of electric and magnetic charges, and the Compton Effect. In passing we also present an alternative form of the ModMax Lagrangian in terms of the coupling of Maxwell's theory to axion-dilaton-like auxiliary scalar fields which may be relevant for revealing the effective field theory origin of ModMax., Comment: 20 pages, v2: title changed to match the PRD version, typos corrected, references updated