Search

Extended Dynamic Mode Decomposition (eDMD) is a powerful tool to generate data-driven surrogate models for the prediction and control of nonlinear dynamical systems in the Koopman framework. In eDMD a compression of the lifted system dynamics on the space spanned by finitely many observables is computed, in which the original space is embedded as a low-dimensional manifold. While this manifold is invariant for the infinite-dimensional Koopman operator, this invariance is typically not preserved for its eDMD-based approximation. Hence, an additional (re-)projection step is often tacitly incorporated to improve the prediction capability. We propose a novel framework for consistent reprojectors respecting the underlying manifold structure. Further, we present a new geometric reprojector based on maximum-likelihood arguments, which significantly enhances the approximation accuracy and preserves known finite-data error bounds., Comment: 15 pages, 7 figures

FRB 20210912A is a fast radio burst (FRB), detected and localised to sub-arcsecond precision by the Australian Square Kilometre Array Pathfinder. No host galaxy has been identified for this burst despite the high precision of its localisation and deep optical and infrared follow-up, to 5-$\sigma$ limits of $R=26.7$ mag and $K_\mathrm{s}=24.9$ mag with the Very Large Telescope. The combination of precise radio localisation and deep optical imaging has almost always resulted in the secure identification of a host galaxy, and this is the first case in which the line-of-sight is not obscured by the Galactic disk. The dispersion measure of this burst, $\mathrm{DM_{FRB}}=1233.696\pm0.006~\mathrm{pc}\ \mathrm{cm}^{-3}$, allows for a large source redshift of $z>1$ according to the Macquart relation. It could thus be that the host galaxy is consistent with the known population of FRB hosts, but is too distant to detect in our observations ($z>0.7$ for a host like that of the first repeating FRB source, FRB 20121102A); that it is more nearby with a significant excess in $\mathrm{DM_{host}}$, and thus dimmer than any known FRB host; or, least likely, that the FRB is truly hostless. We consider each possibility, making use of the population of known FRB hosts to frame each scenario. The fact of the missing host has ramifications for the FRB field: even with high-precision localisation and deep follow-up, some FRB hosts may be difficult to detect, with more distant hosts being the less likely to be found. This has implications for FRB cosmology, in which high-redshift detections are valuable., Comment: 14 pages, 6 figures. Revised based on referee's comments and accepted to MNRAS

Event-based cameras are popular for tracking fast-moving objects due to their high temporal resolution, low latency, and high dynamic range. In this paper, we propose a novel algorithm for tracking event blobs using raw events asynchronously in real time. We introduce the concept of an event blob as a spatio-temporal likelihood of event occurrence where the conditional spatial likelihood is blob-like. Many real-world objects such as car headlights or any quickly moving foreground objects generate event blob data. The proposed algorithm uses a nearest neighbour classifier with a dynamic threshold criteria for data association coupled with an extended Kalman filter to track the event blob state. Our algorithm achieves highly accurate blob tracking, velocity estimation, and shape estimation even under challenging lighting conditions and high-speed motions (> 11000 pixels/s). The microsecond time resolution achieved means that the filter output can be used to derive secondary information such as time-to-contact or range estimation, that will enable applications to real-world problems such as collision avoidance in autonomous driving., Comment: 18 pages, 16 figures. The manuscript was accepted on August 7, 2024, by IEEE Transactions on Robotics

We present high-resolution 1.5 $-$ 6 GHz Karl G. Jansky Very Large Array (VLA) and Hubble Space Telescope (HST) optical and infrared observations of the extremely active repeating fast radio burst (FRB) FRB 20201124A and its barred spiral host galaxy. We constrain the location and morphology of star formation in the host and search for a persistent radio source (PRS) coincident with FRB 20201124A. We resolve the morphology of the radio emission across all frequency bands and measure a star formation rate SFR $\approx 8.9\,M_{\odot}$ yr$^{-1}$, approximately $\approx 2.5-6$ times larger than optically-inferred SFRs, demonstrating dust-obscured star formation throughout the host. Compared to a sample of all known FRB hosts with radio emission, the host of FRB 20201124A has the most significantly obscured star formation. While HST observations show the FRB to be offset from the bar or spiral arms, the radio emission extends to the FRB location. We propose that the FRB progenitor could have formed in situ (e.g., a magnetar born from a massive star explosion). It is still plausible, although less likely, that the progenitor of FRB 20201124A migrated from the central bar of the host. We further place a limit on the luminosity of a putative PRS at the FRB position of $L_{\rm 6.0 \ GHz}$ $\lesssim$ 1.8 $\times 10^{27}$ erg s$^{-1}$ Hz$^{-1}$, among the deepest PRS luminosity limits to date. However, this limit is still broadly consistent with both magnetar nebulae and hypernebulae models assuming a constant energy injection rate of the magnetar and an age of $\gtrsim 10^{5}$ yr in each model, respectively., Comment: 21 pages, 6 figures, 3 tables, Accepted to ApJ; doi:10.3847/1538-4357/ad0cbd

Interpreting the inner workings of deep learning models is crucial for establishing trust and ensuring model safety. Concept-based explanations have emerged as a superior approach that is more interpretable than feature attribution estimates such as pixel saliency. However, defining the concepts for the interpretability analysis biases the explanations by the user's expectations on the concepts. To address this, we propose a novel post-hoc unsupervised method that automatically uncovers the concepts learned by deep models during training. By decomposing the latent space of a layer in singular vectors and refining them by unsupervised clustering, we uncover concept vectors aligned with directions of high variance that are relevant to the model prediction, and that point to semantically distinct concepts. Our extensive experiments reveal that the majority of our concepts are readily understandable to humans, exhibit coherency, and bear relevance to the task at hand. Moreover, we showcase the practical utility of our method in dataset exploration, where our concept vectors successfully identify outlier training samples affected by various confounding factors. This novel exploration technique has remarkable versatility to data types and model architectures and it will facilitate the identification of biases and the discovery of sources of error within training data.

The putative host galaxy of FRB 20171020A was first identified as ESO 601-G036 in 2018, but as no repeat bursts have been detected, direct confirmation of the host remains elusive. In light of recent developments in the field, we re-examine this host and determine a new association confidence level of 98%. At 37 Mpc, this makes ESO 601-G036 the third closest FRB host galaxy to be identified to date and the closest to host an apparently non-repeating FRB (with an estimated repetition rate limit of < 0.011 bursts per day above 10 erg). Due to its close distance, we are able to perform detailed multi-wavelength analysis on the ESO 601-G036 system. Follow-up observations confirm ESO 601-G036 to be a typical star-forming galaxy with HI and stellar masses of log(M_HI/M_sol) ~ 9.2 and log(M_*/M_sol) = 8.64, and a star formation rate of SFR = 0.09 +/- 0.01 M_sol/yr. We detect, for the first time, a diffuse gaseous tail (log(M_HI/M_sol) ~ 8.3) extending to the south-west that suggests recent interactions, likely with the confirmed nearby companion ESO 601-G037. ESO 601-G037 is a stellar shred located to the south of ESO 601-G036 that has an arc-like morphology, is about an order of magnitude less massive, and has a lower gas metallicity that is indicative of a younger stellar population. The properties of the ESO 601-G036 system indicate an ongoing minor merger event, which is affecting the overall gaseous component of the system and the stars within ESO 601-G037. Such activity is consistent with current FRB progenitor models involving magnetars and the signs of recent interactions in other nearby FRB host galaxies., Comment: 16 pages, 9 figures, accepted for publication in PASA

We present a joint cosmic shear analysis of the Dark Energy Survey (DES Y3) and the Kilo-Degree Survey (KiDS-1000) in a collaborative effort between the two survey teams. We find consistent cosmological parameter constraints between DES Y3 and KiDS-1000 which, when combined in a joint-survey analysis, constrain the parameter $S_8 = \sigma_8 \sqrt{\Omega_{\rm m}/0.3}$ with a mean value of $0.790^{+0.018}_{-0.014}$. The mean marginal is lower than the maximum a posteriori estimate, $S_8=0.801$, owing to skewness in the marginal distribution and projection effects in the multi-dimensional parameter space. Our results are consistent with $S_8$ constraints from observations of the cosmic microwave background by Planck, with agreement at the $1.7\sigma$ level. We use a Hybrid analysis pipeline, defined from a mock survey study quantifying the impact of the different analysis choices originally adopted by each survey team. We review intrinsic alignment models, baryon feedback mitigation strategies, priors, samplers and models of the non-linear matter power spectrum., Comment: 40 pages, 21 figures, 15 tables, accepted Open Journal of Astrophysics. Download the chains from https://des.ncsa.illinois.edu/releases/y3a2/Y3key-joint-des-kids or create your own chains with CosmoSIS using https://github.com/joezuntz/cosmosis-standard-library/blob/main/examples/des-y3_and_kids-1000.ini Watch the core team discuss this analysis at https://cosmologytalks.com/2023/05/26/des-kids

The current transformation towards smart manufacturing has led to a growing demand for human-robot collaboration (HRC) in the manufacturing process. Perceiving and understanding the human co-worker's behaviour introduces challenges for collaborative robots to efficiently and effectively perform tasks in unstructured and dynamic environments. Integrating recent data-driven machine vision capabilities into HRC systems is a logical next step in addressing these challenges. However, in these cases, off-the-shelf components struggle due to generalisation limitations. Real-world evaluation is required in order to fully appreciate the maturity and robustness of these approaches. Furthermore, understanding the pure-vision aspects is a crucial first step before combining multiple modalities in order to understand the limitations. In this paper, we propose GoferBot, a novel vision-based semantic HRC system for a real-world assembly task. It is composed of a visual servoing module that reaches and grasps assembly parts in an unstructured multi-instance and dynamic environment, an action recognition module that performs human action prediction for implicit communication, and a visual handover module that uses the perceptual understanding of human behaviour to produce an intuitive and efficient collaborative assembly experience. GoferBot is a novel assembly system that seamlessly integrates all sub-modules by utilising implicit semantic information purely from visual perception.

Feature-based homography estimation approaches rely on extensive image processing for feature extraction and matching, and do not adequately account for the information provided by the image. Therefore, developing efficient direct techniques to extract the homography from images is essential. This paper presents a novel nonlinear direct homography observer that exploits the Lie group structure of $\mathbf{SL}(3)$ and its action on the space of image maps. Theoretical analysis demonstrates local asymptotic convergence of the observer. The observer design is also extended for partial measurements of velocity under the assumption that the unknown component is constant or slowly time-varying. Finally, simulation results demonstrate the performance of the proposed solutions on real images., Comment: 11 pages, 3 figures, to appear in Proceedings of IFAC World Congress 2023

We present new upgraded Giant Metrewave Radio Telescope (uGMRT) HI 21-cm observations of the ultra-luminous infrared galaxy IRAS 10565+2448, previously reported to show blueshifted, broad, and shallow HI absorption indicating an outflow. Our higher spatial resolution observations have localised this blueshifted outflow, which is $\sim$ 1.36 kpc southwest of the radio centre and has a blueshifted velocity of $\sim 148\,\rm km\,s^{-1}$ and a full width at half maximum (FWHM) of $\sim 581\,\rm km\,s^{-1}$. The spatial extent and kinematic properties of the HI outflow are consistent with the previously detected cold molecular outflows in IRAS 10565+2448, suggesting that they likely have the same driving mechanism and are tracing the same outflow. By combining the multi-phase gas observations, we estimate a total outflowing mass rate of at least $140\, \rm M_\odot \,yr^{-1}$ and a total energy loss rate of at least $8.9\times10^{42}\,\rm erg\,s^{-1}$, where the contribution from the ionised outflow is negligible, emphasising the importance of including both cold neutral and molecular gas when quantifying the impact of outflows. We present evidence of the presence of a radio jet and argue that this may play a role in driving the observed outflows. The modest radio luminosity $L_{\rm1.4GHz}$ $\sim1.3\times10^{23}\,{\rm W\,Hz^{-1}}$ of the jet in IRAS 10565+2448 implies that the jet contribution to driving outflows should not be ignored in low radio luminosity AGN., Comment: 12 pages, 9 figures, accepted for publication in MNRAS

Gyres are central features of large-scale ocean circulation and are involved in transporting tracers such as heat, nutrients, and carbon-dioxide within and across ocean basins. Traditionally, the gyre circulation is thought to be driven by surface winds and quantified via Sverdrup balance, but it has been proposed that surface buoyancy fluxes may also contribute to gyre forcing. Through a series of eddy-permitting global ocean model simulations with perturbed surface forcing, the relative contribution of wind stress and surface heat flux forcing to the large-scale ocean circulation is investigated, focusing on the subtropical gyres. In addition to gyre strength being linearly proportional to wind stress, it is shown that the gyre circulation is strongly impacted by variations in the surface heat flux (specifically, its meridional gradient) through a rearrangement of the ocean's buoyancy structure. On shorter timescales ($\sim$ decade), the gyre circulation anomalies are proportional to the magnitude of the surface heat flux gradient perturbation, with up to $\sim 0.15\,\mathrm{Sv}$ anomaly induced per $\mathrm{W}\,\mathrm{m}^{-2}$ change in the surface heat flux. On timescales longer than a decade, the gyre response to surface buoyancy flux gradient perturbations becomes non-linear as ocean circulation anomalies feed back onto the buoyancy structure induced by the surface buoyancy fluxes. These interactions complicate the development of a buoyancy-driven theory for the gyres to complement the Sverdrup relation. The flux-forced simulations underscore the importance of surface buoyancy forcing in steering the large-scale ocean circulation., Comment: Submitted to the Journal of Physical Oceanography

Accurate labeling of phase connectivity in electrical distribution systems is important for maintenance and operations but is often erroneous or missing. In this paper, we present a process to identify which smart meters must be in the same phase using a hierarchical clustering method on voltage time series data. Instead of working with the time series data directly, we apply the Fourier transform to represent the data in their frequency domain, remove $98\%$ of the Fourier coefficients, and use the remaining coefficients to cluster the meters are in the same phase. Result of this process is validated by confirming that cluster (phase) membership of meters does not change over two monthly periods. In addition, we also confirm that meters that belong to the same feeder within the distribution network are correctly classified into the same cluster, that is, assigned to the same phase., Comment: 5 pages, 6 figures, 4 tables

We present WALLABY pilot data release 1, the first public release of HI pilot survey data from the Wide-field ASKAP L-band Legacy All-sky Blind Survey (WALLABY) on the Australian Square Kilometre Array Pathfinder. Phase 1 of the WALLABY pilot survey targeted three $60~{\rm deg}^2$ regions on the sky in the direction of the Hydra and Norma galaxy clusters and the NGC 4636 galaxy group, covering the redshift range of z < 0.08. The source catalogue, images and spectra of nearly 600 extragalactic HI detections and kinematic models for 109 spatially resolved galaxies are available. As the pilot survey targeted regions containing nearby group and cluster environments, the median redshift of the sample of z ~ 0.014 is relatively low compared to the full WALLABY survey. The median galaxy HI mass is $2.3 \times 10^{9}~M_{\odot}$. The target noise level of 1.6 mJy per $30''$ beam and 18.5 kHz channel translates into a $5\sigma$ HI mass sensitivity for point sources of about $5.2 \times 10^{8} \, (D_{\rm L} / \mathrm{100~Mpc})^{2} \, M_{\odot}$ across 50 spectral channels (~200 km/s) and a $5\sigma$ HI column density sensitivity of about $8.6 \times 10^{19} \, (1 + z)^{4}~\mathrm{cm}^{-2}$ across 5 channels (~20 km/s) for emission filling the $30''$ beam. As expected for a pilot survey, several technical issues and artefacts are still affecting the data quality. Most notably, there are systematic flux errors of up to several 10% caused by uncertainties about the exact size and shape of each of the primary beams as well as the presence of sidelobes due to the finite deconvolution threshold. In addition, artefacts such as residual continuum emission and bandpass ripples have affected some of the data. The pilot survey has been highly successful in uncovering such technical problems, most of which are expected to be addressed and rectified before the start of the full WALLABY survey., Comment: 24 pages, 13 figures, 5 tables, accepted for publication in PASA

Accurate tracking of planned trajectories in the presence of perturbations is an important problem in control and robotics. Symmetry is a fundamental mathematical feature of many dynamical systems and exploiting this property offers the potential of improved tracking performance. In this paper, we investigate the tracking problem for systems on homogeneous spaces, manifolds which admit symmetries with transitive group actions. We show that there is natural manner to lift any desired trajectory of such a system to a lifted trajectory on the symmetry group. This construction allows us to define a global tracking error and apply LQR design to obtain an approximately optimal control in a single coordinate chart. The resulting control is then applied to the original plant and shown to yield excellent tracking performance. We term the resulting design methodology the Equivariant Regulator (EqR). We provide an example system posed on a homogeneous space, derive the trajectory linearisation in error coordinates and demonstrate the effectiveness of EqR compared to standard approaches in simulation.

Observers for systems with Lie group symmetries are an active area of research that is seeing significant impact in a number of practical domains, including aerospace, robotics, and mechatronics. This paper builds on the theory of the recently proposed Equivariant Filter (EqF), which is a general observer design for systems on homogeneous spaces that takes advantage of symmetries to yield significant performance advantages. It is shown that the EqF error dynamics are invariant to transformation of the input signal and equivariant as a parametrised vector field. The main theorem shows that two EqF's with different choices of local coordinates and origins and with equivalent noise modelling yield identical performance. In other words, the EqF is intrinsic to the system equations and symmetry. This is verified in a simulation of a 2D robot localisation problem, which also shows how the ability to choose an origin for the EqF can yield practical performance advantages by mitigating floating point precision errors., Comment: 11 pages, 2 figures, accepted for publication in IFAC NOLCOS 2022

We present early science results from Deep Investigation of Neutral Gas Origins (DINGO), an HI survey using the Australian Square Kilometre Array Pathfinder (ASKAP). Using ASKAP sub-arrays available during its commissioning phase, DINGO early science data were taken over $\sim$ 60 deg$^{2}$ of the Galaxy And Mass Assembly (GAMA) 23 h region with 35.5 hr integration time. We make direct detections of six known and one new sources at $z < 0.01$. Using HI spectral stacking, we investigate the HI gas content of galaxies at $0.04 < z< 0.09$ for different galaxy colours. The results show that galaxy morphology based on optical colour is strongly linked to HI gas properties. To examine environmental impacts on the HI gas content of galaxies, three sub-samples are made based on the GAMA group catalogue. The average HI mass of group central galaxies is larger than those of satellite and isolated galaxies, but with a lower HI gas fraction. We derive a variety of HI scaling relations for physical properties of our sample, including stellar mass, stellar mass surface density, $NUV-r$ colour, specific star formation rate, and halo mass. We find that the derived HI scaling relations are comparable to other published results, with consistent trends also observed to $\sim$0.5 dex lower limits in stellar mass and stellar surface density. The cosmic HI densities derived from our data are consistent with other published values at similar redshifts. DINGO early science highlights the power of HI spectral stacking techniques with ASKAP., Comment: 27 pages, 25 figures, 10 tables, accepted for publication in MNRAS

We present constraints on the flat $\Lambda$CDM cosmological model through a joint analysis of galaxy abundance, galaxy clustering and galaxy-galaxy lensing observables with the Kilo-Degree Survey. Our theoretical model combines a flexible conditional stellar mass function, to describe the galaxy-halo connection, with a cosmological N-body simulation-calibrated halo model to describe the non-linear matter field. Our magnitude-limited bright galaxy sample combines 9-band optical-to-near-infrared photometry with an extensive and complete spectroscopic training sample to provide accurate redshift and stellar mass estimates. Our faint galaxy sample provides a background of accurately calibrated lensing measurements. We constrain the structure growth parameter $S_8=\sigma_8\sqrt{\Omega_{\mathrm{m}}/0.3}=0.773^{+0.028}_{-0.030}$, and the matter density parameter $\Omega_{\mathrm{m}}=0.290^{+0.021}_{-0.017}$. The galaxy-halo connection model adopted in the work is shown to be in agreement with previous studies. Our constraints on cosmological parameters are comparable to, and consistent with, joint $3\times2{\mathrm{pt}}$ clustering-lensing analyses that additionally include a cosmic shear observable. This analysis therefore brings attention to the significant constraining power in the often-excluded non-linear scales for galaxy clustering and galaxy-galaxy lensing observables. By adopting a theoretical model that accounts for non-linear halo bias, halo exclusion, scale-dependent galaxy bias and the impact of baryon feedback, this work demonstrates the potential and a way forward to include non-linear scales in cosmological analyses. Varying the width of the satellite galaxy distribution with an additional parameter yields a strong preference for sub-Poissonian variance, improving the goodness of fit by 0.18 in reduced $\chi^{2}$ value compared to a fixed Poisson distribution., Comment: 25 pages, 16 figures, accepted for publication in A&A, v3 with updated figure from Corrigendum (https://www.aanda.org/articles/aa/full_html/2024/08/aa50702e-24/aa50702e-24.html, https://doi.org/10.1051/0004-6361/202450702e)

Stochastic filters for on-line state estimation are a core technology for autonomous systems. The performance of such filters is one of the key limiting factors to a system's capability. Both asymptotic behavior (e.g.,~for regular operation) and transient response (e.g.,~for fast initialization and reset) of such filters are of crucial importance in guaranteeing robust operation of autonomous systems. This paper introduces a new generic formulation for a gyroscope aided attitude estimator using N direction measurements including both body-frame and reference-frame direction type measurements. The approach is based on an integrated state formulation that incorporates navigation, extrinsic calibration for all direction sensors, and gyroscope bias states in a single equivariant geometric structure. This newly proposed symmetry allows modular addition of different direction measurements and their extrinsic calibration while maintaining the ability to include bias states in the same symmetry. The subsequently proposed filter-based estimator using this symmetry noticeably improves the transient response, and the asymptotic bias and extrinsic calibration estimation compared to state-of-the-art approaches. The estimator is verified in statistically representative simulations and is tested in real-world experiments., Comment: to be published in Robotics and Automation Letters

The kinematics of many nonlinear control systems, especially in the robotics field, admit a transitive Lie-group symmetry, which is useful in high performance observer design. The recently proposed equivariant filter (EqF) exploits equivariance to generate high performance filters for a wide range of real-world systems. However, existing work on the equivariant filter, and equivariance of control systems in general, is based on a continuous-time formulation. In this paper, we first present the equivariant structure of a discrete-time system. We then use this to propose a discrete-time version of the equivariant filter. A novelty of the approach is that the geometry of the symmetry group naturally appears as parallel transport in the reset step of the filter. Preliminary results for linear second order kinematics with separate bearing and range measurements indicate that the discrete EqF significantly outperforms both a discretized version of the continuous EqF and a classical discrete EKF., Comment: 13 pages, 2 figures. Accepted for presentation in 2022 61st IEEE Conference on Decision and Control (CDC), IEEE, 2022

Inertial Velocity-Aided Attitude (VAA) is an important problem in the control of Remotely Piloted Aerial Systems (RPAS), and involves estimating the velocity and attitude of a vehicle using gyroscope, accelerometer, and inertial-frame velocity (e.g. GPS velocity) measurements. Existing solutions tend to be complex and provide limited stability guarantees, relying on either high gain designs or assuming constant acceleration of the vehicle. This paper proposes a novel observer for inertial VAA that exploits Lie group symmetries of the system dynamics, and shows that the observer is synchronous with the system trajectories. This is achieved by adding a virtual state of only three dimensions, in contrast to the larger virtual states typically used in the literature. The error dynamics of the observer are shown to be almost globally asymptotically and locally exponentially stable. Finally, the observer is verified in simulation, where it is shown that the estimation error converges to zero even with an extremely poor initial condition., Comment: 11 pages, 2 figures, submitted to NOLCOS 2022

The deposition of dielectric materials on graphene is one of the bottlenecks for unlocking the potential of graphene in electronic applications. In this paper we demonstrate the plasma enhanced atomic layer deposition of 10 nm thin high quality Al$_2$O$_3$ on graphene using a monolayer of hBN as protection layer. Raman spectroscopy was performed to analyze possible structural changes of the graphene lattice caused by the plasma deposition. The results show that a monolayer of hBN in combination with an optimized deposition process can effectively protect graphene from damage, while significant damage was observed without an hBN layer. Electrical characterization of double gated graphene field effect devices confirms that the graphene did not degrade during the plasma deposition of Al$_2$O$_3$. The leakage current densities were consistently below 1 nA/mm for electric fields across the insulators of up to 8 MV/cm, with irreversible breakdown happening above. Such breakdown electric fields are typical for Al$_2$O$_3$ and can be seen as an indicator for high quality dielectric films.

Fine-Grained Change Detection and Regression Analysis are essential in many applications of ArtificialIntelligence. In practice, this task is often challenging owing to the lack of reliable ground truth information andcomplexity arising from interactions between the many underlying factors affecting a system. Therefore,developing a framework which can represent the relatedness and reliability of multiple sources of informationbecomes critical. In this paper, we investigate how techniques in multi-task metric learning can be applied for theregression of fine-grained change in real data.The key idea is that if we incorporate the incremental change in a metric of interest between specific instancesof an individual object as one of the tasks in a multi-task metric learning framework, then interpreting thatdimension will allow the user to be alerted to fine-grained change invariant to what the overall metric isgeneralised to be. The techniques investigated are specifically tailored for handling heterogeneous data sources,i.e. the input data for each of the tasks might contain missing values, the scale and resolution of the values is notconsistent across tasks and the data contains non-independent and identically distributed (non-IID) instances. Wepresent the results of our initial experimental implementations of this idea and discuss related research in thisdomain which may offer direction for further research.

We present the results of a search for associated 21 cm HI absorption at redshift 0.42 < z < 1.00 in radio-loud galaxies from three Galaxy And Mass Assembly (GAMA) survey fields. These observations were carried out as part of a pilot survey for the ASKAP First Large Absorption Survey in HI (FLASH). From a sample of 326 radio sources with 855.5 MHz peak flux density above 10 mJy we detected two associated HI absorption systems, in SDSS J090331+010847 at z= 0.522 and SDSS J113622+004852 at z= 0.563. Both galaxies are massive (stellar mass > 10$^{11}$ M$_{sun}$) and have optical spectra characteristic of luminous red galaxies,though SED fitting implies that SDSS J113622+004852 contains a dust-obscured starburst with SFR ~ 69 M$_{sun}$ yr$^{-1}$. The HI absorption lines have a high optical depth, with $\tau_{pk}$ of 1.77 $\pm$ 0.16 for SDSS J090331+010847 (the highest value for any z > 0.1 associated system found to date) and 0.14 $\pm$ 0.01 for SDSS J113622+004852. In the redshift range probed by our ASKAP observations, the detection rate for associated HI absorption lines (with $\tau_{pk}$ > 0.1 and at least 3$\sigma$ significance) is 2.9 (+9.7 -2.6) percent. Although the current sample is small, this rate is consistent with a trend seen in other studies for a lower detection rate of associated 21 cm HI absorption systems at higher redshift. We also searched for OH absorption lines at 0.67 < z < 1.34, but no detection was made in the 145 radio sources searched., Comment: 26 pages, 15 figures, accepted for publication in MNRAS

Event cameras are bio-inspired dynamic vision sensors that respond to changes in image intensity with a high temporal resolution, high dynamic range and low latency. These sensor characteristics are ideally suited to enable visual target tracking in concert with a broadcast visual communication channel for smart visual beacons with applications in distributed robotics. Visual beacons can be constructed by high-frequency modulation of Light Emitting Diodes (LEDs) such as vehicle headlights, Internet of Things (IoT) LEDs, smart building lights, etc., that are already present in many real-world scenarios. The high temporal resolution characteristic of the event cameras allows them to capture visual signals at far higher data rates compared to classical frame-based cameras. In this paper, we propose a novel smart visual beacon architecture with both LED modulation and event camera demodulation algorithms. We quantitatively evaluate the relationship between LED transmission rate, communication distance and the message transmission accuracy for the smart visual beacon communication system that we prototyped. The proposed method achieves up to 4 kbps in an indoor environment and lossless transmission over a distance of 100 meters, at a transmission rate of 500 bps, in full sunlight, demonstrating the potential of the technology in an outdoor environment., Comment: 7 pages, 8 figures, accepted by IEEE International Conference on Intelligent Robots and Systems (IROS) 2022