Search

Scaling robot learning requires data collection pipelines that scale favorably with human effort. In this work, we propose Crowdsourcing and Amortizing Human Effort for Real-to-Sim-to-Real(CASHER), a pipeline for scaling up data collection and learning in simulation where the performance scales superlinearly with human effort. The key idea is to crowdsource digital twins of real-world scenes using 3D reconstruction and collect large-scale data in simulation, rather than the real-world. Data collection in simulation is initially driven by RL, bootstrapped with human demonstrations. As the training of a generalist policy progresses across environments, its generalization capabilities can be used to replace human effort with model generated demonstrations. This results in a pipeline where behavioral data is collected in simulation with continually reducing human effort. We show that CASHER demonstrates zero-shot and few-shot scaling laws on three real-world tasks across diverse scenarios. We show that CASHER enables fine-tuning of pre-trained policies to a target scenario using a video scan without any additional human effort. See our project website: https://casher-robot-learning.github.io/CASHER/

We present the first numerical investigation of the method proposed in Ref. [1] to utilize gradient flow to obtain precise determinations of higher moments of PDFs from lattice QCD, circumventing power divergent mixing with lower dimensional operators. We apply this method to obtain moments of the isovector PDF of the pion using four Stabilized Wilson Fermion ensembles with $m_{\pi}\simeq 411$ MeV and lattice spacings $a \simeq 0.064, 0.077, 0.094$, and $0.12$ fm. We present preliminary results of ratios of three-point functions as a function of flow time, which can be used to extract the ratios $\left\langle x^2 \right\rangle/\left\langle x \right\rangle$ and $\left\langle x^3 \right\rangle/\left\langle x \right\rangle$. We find that a significantly higher precision can be achieved with this method compared to the canonical approach, which requires boosting and cannot reach higher than the $\left\langle x^3 \right\rangle$ moment., Comment: 15 pages, 2 figures, contributions to the proceedings of the 41st International Symposium on Lattice Field Theory (LATTICE2024)

We investigate the impact of Mutual Coupling (MC) between antennas on the time-delay power spectrum response of the core of the SKA-Low radio telescope. Using two in-house tools - Fast Array Simulation Tool (FAST) (a fast full-wave electromagnetic solver) and OSKAR (a GPU-accelerated radio telescope simulator) - we simulate station beams and compute visibilities for various array layouts (regular, sunflower, and random). Simulations are conducted in an Epoch of Reionisation subband between 120-150~MHz, with a fine frequency resolution of 100~kHz, enabling the investigation of longer delays. Our results show that MC effects significantly increase foreground leakage into longer delays, especially for regular station layouts. For 21-cm science, foreground spill-over into the 21-cm window extends beyond $k_{\parallel} \sim 2$~h$^{-1}$Mpc for all station layouts and across all $k_{\perp}$ modes, completely obscuring the detection window. We find that attempting to remove the foreground contribution from the visibilities using an approximated beam model, based on the average embedded element pattern or interpolating the embedded element patterns from a coarse channel rate of 781~kHz, results in residuals around 1% ($\sim 10^{11}~\mathrm{mK}^2$h$^{-3}\mathrm{Mpc}^3$) which is still around 7 orders of magnitude brighter than the expected level of the EoR signal ($\sim 10^{4}~\mathrm{mK}^2$h$^{-3}\mathrm{Mpc}^3$). We also find that station beam models with at least 4-5 significant digits in the far-field pattern and high spectral resolution are needed for effective foreground removal. Our research provides critical insights into the role of MC in SKA-Low experiments and highlights the computational challenges of fully integrating array patterns that account for MC effects into processing pipelines., Comment: 18 pages, 17 figures

As prompt engineering research rapidly evolves, evaluations beyond accuracy are crucial for developing cost-effective techniques. We present the Economical Prompting Index (EPI), a novel metric that combines accuracy scores with token consumption, adjusted by a user-specified cost concern level to reflect different resource constraints. Our study examines 6 advanced prompting techniques, including Chain-of-Thought, Self-Consistency, and Tree of Thoughts, across 10 widely-used language models and 4 diverse datasets. We demonstrate that approaches such as Self-Consistency often provide statistically insignificant gains while becoming cost-prohibitive. For example, on high-performing models like Claude 3.5 Sonnet, the EPI of simpler techniques like Chain-of-Thought (0.72) surpasses more complex methods like Self-Consistency (0.64) at slight cost concern levels. Our findings suggest a reevaluation of complex prompting strategies in resource-constrained scenarios, potentially reshaping future research priorities and improving cost-effectiveness for end-users., Comment: 5 pages (excluding references), accepted to Coling 2025

1. Obtaining data to train robust artificial intelligence (AI)-based models for species classification can be challenging, particularly for rare species. Data augmentation can boost classification accuracy by increasing the diversity of training data and is cheaper to obtain than expert-labelled data. However, many classic image-based augmentation techniques are not suitable for audio spectrograms. 2. We investigate two generative AI models as data augmentation tools to synthesise spectrograms and supplement audio data: Auxiliary Classifier Generative Adversarial Networks (ACGAN) and Denoising Diffusion Probabilistic Models (DDPMs). The latter performed particularly well in terms of both realism of generated spectrograms and accuracy in a resulting classification task. 3. Alongside these new approaches, we present a new audio data set of 640 hours of bird calls from wind farm sites in Ireland, approximately 800 samples of which have been labelled by experts. Wind farm data are particularly challenging for classification models given the background wind and turbine noise. 4. Training an ensemble of classification models on real and synthetic data combined gave 92.6% accuracy (and 90.5% with just the real data) when compared with highly confident BirdNET predictions. 5. Our approach can be used to augment acoustic signals for more species and other land-use types, and has the potential to bring about a step-change in our capacity to develop reliable AI-based detection of rare species. Our code is available at https://github.com/gibbona1/ SpectrogramGenAI., Comment: 18 pages, 3 tables, 5 figures

We demonstrate the advantages of THz frequency combs for high-resolution spectroscopy. This benefits from wide spectral coverage and the exact knowledge of the frequency position of each comb component. Heterodyne detection combined with a fast Fourier spectrometer enables rapid and simultaneous measurement of more than 80 frequency comb modes covering a 7.5 GHz bandwidth. A spectrum is obtained in under 20 minutes yielding a uniform resolution of 70 kHz. This new setup has been validated by recording more than 150 lines of methanol around 723 GHz, and represents a new solution to exploit THz frequency combs for high-resolution spectroscopy.

Dielectric barrier discharge (DBD) plasma actuator arrays have been suggested as active flow control devices due to the robust electrohydrodynamic (EHD) force generation in variable atmospheric conditions. DBD plasma augmentation schemes allow for significant performance improvements. However, the transitions to sliding discharge or counter-flow discharge limit their use in high-power arrays. Here, we experimentally demonstrate the performance of a scalable DBD array for two alternating phases of air-exposed electrode configuration. Plasma emissions, direct thrust, velocity profiles, and power consumption measurements of the DBD array reveal that cross-talk between DBD stages can be eliminated to create high-power density actuators. AC augmentation of plasma provides additional gains in thrust; however, the transition to sliding and filamentary discharge reveals geometric limits when increasing the array power density. Introducing a segmented electrode with a resistor delays the onset of adverse sliding and filamentary discharge, allowing it to operate at higher voltage inputs. An optimized four-stage DBD array generated thrust > 250 mN/m with a wall jet thickness > 15 mm, enabling a broader range of flow control applications.

In this paper we investigate a class of integrals that were encountered in the course of a work on statistical plasma physics, in the so-called Sommerfeld temperature-expansion of the electronic entropy. We show that such integrals, involving some parameters, can be fully described in closed form represented by special functions.

Cancer clinical trials often face challenges in recruitment and engagement due to a lack of participant-facing informational and educational resources. This study investigated the potential of Large Language Models (LLMs), specifically GPT4, in generating patient-friendly educational content from clinical trial informed consent forms. Using data from ClinicalTrials.gov, we employed zero-shot learning for creating trial summaries and one-shot learning for developing multiple-choice questions, evaluating their effectiveness through patient surveys and crowdsourced annotation. Results showed that GPT4-generated summaries were both readable and comprehensive, and may improve patients' understanding and interest in clinical trials. The multiple-choice questions demonstrated high accuracy and agreement with crowdsourced annotators. For both resource types, hallucinations were identified that require ongoing human oversight. The findings demonstrate the potential of LLMs "out-of-the-box" to support the generation of clinical trial education materials with minimal trial-specific engineering, but implementation with a human-in-the-loop is still needed to avoid misinformation risks.

Observations of tidal disruption events (TDEs) on a timescale of years after the main flare show evidence of continued activity in the form of optical/UV emission, quasi-periodic eruptions, and delayed radio flares. Motivated by this, we explore the time evolution of these disks using semi-analytic models to follow the changing disk properties and feeding rate to the central black hole (BH). We find that thermal instabilities typically begin $\sim150-250\,{\rm days}$ after the TDE, causing the disk to cycle between high and low accretion states for up to $\sim10-20\,{\rm yrs}$. The high state is super-Eddington, which may be associated with outflows that eject $\sim10^{-3}-10^{-1}\,M_\odot$ with a range of velocities of $\sim0.03-0.3c$ over a span of a couple of days and produce radio flares. In the low state, the accretion rate slowly grows over many months to years as continued fallback accretion builds the mass of the disk. In this phase, the disk may reach luminosities of $\sim10^{41}-10^{42}\,{\rm erg\,s^{-1}}$ in the UV as seen in some late-time observations. We highlight the importance of the iron-opacity "bump" at $\approx2\times10^5\,{\rm K}$ in generating sufficiently high luminosities. This work suggests that joint optical/UV observations with radio monitoring could be key for following the disk state as the radio flares are produced., Comment: 15 pages, 14 figures, submitted to ApJ

The river Ganga is one of the Earth's most critically important river basins, yet it faces significant pollution challenges, making it crucial to evaluate its vulnerability for effective and targeted remediation efforts. While the Analytic Hierarchy Process (AHP) is widely regarded as the standard in decision making methodologies, uncertainties arise from its dependence on expert judgments, which can introduce subjectivity, especially when applied to remote sensing data, where expert knowledge might not fully capture spatial and spectral complexities inherent in such data. To address that, in this paper, we applied AHP alongside a suite of alternative existing and novel variants of AHP-based decision analysis on remote sensing data to assess the vulnerability of the river Ganga to pollution. We then compared the areas where the outputs of each variant may provide additional insights over AHP. Lastly, we utilized our learnings to design a composite variable to robustly define the vulnerability of the river Ganga to pollution. This approach contributes to a more comprehensive understanding of remote sensing data applications in environmental assessment, and these decision making variants can also have broader applications in other areas of environment management and sustainability, facilitating more precise and adaptable decision support frameworks.

The Wide Area Linear Optical Polarimeter North (WALOP-North) is an optical polarimeter designed for the needs of the PASIPHAE survey. It will be installed on the 1.3m telescope at the Skinakas Observatory in Crete, Greece. After commissioning, it will measure the polarization of millions of stars at high Galactic latitude, aiming to measure hundreds of stars per $deg^2$. The astronomical filter used in the instrument is a modified, polarimetrically-neutral broadband SDSS-r. This instrument will be pioneering one due to its large field-of-view (FoV) of $30\times 30$ $arcmin^2$ and high accuracy polarimetry measurements. The accuracy and sensitivity of the instrument in polarization fraction will be at the 0.1\% and 0.05\% level, respectively. Four separate 4k$\times$4k CCDs will be used as the instrument detectors, each imaging one of the $0\deg{}, 45\deg{}, 90\deg{}$ and $135\deg{}$ polarized FoV separately, therefore making the instrument a four-channel, one-shot polarimeter. Here, we present the overall optical design of the instrument, emphasizing on the aspects of the instrument that are different from WALOP-South. We also present a novel design of filters appropriate for polarimetry along with details on the management of the instrument size and its polarimetric calibration., Comment: 26 pages, 31 figures, 7 tables

Background: Formulation, associated with suicide risk assessment, is an individualised process that seeks to understand the idiosyncratic nature and development of an individual's problems. Auditing clinical documentation on an electronic health record (EHR) is challenging as it requires resource-intensive manual efforts to identify keywords in relevant sections of specific forms. Furthermore, clinicians and healthcare professionals often do not use keywords; their clinical language can vary greatly and may contain various jargon and acronyms. Also, the relevant information may be recorded elsewhere. This study describes how we developed advanced Natural Language Processing (NLP) algorithms, a branch of Artificial Intelligence (AI), to analyse EHR data automatically. Method: Advanced Optical Character Recognition techniques were used to process unstructured data sets, such as portable document format (pdf) files. Free text data was cleaned and pre-processed using Normalisation of Free Text techniques. We developed algorithms and tools to unify the free text. Finally, the formulation was checked for the presence of each concept based on similarity using NLP-powered semantic matching techniques. Results: We extracted information indicative of formulation and assessed it to cover the relevant concepts. This was achieved using a Weighted Score to obtain a Confidence Level. Conclusion: The rigour to which formulation is completed is crucial to effectively using EHRs, ensuring correct and timely identification, engagement and interventions that may potentially avoid many suicide attempts and suicides., Comment: Have some incorrect information

Animal excretions in form of urine puddles and feces are a significant source of emissions in livestock farming. Automated detection of soiled floor in barns can contribute to improved management processes but also the derived information can be used to model emission dynamics. Previous research approaches to determine the puddle area require manual detection of the puddle in the barn. While humans can detect animal excretions on thermal images of a livestock barn, automated approaches using thresholds fail due to other objects of the same temperature, such as the animals themselves. In addition, various parameters such as the type of housing, animal species, age, sex, weather and unknown factors can influence the type and shape of excretions. Due to this heterogeneity, a method for automated detection of excretions must therefore be not only be accurate but also robust to varying conditions. These requirements can be met by using contemporary deep learning models from the field of artificial intelligence. This work is the first to investigate the suitability of different deep learning models for the detection of excretions in pigsties, thereby comparing established convolutional architectures with recent transformer-based approaches. The detection models Faster R-CNN, YOLOv8, DETR and DAB-DETR are compared and statistically assessed on two created training datasets representing two pig houses. We apply a method derived from nested cross-validation and report on the results in terms of eight common detection metrics. Our work demonstrates that all investigated deep learning models are generally suitable for reliably detecting excretions with an average precision of over 90%. The models also show robustness on out of distribution data that possesses differences from the conditions in the training data, however, with expected slight decreases in the overall detection performance., Comment: Keywords: Artificial Intelligence, Objected detection, Pig, Urine puddle, Thermal IR data, CNN vs Transformer, Precision Livestock Farming; Stats: 54 pages, 13 figures, 1 graphical abstract

By a systematic development of fundamental concepts of conformable calculus we establish conformable divergence theorem and Green's identities which we combine with some new anisotropic Picone type identities to derive a generalized anisotropic Hardy type inequality with weights and conformable fractional differential operators. As a consequence, several Hardy type inequalities and Heisenberg Pauli-Weyl uncertainty principles are obtained., Comment: 15 pages

Qualitative Spatial Reasoning is a well explored area of Knowledge Representation and Reasoning and has multiple applications ranging from Geographical Information Systems to Robotics and Computer Vision. Recently, many claims have been made for the reasoning capabilities of Large Language Models (LLMs). Here, we investigate the extent to which a set of representative LLMs can perform classical qualitative spatial reasoning tasks on the mereotopological Region Connection Calculus, RCC-8. We conduct three pairs of experiments (reconstruction of composition tables, alignment to human composition preferences, conceptual neighbourhood reconstruction) using state-of-the-art LLMs; in each pair one experiment uses eponymous relations and one, anonymous relations (to test the extent to which the LLM relies on knowledge about the relation names obtained during training). All instances are repeated 30 times to measure the stochasticity of the LLMs., Comment: 13 pages. arXiv admin note: text overlap with arXiv:2309.15577

Flow cytometry measurements are widely used in diagnostics and medical decision making. Incomplete understanding of sources of measurement uncertainty can make it difficult to distinguish autofluorescence and background sources from signals of interest. Moreover, established methods for modeling uncertainty overlook the fact that the apparent distribution of measurements is a convolution of the inherent the population variability (e.g., associated with calibration beads or cells) and instrument induced-effects. Such issues make it difficult, for example, to identify signals from small objects such as extracellular vesicles. To overcome such limitations, we formulate an explicit probabilistic measurement model that accounts for volume and labeling variation, background signals and fluorescence shot noise. Using raw data from routine per-event calibration measurements, we use this model to separate the aforementioned sources of uncertainty and demonstrate how such information can be used to facilitate decision-making and instrument characterization.

Clinical language models are important for many applications in healthcare, but their development depends on access to extensive clinical text for pretraining. However, obtaining clinical notes from electronic health records (EHRs) at scale is challenging due to patient privacy concerns. In this study, we rephrase existing clinical notes using LLMs to generate synthetic pretraining corpora, drawing inspiration from previous work on rephrasing web data. We examine four popular small-sized LLMs (<10B) to create synthetic clinical text to pretrain both decoder-based and encoder-based language models. The method yields better results in language modeling and downstream tasks than previous synthesis approaches without referencing real clinical text. We find that augmenting original clinical notes with synthetic corpora from different LLMs improves performances even at a small token budget, showing the potential of this method to support pretraining at the institutional level or be scaled to synthesize large-scale clinical corpora.

Adversarial input image perturbation attacks have emerged as a significant threat to machine learning algorithms, particularly in image classification setting. These attacks involve subtle perturbations to input images that cause neural networks to misclassify the input images, even though the images remain easily recognizable to humans. One critical area where adversarial attacks have been demonstrated is in automotive systems where traffic sign classification and recognition is critical, and where misclassified images can cause autonomous systems to take wrong actions. This work presents a new class of adversarial attacks. Unlike existing work that has focused on adversarial perturbations that leverage human-made artifacts to cause the perturbations, such as adding stickers, paint, or shining flashlights at traffic signs, this work leverages nature-made artifacts: tree leaves. By leveraging nature-made artifacts, the new class of attacks has plausible deniability: a fall leaf stuck to a street sign could come from a near-by tree, rather than be placed there by an malicious human attacker. To evaluate the new class of the adversarial input image perturbation attacks, this work analyses how fall leaves can cause misclassification in street signs. The work evaluates various leaves from different species of trees, and considers various parameters such as size, color due to tree leaf type, and rotation. The work demonstrates high success rate for misclassification. The work also explores the correlation between successful attacks and how they affect the edge detection, which is critical in many image classification algorithms.

Precise weed management is essential for sustaining crop productivity and ecological balance. Traditional herbicide applications face economic and environmental challenges, emphasizing the need for intelligent weed control systems powered by deep learning. These systems require vast amounts of high-quality training data. The reality of scarcity of well-annotated training data, however, is often addressed through generating more data using data augmentation. Nevertheless, conventional augmentation techniques such as random flipping, color changes, and blurring lack sufficient fidelity and diversity. This paper investigates a generative AI-based augmentation technique that uses the Stable Diffusion model to produce diverse synthetic images that improve the quantity and quality of training datasets for weed detection models. Moreover, this paper explores the impact of these synthetic images on the performance of real-time detection systems, thus focusing on compact CNN-based models such as YOLO nano for edge devices. The experimental results show substantial improvements in mean Average Precision (mAP50 and mAP50-95) scores for YOLO models trained with generative AI-augmented datasets, demonstrating the promising potential of synthetic data to enhance model robustness and accuracy.

We develop a theoretical framework to investigate the link between rising scale economies and stagnating productivity. Our model features heterogeneous firms, imperfect competition, and firm selection. We demonstrate that scale economies generated by fixed costs have distinct impacts on aggregate productivity compared to those driven by returns to scale. Using UK data, we estimate long-run increases in both fixed costs and returns to scale. Our model implies that this should increase aggregate productivity through improved firm selection and resource allocation. However, increasing markups can offset the productivity gain. Higher markups cushion low-productivity firms' revenues, allowing them to survive, and constrain firm output, which limits exploitation of scale economies.

The authors here show that the partition $P_{k,l}(Q)$ in the table $\mathcal T(Q)$ of partitions having maximal nilpotent commutator a given stable partition $Q$, defined in [IKVZ2], is identical to the analogous partition $P_{k,l}^Q$ defined by the authors in [BIK] using the Burge correspondence., Comment: 15 pages, 3 figures, Overlap with arXiv 2409.13553 (this is [BIK]) section 2, Notations and Background, in order for this note to be self contained

We present q2-fmt, a QIIME 2 plugin that provides diverse methods for assessing the extent of microbiome engraftment following fecal microbiota transplant. The methods implemented here were informed by a recent literature review on approaches for assessing FMT engraftment, and cover aspects of engraftment including Chimeric Asymmetric Community Coalescence, Donated Microbiome Indicator Features, and Temporal Stability. q2-fmt is free for all use, and detailed documentation illustrating worked examples on a real-world data set are provided in the project's documentation., Comment: 5 pages, 1 figure

The Affleck-Kennedy-Lieb-Tasaki (AKLT) point of the bilinear-biquadratic spin-1 chain is a cornerstone example of a disorder point where short-range correlations become incommensurate, and correlation lengths and momenta are non-analytic. While the presence of singularities appears to be generic for AKLT points, we show that for a family of SU(N) models, the AKLT point is not a disorder point: It occurs entirely within an incommensurate phase yet the wave vector remains singular on both sides of the AKLT point. We conjecture that this new possibility is generic for models where the representation is not self-conjugate and the transfer matrix non-Hermitian, while for self-conjugate representations the AKLT points remain disorder points., Comment: 4.5p + Appendices, comments welcome

Time series data has become increasingly prevalent across numerous domains, driving a growing demand for time series machine learning techniques. Among these, time series clustering (TSCL) stands out as one of the most popular machine learning tasks. TSCL serves as a powerful exploratory analysis tool and is also employed as a preprocessing step or subroutine for various tasks, including anomaly detection, segmentation, and classification. The most popular TSCL algorithms are either fast (in terms of run time) but perform poorly on benchmark problems, or perform well on benchmarks but scale poorly. We present a new TSCL algorithm, the $k$-means (K) accelerated (A) Stochastic subgradient (S) Barycentre (B) Average (A) (KASBA) clustering algorithm. KASBA is a $k$-means clustering algorithm that uses the Move-Split-Merge (MSM) elastic distance at all stages of clustering, applies a randomised stochastic subgradient gradient descent to find barycentre centroids, links each stage of clustering to accelerate convergence and exploits the metric property of MSM distance to avoid a large proportion of distance calculations. It is a versatile and scalable clusterer designed for real-world TSCL applications. It allows practitioners to balance run time and clustering performance. We demonstrate through extensive experimentation that KASBA produces significantly better clustering than the faster state of the art clusterers and is offers orders of magnitude improvement in run time over the most performant $k$-means alternatives.

Dielectric barrier discharge (DBD) plasma actuators generate an electrohydrodynamic (EHD) force through the ionization and acceleration of charged species. Most active flow control DBD applications are only practical at lower Reynolds numbers, and increasing the momentum injection can extend the practical uses of the technology. Here, we experimentally demonstrate improvement in the performance of a planar DBD actuator by utilizing an AC-augmented electrical field in a three-electrode geometry. Time-resolved electrical and optical measurements, velocity profiles, and direct thrust measurements were used to characterize the EHD augmentation. Varying phase shift and E-field strength between the two air-exposed DBD electrodes can accelerate EHD flow and increase EHD forcing by up to ~ 40%. At the most favorable conditions, the maximum thrust was 54 mN/m when the air-exposed electrodes were out of phase. In-phase operation of the exposed electrodes at high E-field conditions can induce adverse effects and sliding discharge. Mechanistically, the performance improvements in the AC-augmented DBD actuator primarily come from the additional charge pull action by the third electrode. The insight into the AC-augmented DBD mechanism allows for developing multi-stage arrays capable of further increasing EHD forces.

Physical vapor deposition can prepare organic glasses with high kinetic stability. When heated, these glassy solids slowly transform into the supercooled liquid in a process known as rejuvenation. In this study, we anneal vapor-deposited glasses of methyl-m-toluate (MMT) for six hours at 0.98Tg to observe rejuvenation using dielectric spectroscopy. Glasses of moderate stability exhibited partial or full rejuvenation in six hours. For highly stable glasses, prepared at substrate temperatures of 0.85Tg and 0.80Tg, the six-hour annealing time is ~2% of the estimated transformation time, and no change in the onset temperature for the {\alpha} relaxation process was observed, as expected. Surprisingly, for these highly stable glasses, annealing resulted in significant increases in the storage component of the dielectric susceptibility, without corresponding increases in the loss component. These changes are interpreted to indicate that short-term annealing rejuvenates a high frequency relaxation (e.g., the boson peak) within the stable glass. We compare these results to computer simulations of the rejuvenation of highly stable glasses generated by the swap Monte Carlo algorithm. The in silico glasses, in contrast to the experiment, show no evidence of rejuvenation within the stable glass at times shorter than the alpha relaxation process.

Recent advances in radiance field reconstruction, such as 3D Gaussian Splatting (3DGS), have achieved high-quality novel view synthesis and fast rendering by representing scenes with compositions of Gaussian primitives. However, 3D Gaussians present several limitations for scene reconstruction. Accurately capturing hard edges is challenging without significantly increasing the number of Gaussians, creating a large memory footprint. Moreover, they struggle to represent flat surfaces, as they are diffused in space. Without hand-crafted regularizers, they tend to disperse irregularly around the actual surface. To circumvent these issues, we introduce a novel method, named 3D Convex Splatting (3DCS), which leverages 3D smooth convexes as primitives for modeling geometrically-meaningful radiance fields from multi-view images. Smooth convex shapes offer greater flexibility than Gaussians, allowing for a better representation of 3D scenes with hard edges and dense volumes using fewer primitives. Powered by our efficient CUDA-based rasterizer, 3DCS achieves superior performance over 3DGS on benchmarks such as Mip-NeRF360, Tanks and Temples, and Deep Blending. Specifically, our method attains an improvement of up to 0.81 in PSNR and 0.026 in LPIPS compared to 3DGS while maintaining high rendering speeds and reducing the number of required primitives. Our results highlight the potential of 3D Convex Splatting to become the new standard for high-quality scene reconstruction and novel view synthesis. Project page: convexsplatting.github.io., Comment: 13 pages, 13 figures, 10 tables

Characterizing domains is essential for models analyzing dynamic environments, as it allows them to adapt to evolving conditions or to hand the task over to backup systems when facing conditions outside their operational domain. Existing solutions typically characterize a domain by solving a regression or classification problem, which limits their applicability as they only provide a limited summarized description of the domain. In this paper, we present a novel approach to domain characterization by characterizing domains as probability distributions. Particularly, we develop a method to predict the likelihood of different weather conditions from images captured by vehicle-mounted cameras by estimating distributions of physical parameters using normalizing flows. To validate our proposed approach, we conduct experiments within the context of autonomous vehicles, focusing on predicting the distribution of weather parameters to characterize the operational domain. This domain is characterized by physical parameters (absolute characterization) and arbitrarily predefined domains (relative characterization). Finally, we evaluate whether a system can safely operate in a target domain by comparing it to multiple source domains where safety has already been established. This approach holds significant potential, as accurate weather prediction and effective domain adaptation are crucial for autonomous systems to adjust to dynamic environmental conditions.

In this technical report, we present the Zamba2 series -- a suite of 1.2B, 2.7B, and 7.4B parameter hybrid Mamba2-transformer models that achieve state of the art performance against the leading open-weights models of their class, while achieving substantial gains in inference latency, throughput, and memory efficiency. The Zamba2 series builds upon our initial work with Zamba1-7B, optimizing its architecture, training and annealing datasets, and training for up to three trillion tokens. We provide open-source weights for all models of the Zamba2 series as well as instruction-tuned variants that are strongly competitive against comparable instruct-tuned models of their class. We additionally open-source the pretraining dataset, which we call Zyda-2, used to train the Zamba2 series of models. The models and datasets used in this work are openly available at https://huggingface.co/Zyphra, Comment: 21/11/24 initial upload

The energy-efficient trip allocation of mobile robots employing differential drives for data retrieval from stationary sensor locations is the scope of this article. Given a team of robots and a set of targets (wireless sensor nodes), the planner computes all possible tours that each robot can make if it needs to visit a part of or the entire set of targets. Each segment of the tour relies on a minimum energy path planning algorithm. After the computation of all possible tour-segments, a utility function penalizing the overall energy consumption is formed. Rather than relying on the NP-hard Mobile Element Scheduling (MES) MILP problem, an approach using elements from game theory is employed. The suggested approach converges fast for most practical reasons thus allowing its utilization in near real time applications. Simulations are offered to highlight the efficiency of the developed algorithm., Comment: In proceedings of the 5th International Conference on Control, Decision and Information Technologies, 2018. 6 pages, 5 figures

Wall-based active and passive flow control for drag reduction in low Reynolds number (Re) turbulent flows can lead to three typical phenomena: i) attenuation or ii) amplification of the near-wall cycle, and iii) generation of spanwise rollers. The present study conducts direct numerical simulations (DNS) of a low Re turbulent channel flow and demonstrates that each flow response can be generated with a wall transpiration at two sets of spatial scales, termed "streak" and "roller" scales. The effect of the transpiration is controlled by its relative phase to the background flow, which can be parametrized by the wall pressure. Streak scales i) attenuate the near-wall cycle if transpiration and wall-pressure are approximately in-phase or ii) amplify it otherwise, and iii) roller scales energize spanwise rollers when transpiration and wall pressure are out-of-phase. The dynamics of the wall pressure and transpiration are coupled and robust relative phase relations, which are required to trigger the flow responses, can result if the source term of the linear fast or nonlinear slow pressure correlates with the wall transpiration over a scale-dependent height or if the temporal frequency content of the wall transpiration is approximately sparse. The importance of each condition depends on the relative magnitude of the pressure components, which is significantly altered by the transpiration. The analogy in flow response suggests that transpiration with the two scale families and their phase relations to the wall pressure represent fundamental building blocks for flows over tailored surfaces including riblets, porous, and permeable walls., Comment: 40 pages, 15 figures

Large Language Models (LLMs) demonstrate strong abilities in common-sense reasoning and interactive decision-making, but often struggle with complex, long-horizon planning tasks. Recent techniques have sought to structure LLM outputs using control flow and other code-adjacent techniques to improve planning performance. These techniques include using variables (to track important information) and functions (to divide complex tasks into smaller re-usable sub-tasks). However, purely code-based approaches can be error-prone and insufficient for handling ambiguous or unstructured data. To address these challenges, we propose REPL-Plan, an LLM planning approach that is fully code-expressive (it can utilize all the benefits of code) while also being dynamic (it can flexibly adapt from errors and use the LLM for fuzzy situations). In REPL-Plan, an LLM solves tasks by interacting with a Read-Eval-Print Loop (REPL), which iteratively executes and evaluates code, similar to language shells or interactive code notebooks, allowing the model to flexibly correct errors and handle tasks dynamically. We demonstrate that REPL-Plan achieves strong results across various planning domains compared to previous methods.

We introduce the Situated Corpus Of Understanding Transactions (SCOUT), a multi-modal collection of human-robot dialogue in the task domain of collaborative exploration. The corpus was constructed from multiple Wizard-of-Oz experiments where human participants gave verbal instructions to a remotely-located robot to move and gather information about its surroundings. SCOUT contains 89,056 utterances and 310,095 words from 278 dialogues averaging 320 utterances per dialogue. The dialogues are aligned with the multi-modal data streams available during the experiments: 5,785 images and 30 maps. The corpus has been annotated with Abstract Meaning Representation and Dialogue-AMR to identify the speaker's intent and meaning within an utterance, and with Transactional Units and Relations to track relationships between utterances to reveal patterns of the Dialogue Structure. We describe how the corpus and its annotations have been used to develop autonomous human-robot systems and enable research in open questions of how humans speak to robots. We release this corpus to accelerate progress in autonomous, situated, human-robot dialogue, especially in the context of navigation tasks where details about the environment need to be discovered., Comment: 14 pages, 7 figures

In this paper, we describe the development of symbolic representations annotated on human-robot dialogue data to make dimensions of meaning accessible to autonomous systems participating in collaborative, natural language dialogue, and to enable common ground with human partners. A particular challenge for establishing common ground arises in remote dialogue (occurring in disaster relief or search-and-rescue tasks), where a human and robot are engaged in a joint navigation and exploration task of an unfamiliar environment, but where the robot cannot immediately share high quality visual information due to limited communication constraints. Engaging in a dialogue provides an effective way to communicate, while on-demand or lower-quality visual information can be supplemented for establishing common ground. Within this paradigm, we capture propositional semantics and the illocutionary force of a single utterance within the dialogue through our Dialogue-AMR annotation, an augmentation of Abstract Meaning Representation. We then capture patterns in how different utterances within and across speaker floors relate to one another in our development of a multi-floor Dialogue Structure annotation schema. Finally, we begin to annotate and analyze the ways in which the visual modalities provide contextual information to the dialogue for overcoming disparities in the collaborators' understanding of the environment. We conclude by discussing the use-cases, architectures, and systems we have implemented from our annotations that enable physical robots to autonomously engage with humans in bi-directional dialogue and navigation., Comment: 52 pages, 14 figures

Single-photon detectors are ``blind" after the detection of a photon, and thereafter display a characteristic recovery in efficiency, during which the number of undetected photons depends on the statistics of the incident light. We show how the efficiency-recovery, photon statistics and intensity have an interdependent relationship which suppresses a detector's ability to count photons and measure correlations. We also demonstrate this effect with an experiment using $n$ such detectors to determine the $n^{\mathrm{th}}$ order correlation function with pseudothermal light., Comment: 8 pages, 6 figures including appendix

Large language models are increasingly becoming a cornerstone technology in artificial intelligence, the sciences, and society as a whole, yet the optimal strategies for dataset composition and filtering remain largely elusive. Many of the top-performing models lack transparency in their dataset curation and model development processes, posing an obstacle to the development of fully open language models. In this paper, we identify three core data-related challenges that must be addressed to advance open-source language models. These include (1) transparency in model development, including the data curation process, (2) access to large quantities of high-quality data, and (3) availability of artifacts and metadata for dataset curation and analysis. To address these challenges, we release RedPajama-V1, an open reproduction of the LLaMA training dataset. In addition, we release RedPajama-V2, a massive web-only dataset consisting of raw, unfiltered text data together with quality signals and metadata. Together, the RedPajama datasets comprise over 100 trillion tokens spanning multiple domains and with their quality signals facilitate the filtering of data, aiming to inspire the development of numerous new datasets. To date, these datasets have already been used in the training of strong language models used in production, such as Snowflake Arctic, Salesforce's XGen and AI2's OLMo. To provide insight into the quality of RedPajama, we present a series of analyses and ablation studies with decoder-only language models with up to 1.6B parameters. Our findings demonstrate how quality signals for web data can be effectively leveraged to curate high-quality subsets of the dataset, underscoring the potential of RedPajama to advance the development of transparent and high-performing language models at scale., Comment: 38th Conference on Neural Information Processing Systems (NeurIPS 2024) Track on Datasets and Benchmarks

JWST has revealed a large population of compact, red galaxies at $z>4$ known as Little Red Dots (LRDs). We analyze the spectral energy distributions (SEDs) of 95 LRDs from the JWST PRIMER survey with complete photometric coverage from $1-18\ \mu$m using NIRCam and MIRI imaging, representing the most extensive SED analysis on a large LRD sample with long-wavelength MIRI data. We examine SED models in which either galaxy or active galactic nucleus (AGN) emission dominates the rest-frame UV or optical continuum, extracting physical properties to explore each scenario's implications. In the galaxy-only model, we find massive, dusty stellar populations alongside unobscured, low-mass components, hinting at inhomogeneous obscuration. The AGN-only model indicates dusty, luminous AGNs with low hot dust fractions compared to typical quasars. A hybrid AGN and galaxy model suggests low-mass, unobscured galaxies in the UV, with stellar mass estimates spanning $\sim$2 dex across the different models, underscoring the need for caution in interpreting LRD stellar masses. With MIRI photometry, the galaxy-only model produces stellar masses within cosmological limits, but extremely high stellar mass densities are inferred. The hybrid model infers highly overmassive black holes exceeding those in recently reported high-redshift AGNs, hinting at a partial AGN contribution to the rest-optical continuum or widespread super-Eddington accretion. Our findings highlight the extreme conditions required for both AGN or galaxy dominated scenarios in LRDs, supporting a mixed contribution to the red continuum, or novel scenarios to explain the observed emission., Comment: 22 pages, 10 figures, submitted to ApJ. Machine-readable form of Table 2 available at: https://github.com/geneckleung/lrd_primer_miri

Graphical User Interface (GUI) grounding plays a crucial role in enhancing the capabilities of Vision-Language Model (VLM) agents. While general VLMs, such as GPT-4V, demonstrate strong performance across various tasks, their proficiency in GUI grounding remains suboptimal. Recent studies have focused on fine-tuning these models specifically for zero-shot GUI grounding, yielding significant improvements over baseline performance. We introduce a visual prompting framework that employs an iterative narrowing mechanism to further improve the performance of both general and fine-tuned models in GUI grounding. For evaluation, we tested our method on a comprehensive benchmark comprising various UI platforms and provided the code to reproduce our results., Comment: Code available at https://github.com/ant-8/GUI-Grounding-via-Iterative-Narrowing

This paper investigates the role of communication in improving coordination within robot swarms, focusing on a paradigm where learning and execution occur simultaneously in a decentralized manner. We highlight the role communication can play in addressing the credit assignment problem (individual contribution to the overall performance), and how it can be influenced by it. We propose a taxonomy of existing and future works on communication, focusing on information selection and physical abstraction as principal axes for classification: from low-level lossless compression with raw signal extraction and processing to high-level lossy compression with structured communication models. The paper reviews current research from evolutionary robotics, multi-agent (deep) reinforcement learning, language models, and biophysics models to outline the challenges and opportunities of communication in a collective of robots that continuously learn from one another through local message exchanges, illustrating a form of social learning., Comment: 17 pages, 3 Figures

Physics Informed Machine Learning has emerged as a popular approach in modelling and simulation for digital twins to generate accurate models of processes and behaviours of real-world systems. However, despite their success in generating accurate and reliable models, the existing methods either use simple regularizations in loss functions to offer limited physics integration or are too specific in architectural definitions to be generalized to a wide variety of physical systems. This paper presents a generic approach based on a novel physics-encoded residual neural network architecture to combine data-driven and physics-based analytical models to address these limitations. Our method combines physics blocks as mathematical operators from physics-based models with learning blocks comprising feed-forward layers. Intermediate residual blocks are incorporated for stable gradient flow as they train on physical system observation data. This way, the model learns to comply with the geometric and kinematic aspects of the physical system. Compared to conventional neural network-based methods, our method improves generalizability with substantially low data requirements and model complexity in terms of parameters, especially in scenarios where prior physics knowledge is either elementary or incomplete. We investigate our approach in two application domains. The first is a basic robotic motion model using Euler Lagrangian equations of motion as physics prior. The second application is a complex scenario of a steering model for a self-driving vehicle in a simulation. In both applications, our method outperforms both conventional neural network based approaches as-well as state-of-the-art Physics Informed Machine Learning methods.

Seysen's Python package mmgroup provides functionality for fast computations within the sporadic simple group $\mathbb{M}$, the Monster. The aim of this work is to present an mmgroup database of maximal subgroups of $\mathbb{M}$: for each conjugacy class $C$ of maximal subgroups in $\mathbb{M}$, we construct explicit group elements in mmgroup and prove that these elements generate a group in $C$. Our generators and the computations verifying correctness are available in accompanying code. The maximal subgroups of $\mathbb{M}$ have been classified in a number of papers spanning several decades; our work constitutes an independent verification of these constructions. We also correct the claim that $\mathbb{M}$ has a maximal subgroup $\mathrm{PSL}_2({59})$, and hence identify a new maximal subgroup $59{:}29$., Comment: 22 pages

We show how curing an anomaly of the twistor uplift of self-dual Yang-Mills theory implies linear relations among one-loop, $n$-gluon, color-ordered subamplitudes in QCD, when all $n$ gluon helicities are positive, or when exactly one is negative. We compute the number of linearly independent subamplitudes as determined by these relations, in terms of unsigned Stirling numbers. Then we use a momentum-twistor parametrization to show that there are no further linear dependencies.

Artificial Intelligence models encoding biology and chemistry are opening new routes to high-throughput and high-quality in-silico drug development. However, their training increasingly relies on computational scale, with recent protein language models (pLM) training on hundreds of graphical processing units (GPUs). We introduce the BioNeMo Framework to facilitate the training of computational biology and chemistry AI models across hundreds of GPUs. Its modular design allows the integration of individual components, such as data loaders, into existing workflows and is open to community contributions. We detail technical features of the BioNeMo Framework through use cases such as pLM pre-training and fine-tuning. On 256 NVIDIA A100s, BioNeMo Framework trains a three billion parameter BERT-based pLM on over one trillion tokens in 4.2 days. The BioNeMo Framework is open-source and free for everyone to use.

A Hamiltonian lattice formulation of lattice gauge theories opens the possibility for quantum simulations of the non-perturbative dynamics of QCD. By parametrizing the gauge invariant Hilbert space in terms of plaquette degrees of freedom, we show how the Hilbert space and interactions can be expanded in inverse powers of $N_c$. At leading order in this expansion, the Hamiltonian simplifies dramatically, both in the required size of the Hilbert space as well as the type of interactions involved. Adding a truncation of the resulting Hilbert space in terms of local electric energy states we give explicit constructions that allow simple representations of SU(3) gauge fields on qubits and qutrits to leading order in large $N_c$, Comment: 12 pages, 4 figures, Contribution to the 41st International Symposium on Lattice Field theory (LATTICE2024), 28th July, 2024, Liverpool, United Kingdom

We present complete results for the hadronic vacuum polarization (HVP) contribution to the muon anomalous magnetic moment $a_\mu$ in the short- and intermediate-distance window regions, which account for roughly 10% and 35% of the total HVP contribution to $a_\mu$, respectively. In particular, we perform lattice-QCD calculations for the isospin-symmetric connected and disconnected contributions, as well as corrections due to strong isospin-breaking. For the short-distance window observables, we investigate the so-called log-enhancement effects as well as the significant oscillations associated with staggered quarks in this region. For the dominant, isospin-symmetric light-quark connected contribution, we obtain $a^{ll,\,{\mathrm{SD}}}_{\mu}(\mathrm{conn.}) = 48.116(16)(94)[96] \times 10^{-10}$ and $a^{ll,\,{\mathrm{W}}}_{\mu}(\mathrm{conn.}) = 207.06(17)(63)[66] \times 10^{-10}$. We use Bayesian model averaging combined with a global bootstrap to fully estimate the covariance matrix between the individual contributions. Our determinations of the complete window contributions are $a^{{\mathrm{SD}}}_{\mu} = 69.01(2)(21)[21] \times 10^{-10}$ and $a^{{\mathrm{W}}}_{\mu} = 236.57(20)(94)[96] \times 10^{-10}$. This work is part of our ongoing effort to compute all contributions to HVP with an overall uncertainty at the few permille level.

One way to define a sub-Riemannian metric is as the limit of a Riemannian metric. Consider a Riemannian structure depending on a parameter $s$ such that its limit defines a sub-Riemannian metric when $s \to \infty$, assuming that the Riemannian geodesic flow is integrable for all $s$. An interesting question is: Can we determine the integrability of the sub-Riemannian geodesic flow as the limit of the integrals of motion of the Riemannian geodesic flow? The paper's main contribution is to provide a positive answer to this question in the special orthogonal group. Theorem 1.1 states that the Riemannian geodesic flow is Liuville integrable: The Manakov integrals' limit suggests the existence of a Lax pair formulation of the Riemannian geodesic equations, and the proof of Theorem 1.1 relies on this Lax pair.

The description of a closed quantum system is extended with the identification of an underlying substructure in which an expanded formulation of dynamics in the Heisenberg picture is given. Between measurements a ``state point" moves in an underlying multi-dimensional complex, projective space with constant velocity determined by the quantum state vector. Following a measurement, the point changes direction and moves with new constant velocity along one of several possible new orthogonal paths, with probabilities determined by Born's Rule. A possible generalization to a Riemannian substructure is speculated upon, suggesting an unexpected interaction with the background gravitational field., Comment: Typos and minor errors corrected; section headings refined; short concluding remarks added

We characterize the power of constant-depth Boolean circuits in generating uniform symmetric distributions. Let $f\colon\{0,1\}^m\to\{0,1\}^n$ be a Boolean function where each output bit of $f$ depends only on $O(1)$ input bits. Assume the output distribution of $f$ on uniform input bits is close to a uniform distribution $D$ with a symmetric support. We show that $D$ is essentially one of the following six possibilities: (1) point distribution on $0^n$, (2) point distribution on $1^n$, (3) uniform over $\{0^n,1^n\}$, (4) uniform over strings with even Hamming weights, (5) uniform over strings with odd Hamming weights, and (6) uniform over all strings. This confirms a conjecture of Filmus, Leigh, Riazanov, and Sokolov (RANDOM 2023)., Comment: This version improves the main result by removing dependence on d from the final distance bound

Binaries are known to play a key role in the mass loss and dynamical environments of evolved stars. Stellar and sub-stellar companion interactions produce complex wind morphologies including rotating/expanding disks, bipolar outflows, and spiral wind patterns; however, the connection between these many structures and the gas phase chemistry they harbor is not well-constrained. To expand the sample of chemical inventories in interacting systems, we present a detailed spectroscopic case study of the binary C-rich Asymptotic Giant Branch (AGB) star V Hya. Using spatially resolved ALMA observations at Bands 3, 6 and 7, we characterize the rotational emission lines and distributions of molecules in its surrounding disk undergoing dynamical expansion (DUDE). We detect emission from over 15 molecules and isotopologues toward this source, and present resolved maps for the brightest tracers of carbonaceous chemistry (e.g. CCH, C4H, HC5N, HNC, CH3CN). Employing LTE and non-LTE models of emission from the DUDE, we estimate the abundance distributions for optically thin species, and compare them with prototypical carbon-rich AGB envelopes. We find that the average abundances of detected species are within a factor of ${\sim}5$ from sources with similar mass-loss rates; however, the distribution of daughter species in V Hya is much more compact, with carbon chain species (CCH, C4H, HC3N) appearing with abundances $>$10$^{-7}$ even in the innermost sampled regions (200 au) of the disk., Comment: 40 pages, 27 figures, 2 appendices. Accepted for publication in ApJ