Search

We search for excess in-transit absorption of neutral helium at 1.083 $\mu$m in the atmospheres of the young (<800 Myr) sub-Jovian (0.2-0.5 $\rm R_{J}$) planets HD 63433b, K2-100b, and V1298 Tau c using high-resolution (R~25,000) transit observations taken with Keck II/NIRSPEC. Our observations do not show evidence of helium absorption for any of the planets in our sample. We calculate 3$\sigma$ upper limits on the planets' excess helium absorption of <0.47% for HD 63433b, <0.56% for K2-100b, and <1.13% for V1298 Tau c. In terms of equivalent width, we constrain these to <2.52, <4.44, and <8.49 mA for HD 63433b, K2-100b, and V1298 Tau c, respectively. We fit our transmission spectra with one-dimensional Parker wind models to determine upper limits on the planets' mass-loss rates of <7.9$\times10^{10}$, <1.25$\times10^{11}$, and <$7.9\times10^{11}$g s$^{-1}$. Our non-detections align with expectations from one-dimensional hydrodynamic escape models, magnetic fields, and stellar wind confinement. The upper limits we measure for these planets are consistent with predicted trends in system age and He equivalent width from 1D hydrodynamic models., Comment: 24 pages, 14 figures, accepted for publication in The Astronomical Journal

We consider genealogies arising from a Markov population process in which individuals are categorized into a discrete collection of compartments, with the requirement that individuals within the same compartment are statistically exchangeable. When equipped with a sampling process, each such population process induces a time-evolving tree-valued process defined as the genealogy of all sampled individuals. We provide a construction of this genealogy process and derive exact expressions for the likelihood of an observed genealogy in terms of filter equations. These filter equations can be numerically solved using standard Monte Carlo integration methods. Thus, we obtain statistically efficient likelihood-based inference for essentially arbitrary compartment models based on an observed genealogy of individuals sampled from the population.

While much about Alfven eigenmode (AE) stability has been explored in previous and current tokamaks, open questions remain for future burning plasma experiments, especially regarding exact stability threshold conditions and related isotope effects; the latter, of course, requiring good knowledge of the plasma ion composition. In the JET tokamak, eight in-vessel antennas actively excite stable AEs, from which their frequencies, toroidal mode numbers, and net damping rates are assessed. The effective ion mass can also be inferred using measurements of the plasma density and magnetic geometry. Thousands of AE stability measurements have been collected by the Alfven Eigenmode Active Diagnostic in hundreds of JET plasmas during the recent Hydrogen, Deuterium, Tritium, DT, and Helium-4 campaigns. In this novel AE stability database, spanning all four main ion species, damping is observed to decrease with increasing Hydrogenic mass, but increase for Helium, a trend consistent with radiative damping as the dominant damping mechanism. These data are important for confident predictions of AE stability in both non-nuclear (H/He) and nuclear (D/T) operations in future devices. In particular, if radiative damping plays a significant role in overall stability, some AEs could be more easily destabilized in D/T plasmas than their H/He reference pulses, even before considering fast ion and alpha particle drive. Active MHD spectroscopy is also employed on select HD, HT, and DT plasmas to infer the effective ion mass, thereby closing the loop on isotope analysis and demonstrating a complementary method to typical diagnosis of the isotope ratio.

Face Recognition models are commonly trained with web-scraped datasets containing millions of images and evaluated on test sets emphasizing pose, age and mixed attributes. With train and test sets both assembled from web-scraped images, it is critical to ensure disjoint sets of identities between train and test sets. However, existing train and test sets have not considered this. Moreover, as accuracy levels become saturated, such as LFW $>99.8\%$, more challenging test sets are needed. We show that current train and test sets are generally not identity- or even image-disjoint, and that this results in an optimistic bias in the estimated accuracy. In addition, we show that identity-disjoint folds are important in the 10-fold cross-validation estimate of test accuracy. To better support continued advances in face recognition, we introduce two "Goldilocks" test sets, Hadrian and Eclipse. The former emphasizes challenging facial hairstyles and latter emphasizes challenging over- and under-exposure conditions. Images in both datasets are from a large, controlled-acquisition (not web-scraped) dataset, so they are identity- and image-disjoint with all popular training sets. Accuracy for these new test sets generally falls below that observed on LFW, CPLFW, CALFW, CFP-FP and AgeDB-30, showing that these datasets represent important dimensions for improvement of face recognition. The datasets are available at: \url{https://github.com/HaiyuWu/SOTA-Face-Recognition-Train-and-Test}

Marine debris poses a significant ecological threat to birds, fish, and other animal life. Traditional methods for assessing debris accumulation involve labor-intensive and costly manual surveys. This study introduces a framework that utilizes aerial imagery captured by drones to conduct remote trash surveys. Leveraging computer vision techniques, our approach detects, classifies, and maps marine debris distributions. The framework uses Grounding DINO, a transformer-based zero-shot object detector, and CLIP, a vision-language model for zero-shot object classification, enabling the detection and classification of debris objects based on material type without the need for training labels. To mitigate over-counting due to different views of the same object, Scale-Invariant Feature Transform (SIFT) is employed for duplicate matching using local object features. Additionally, we have developed a user-friendly web application that facilitates end-to-end analysis of drone images, including object detection, classification, and visualization on a map to support cleanup efforts. Our method achieves competitive performance in detection (0.69 mean IoU) and classification (0.74 F1 score) across seven debris object classes without labeled data, comparable to state-of-the-art supervised methods. This framework has the potential to streamline automated trash sampling surveys, fostering efficient and sustainable community-led cleanup initiatives., Comment: 9 pages, 6 figures, 2 tables, In Proceedings of the 7th ACM SIGCAS/SIGCHI Conference on Computing and Sustainable Societies

The choroid is a key vascular layer of the eye, supplying oxygen to the retinal photoreceptors. Non-invasive enhanced depth imaging optical coherence tomography (EDI-OCT) has recently improved access and visualisation of the choroid, making it an exciting frontier for discovering novel vascular biomarkers in ophthalmology and wider systemic health. However, current methods to measure the choroid often require use of multiple, independent semi-automatic and deep learning-based algorithms which are not made open-source. Previously, Choroidalyzer -- an open-source, fully automatic deep learning method trained on 5,600 OCT B-scans from 385 eyes -- was developed to fully segment and quantify the choroid in EDI-OCT images, thus addressing these issues. Using the same dataset, we propose a Robust, Resolution-agnostic and Efficient Attention-based network for CHoroid segmentation (REACH). REACHNet leverages multi-resolution training with domain-specific data augmentation to promote generalisation, and uses a lightweight architecture with resolution-agnostic self-attention which is not only faster than Choroidalyzer's previous network (4 images/s vs. 2.75 images/s on a standard laptop CPU), but has greater performance for segmenting the choroid region, vessels and fovea (Dice coefficient for region 0.9769 vs. 0.9749, vessels 0.8612 vs. 0.8192 and fovea 0.8243 vs. 0.3783) due to its improved hyperparameter configuration and model training pipeline. REACHNet can be used with Choroidalyzer as a drop-in replacement for the original model and will be made available upon publication., Comment: 13 pages, 2 figures, 8 tables (including supplementary material)

Deep learning has demonstrated remarkable success across many domains, including computer vision, natural language processing, and reinforcement learning. Representative artificial neural networks in these fields span convolutional neural networks, Transformers, and deep Q-networks. Built upon unimodal neural networks, numerous multi-modal models have been introduced to address a range of tasks such as visual question answering, image captioning, and speech recognition. The rise of instruction-following robotic policies in embodied AI has spurred the development of a novel category of multi-modal models known as vision-language-action models (VLAs). Their multi-modality capability has become a foundational element in robot learning. Various methods have been proposed to enhance traits such as versatility, dexterity, and generalizability. Some models focus on refining specific components through pretraining. Others aim to develop control policies adept at predicting low-level actions. Certain VLAs serve as high-level task planners capable of decomposing long-horizon tasks into executable subtasks. Over the past few years, a myriad of VLAs have emerged, reflecting the rapid advancement of embodied AI. Therefore, it is imperative to capture the evolving landscape through a comprehensive survey., Comment: 15 pages, a survey of vision-language-action models

The current standard model of cosmology successfully describes a variety of measurements, but the nature of its main ingredients, dark matter and dark energy, remains unknown. Euclid is a medium-class mission in the Cosmic Vision 2015-2025 programme of the European Space Agency (ESA) that will provide high-resolution optical imaging, as well as near-infrared imaging and spectroscopy, over about 14,000 deg^2 of extragalactic sky. In addition to accurate weak lensing and clustering measurements that probe structure formation over half of the age of the Universe, its primary probes for cosmology, these exquisite data will enable a wide range of science. This paper provides a high-level overview of the mission, summarising the survey characteristics, the various data-processing steps, and data products. We also highlight the main science objectives and expected performance., Comment: Accepted for publication in the A&A special issue`Euclid on Sky'

This paper presents the specification, design, and development of the Visible Camera (VIS) on the ESA Euclid mission. VIS is a large optical-band imager with a field of view of 0.54 deg^2 sampled at 0.1" with an array of 609 Megapixels and spatial resolution of 0.18". It will be used to survey approximately 14,000 deg^2 of extragalactic sky to measure the distortion of galaxies in the redshift range z=0.1-1.5 resulting from weak gravitational lensing, one of the two principal cosmology probes of Euclid. With photometric redshifts, the distribution of dark matter can be mapped in three dimensions, and, from how this has changed with look-back time, the nature of dark energy and theories of gravity can be constrained. The entire VIS focal plane will be transmitted to provide the largest images of the Universe from space to date, reaching m_AB>24.5 with S/N >10 in a single broad I_E~(r+i+z) band over a six year survey. The particularly challenging aspects of the instrument are the control and calibration of observational biases, which lead to stringent performance requirements and calibration regimes. With its combination of spatial resolution, calibration knowledge, depth, and area covering most of the extra-Galactic sky, VIS will also provide a legacy data set for many other fields. This paper discusses the rationale behind the VIS concept and describes the instrument design and development before reporting the pre-launch performance derived from ground calibrations and brief results from the in-orbit commissioning. VIS should reach fainter than m_AB=25 with S/N>10 for galaxies of full-width half-maximum of 0.3" in a 1.3" diameter aperture over the Wide Survey, and m_AB>26.4 for a Deep Survey that will cover more than 50 deg^2. The paper also describes how VIS works with the other Euclid components of survey, telescope, and science data processing to extract the cosmological information., Comment: Paper submitted as part of the A&A special issue `Euclid on Sky', which contains Euclid key reference papers and first results from the Euclid Early Release Observations

The Super-Kamiokande and T2K collaborations present a joint measurement of neutrino oscillation parameters from their atmospheric and beam neutrino data. It uses a common interaction model for events overlapping in neutrino energy and correlated detector systematic uncertainties between the two datasets, which are found to be compatible. Using 3244.4 days of atmospheric data and a beam exposure of $19.7(16.3) \times 10^{20}$ protons on target in (anti)neutrino mode, the analysis finds a 1.9$\sigma$ exclusion of CP-conservation (defined as $J_{CP}=0$) and a preference for the normal mass ordering., Comment: 12 pages, 4 figures

Large language models (LLMs) have transformed AI and achieved breakthrough performance on a wide range of tasks that require human intelligence. In science, perhaps the most interesting application of LLMs is for hypothesis formation. A feature of LLMs, which results from their probabilistic structure, is that the output text is not necessarily a valid inference from the training text. These are 'hallucinations', and are a serious problem in many applications. However, in science, hallucinations may be useful: they are novel hypotheses whose validity may be tested by laboratory experiments. Here we experimentally test the use of LLMs as a source of scientific hypotheses using the domain of breast cancer treatment. We applied the LLM GPT4 to hypothesize novel pairs of FDA-approved non-cancer drugs that target the MCF7 breast cancer cell line relative to the non-tumorigenic breast cell line MCF10A. In the first round of laboratory experiments GPT4 succeeded in discovering three drug combinations (out of 12 tested) with synergy scores above the positive controls. These combinations were itraconazole + atenolol, disulfiram + simvastatin and dipyridamole + mebendazole. GPT4 was then asked to generate new combinations after considering its initial results. It then discovered three more combinations with positive synergy scores (out of four tested), these were disulfiram + fulvestrant, mebendazole + quinacrine and disulfiram + quinacrine. A limitation of GPT4 as a generator of hypotheses was that its explanations for them were formulaic and unconvincing. We conclude that LLMs are an exciting novel source of scientific hypotheses., Comment: 13 pages, 6 tables, 1 figure. Supplementary information available

We explore the time evolution of a topological system when the system undergoes a sudden quantum quench within the same nontrivial phase. Using Haldane's honeycomb model as an example, we show that equilibrium states in a topological phase can be distinguished by geometrical features, such as the characteristic momentum at which the half-occupied edge modes cross, the associated edge-mode velocity, and the winding vector about which the normalized pseudospin magnetic field winds along a great circle on the Bloch sphere. We generalize these geometrical quantities for non-equilibrium states and use them to visualize the quench dynamics of the topological system. In general, we find the pre-quench equilibrium state relaxes to the post-quench equilibrium state in an oscillatory fashion, whose amplitude decay as $t^{1/2}$. In the process, however, the characteristic winding vector of the non-equilibrium system can evolve to regimes that are not reachable with equilibrium states., Comment: 13 pages, 11 figures

Over the past year, there has been a robust debate about the benefits and risks of open sourcing foundation models. However, this discussion has often taken place at a high level of generality or with a narrow focus on specific technical attributes. In part, this is because defining open source for foundation models has proven tricky, given its significant differences from traditional software development. In order to inform more practical and nuanced decisions about opening AI systems, including foundation models, this paper presents a framework for grappling with openness across the AI stack. It summarizes previous work on this topic, analyzes the various potential reasons to pursue openness, and outlines how openness varies in different parts of the AI stack, both at the model and at the system level. In doing so, its authors hope to provide a common descriptive framework to deepen a nuanced and rigorous understanding of openness in AI and enable further work around definitions of openness and safety in AI.

One of the keys to the success of collaborative learning is balanced participation by all learners, but this does not always happen naturally. Pedagogical robots have the potential to facilitate balance. However, it remains unclear what participation balance robots should aim at; various metrics have been proposed, but it is still an open question whether we should balance human participation in human-human interactions (HHI) or human-robot interactions (HRI) and whether we should consider robots' participation in collaborative learning involving multiple humans and a robot. This paper examines collaborative learning between a pair of students and a teachable robot that acts as a peer tutee to answer the aforementioned question. Through an exploratory study, we hypothesize which balance metrics in the literature and which portions of dialogues (including vs. excluding robots' participation and human participation in HHI vs. HRI) will better predict learning as a group. We test the hypotheses with another study and replicate them with automatically obtained units of participation to simulate the information available to robots when they adaptively fix imbalances in real-time. Finally, we discuss recommendations on which metrics learning science researchers should choose when trying to understand how to facilitate collaboration., Comment: To appear in Seventeenth International Conference on Educational Data Mining (EDM 2024)

LLM watermarking, which embeds imperceptible yet algorithmically detectable signals in model outputs to identify LLM-generated text, has become crucial in mitigating the potential misuse of large language models. However, the abundance of LLM watermarking algorithms, their intricate mechanisms, and the complex evaluation procedures and perspectives pose challenges for researchers and the community to easily experiment with, understand, and assess the latest advancements. To address these issues, we introduce MarkLLM, an open-source toolkit for LLM watermarking. MarkLLM offers a unified and extensible framework for implementing LLM watermarking algorithms, while providing user-friendly interfaces to ensure ease of access. Furthermore, it enhances understanding by supporting automatic visualization of the underlying mechanisms of these algorithms. For evaluation, MarkLLM offers a comprehensive suite of 12 tools spanning three perspectives, along with two types of automated evaluation pipelines. Through MarkLLM, we aim to support researchers while improving the comprehension and involvement of the general public in LLM watermarking technology, fostering consensus and driving further advancements in research and application. Our code is available at https://github.com/THU-BPM/MarkLLM., Comment: EMNLP 2024 Demo

Low-Rank Adaptation (LoRA) is a widely-used parameter-efficient finetuning method for large language models. LoRA saves memory by training only low rank perturbations to selected weight matrices. In this work, we compare the performance of LoRA and full finetuning on two target domains, programming and mathematics. We consider both the instruction finetuning (approximately 100K prompt-response pairs) and continued pretraining (20B unstructured tokens) data regimes. Our results show that, in the standard low-rank settings, LoRA substantially underperforms full finetuning. Nevertheless, LoRA better maintains the base model's performance on tasks outside the target domain. We show that LoRA mitigates forgetting more than common regularization techniques such as weight decay and dropout; it also helps maintain more diverse generations. Finally, we show that full finetuning learns perturbations with a rank that is 10-100X greater than typical LoRA configurations, possibly explaining some of the reported gaps. We conclude by proposing best practices for finetuning with LoRA., Comment: Final version with new experiments and analyses, as accepted to Transactions on Machine Learning Research, August 2024 (Featured Certification). https://openreview.net/forum?id=aloEru2qCG&noteId=Jb3PQNQDI2

In this paper we show that it is possible to achieve successful hilltop inflation in which the inflaton is identified as the modulus field in a modular invariant theory. The dilaton plays a crucial role in shaping the potential. Modular invariant gaugino condensation provides the mechanism for the modulus stabilisation after inflation. The inflationary trajectory lies on the lower boundary of the fundamental domain of the modulus field $\tau$. Inflation starts near the fixed point $\tau ={\rm i}$, and ends at a point near $\tau = \omega$, which is the global de Sitter vacuum. We investigate the allowed parameter space for successful modular invariant hilltop inflation., Comment: 10 pages, 7 figures, match the published version in JCAP

Despite demonstrating superior rate-distortion (RD) performance, learning-based image compression (LIC) algorithms have been found to be vulnerable to malicious perturbations in recent studies. However, the adversarial attacks considered in existing literature remain divergent from real-world scenarios, both in terms of the attack direction and bitrate. Additionally, existing methods focus solely on empirical observations of the model vulnerability, neglecting to identify the origin of it. These limitations hinder the comprehensive investigation and in-depth understanding of the adversarial robustness of LIC algorithms. To address the aforementioned issues, this paper considers the arbitrary nature of the attack direction and the uncontrollable compression ratio faced by adversaries, and presents two practical rate-distortion attack paradigms, i.e., Specific-ratio Rate-Distortion Attack (SRDA) and Agnostic-ratio Rate-Distortion Attack (ARDA). Using the performance variations as indicators, we evaluate the adversarial robustness of eight predominant LIC algorithms against diverse attacks. Furthermore, we propose two novel analytical tools for in-depth analysis, i.e., Entropy Causal Intervention and Layer-wise Distance Magnify Ratio, and reveal that hyperprior significantly increases the bitrate and Inverse Generalized Divisive Normalization (IGDN) significantly amplifies input perturbations when under attack. Lastly, we examine the efficacy of adversarial training and introduce the use of online updating for defense. By comparing their advantages and disadvantages, we provide a reference for constructing more robust LIC algorithms against the rate-distortion attacks.

For a given polarized toric variety, we define the notion of $\lambda$-stability which is a natural generalization of uniform K-stability. At the neighbourhoods of the vertices of the corresponding moment polytope $\Delta$, we consider appropriate triangulations and give a sufficient criteria for a $\lambda$-stable polarized toric variety $(X,L)$ to be asymptotically Chow polystable when the obstruction of asymptotic Chow semistability (the Futaki-Ono invariant) vanishes. As an application, we prove that any K-semistable polarized smooth toric variety $(X,L)$ with the vanishing Futaki-Ono invariant is asymptotically Chow polystable., Comment: 36pages. Comments are welcome!

Improving calibration performance in deep learning (DL) classification models is important when planning the use of DL in a decision-support setting. In such a scenario, a confident wrong prediction could lead to a lack of trust and/or harm in a high-risk application. We evaluate the impact on accuracy and calibration of two types of approach that aim to improve DL classification model calibration: deterministic uncertainty methods (DUM) and uncertainty-aware training. Specifically, we test the performance of three DUMs and two uncertainty-aware training approaches as well as their combinations. To evaluate their utility, we use two realistic clinical applications from the field of cardiac imaging: artefact detection from phase contrast cardiac magnetic resonance (CMR) and disease diagnosis from the public ACDC CMR dataset. Our results indicate that both DUMs and uncertainty-aware training can improve both accuracy and calibration in both of our applications, with DUMs generally offering the best improvements. We also investigate the combination of the two approaches, resulting in a novel deterministic uncertainty-aware training approach. This provides further improvements for some combinations of DUMs and uncertainty-aware training approaches., Comment: currently under review for publication

An all-too-present bottleneck for text classification model development is the need to annotate training data and this need is multiplied for multilingual classifiers. Fortunately, contemporary machine translation models are both easily accessible and have dependable translation quality, making it possible to translate labeled training data from one language into another. Here, we explore the effects of using machine translation to fine-tune a multilingual model for a classification task across multiple languages. We also investigate the benefits of using a novel technique, originally proposed in the field of image captioning, to account for potential negative effects of tuning models on translated data. We show that translated data are of sufficient quality to tune multilingual classifiers and that this novel loss technique is able to offer some improvement over models tuned without it.

Merkle hash trees are the state-of-the-art method to protect the integrity of storage systems. However, using a hash tree can severely degrade performance, and prior works optimizing them have yet to yield a concrete understanding of the scalability of certain designs in the context of large-scale cloud storage systems. In this paper, we take a first-principles approach to analyzing hash tree performance for storage by introducing a definition of an optimal hash tree and a principled methodology for evaluating hash tree designs. We show that state-of-the-art designs are not scalable; they incur up to 40.1X slowdowns over an insecure baseline and deliver <50% of optimal performance across various experiments. We then exploit the characteristics of optimal hash trees to design Dynamic Hash Trees (DHTs), hash trees that can adapt to workload patterns on-the-fly, delivering >95% of optimal read and write performance and up to 4.2X speedups over the state-of-the art. Our novel methodology and DHT design provides a new foundation in the search for integrity mechanisms that can operate efficiently at scale.

Everyday devices like light bulbs and kitchen appliances are now embedded with so many features and automated behaviors that they have become complicated to actually use. While such "smart" capabilities can better support users' goals, the task of learning the "ins and outs" of different devices is daunting. Voice assistants aim to solve this problem by providing a natural language interface to devices, yet such assistants cannot understand loosely-constrained commands, they lack the ability to reason about and explain devices' behaviors to users, and they rely on connectivity to intrusive cloud infrastructure. Toward addressing these issues, we propose thoughtful things: devices that leverage lightweight, on-device language models to take actions and explain their behaviors in response to unconstrained user commands. We propose an end-to-end framework that leverages formal modeling, automated training data synthesis, and generative language models to create devices that are both capable and thoughtful in the presence of unconstrained user goals and inquiries. Our framework requires no labeled data and can be deployed on-device, with no cloud dependency. We implement two thoughtful things (a lamp and a thermostat) and deploy them on real hardware, evaluating their practical performance., Comment: 24 pages (3 pages of references)

The General Transit Feed Specification (GTFS) is an open standard format for recording transit information, utilized by thousands of transit agencies worldwide. This study introduces GTFS2STN, a novel tool that converts static GTFS transit networks into spatiotemporal networks, connecting bus stops across space and time. This transformation enables comprehensive analysis of transit system accessibility. Additionally, we present a web-based application version of the GTFS2STN tool that allows users to generate spatiotemporal networks online and perform basic analyses, including the creation of isochrone maps from a given origin and the calculation of travel time variability between origin-destination pairs over time. Comparative analysis demonstrates that GTFS2STN produces results similar to those of Mapnificent, an existing open-source tool for generating isochrone maps from GTFS inputs. Compared with Mapnificent, GTFS2STN offers enhanced flexibility for researchers and planners to evaluate transit plans, as it allows users to upload and analyze historical or suggested GTFS feeds from any transit agency. This feature facilitates the assessment of accessibility and travel time variability in transit networks over extended periods, making GTFS2STN a valuable tool for the planning and research for the transit systems., Comment: 13 pages, 12 figures

Demographic shifts have led to an increase in mild cognitive impairment (MCI), and this study investigates the effects of cognitive training (CT) and reminiscence therapy (RT) conducted by humans or socially assistive robots (SARs) on prefrontal cortex activation in elderly individuals with MCI, aiming to determine the most effective therapy-modality combination for promoting cognitive function. This pilot study employs a randomized control trial (RCT) design. Additionally, the study explores the efficacy of Reminiscence Therapy (RT) in comparison to Cognitive Training (CT). Eight MCI subjects, with a mean age of 70.125 years, were randomly assigned to ``human-led'' or ``SAR-led'' groups. Utilizing Functional Near-infrared Spectroscopy (fNIRS) to measure oxy-hemoglobin concentration changes in the dorsolateral prefrontal cortex (DLPFC), the study found no significant differences in the effects of human-led and SAR-led cognitive training on DLPFC activation. However, distinct patterns emerged in memory encoding and retrieval phases between RT and CT, shedding light on the impacts of these interventions on brain activation in the context of MCI., Comment: submitted to IEEE on affective computing

Quantum computation represents a promising frontier in the domain of high-performance computing, blending quantum information theory with practical applications to overcome the limitations of classical computation. This study investigates the challenges of manufacturing high-fidelity and scalable quantum processors. Quantum gate set tomography (QGST) is a critical method for characterizing quantum processors and understanding their operational capabilities and limitations. This paper introduces ML4QGST as a novel approach to QGST by integrating machine learning techniques, specifically utilizing a transformer neural network model. Adapting the transformer model for QGST addresses the computational complexity of modeling quantum systems. Advanced training strategies, including data grouping and curriculum learning, are employed to enhance model performance, demonstrating significant congruence with ground-truth values. We benchmark this training pipeline on the constructed learning model, to successfully perform QGST for $3$ gates on a $1$ qubit system with over-rotation error and depolarizing noise estimation with comparable accuracy to pyGSTi. This research marks a pioneering step in applying deep neural networks to the complex problem of quantum gate set tomography, showcasing the potential of machine learning to tackle nonlinear tomography challenges in quantum computing., Comment: 14 pages

We investigate the reproducing properties of Gabor systems within the context of expansible groups. These properties are established in terms of density conditions. The concept of density that we employ mirrors the well-known Beurling density defined in Euclidean space, which is made possible due to the expansive structure. Along the way, for groups with a compact open subgroup, we demonstrate that modulation spaces are continuously embedded in Wiener spaces. Utilizing this result, we derive the Bessel condition of Gabor systems. We also provide a straightforward proof of the density result for Gabor frames, utilizing a comparison theorem for coherent frames., Comment: This version fixes a previous error and adds more citations

The observed excesses in the search for neutralinos and charginos by ATLAS and CMS can be fitted simultaneously in the minimal supersymmetric standard model (MSSM) assuming a light higgsino mass, of magnitude less than about 250 GeV, and a compressed higgsino dominated neutralino and chargino spectrum, with $5-10\%$ mass splittings. However, light higgsinos as dark matter would have far too large direct detection cross sections. We consider the next-to-MSSM (NMSSM) with an additional singlino-like lightest supersymmetric particle (LSP) a few GeV below the next-to-lightest supersymmetric particle (NLSP). Sparticles prefer to decay first into the NLSP and remnants from the final decay into the LSP are too soft to contribute to the observed signals. Co-annihilation in the higgsino-sector can generate a relic density in the WMAP/Planck window. The singlino-like LSP has automatically a direct detection cross section below present and future sensitivities: a direct detection signal in the near future would exclude this scenario. The singlet-like Higgs scalar of the NMSSM can have a mass around 95 GeV and signal cross sections in the $b\bar{b}$ channel at LEP and in the $\gamma\gamma$ channel at the LHC compatible with the respective observations., Comment: 17 pages, 9 Figures, 4 Tables

Given copies of a quantum state $\rho$, a shadow tomography protocol aims to learn all expectation values from a fixed set of observables, to within a given precision $\epsilon$. We say that a shadow tomography protocol is triply efficient if it is sample- and time-efficient, and only employs measurements that entangle a constant number of copies of $\rho$ at a time. The classical shadows protocol based on random single-copy measurements is triply efficient for the set of local Pauli observables. This and other protocols based on random single-copy Clifford measurements can be understood as arising from fractional colorings of a graph $G$ that encodes the commutation structure of the set of observables. Here we describe a framework for two-copy shadow tomography that uses an initial round of Bell measurements to reduce to a fractional coloring problem in an induced subgraph of $G$ with bounded clique number. This coloring problem can be addressed using techniques from graph theory known as chi-boundedness. Using this framework we give the first triply efficient shadow tomography scheme for the set of local fermionic observables, which arise in a broad class of interacting fermionic systems in physics and chemistry. We also give a triply efficient scheme for the set of all $n$-qubit Pauli observables. Our protocols for these tasks use two-copy measurements, which is necessary: sample-efficient schemes are provably impossible using only single-copy measurements. Finally, we give a shadow tomography protocol that compresses an $n$-qubit quantum state into a $\text{poly}(n)$-sized classical representation, from which one can extract the expected value of any of the $4^n$ Pauli observables in $\text{poly}(n)$ time, up to a small constant error.

Particles that scatter waves interact through the waves they scatter. We establish that wave-mediated interactions between dissimilar particles are generically nonreciprocal. This apparent violation of Newton's third law is allowed because the pair of particles does not constitute a closed system; the balance of the system's momentum is carried away by the scattered wave. Nonreciprocal interactions enable collections of scatterers to transduce energy out of a wave and use it to power their own motion. Such wave-matter composites therefore are a form of active matter. Their activity, however, is not an inherent property of the individual particles, nor are the particles simply driven by the wave. Instead, the particles' collective motion is an emergent property of their state of organization. This kind of emergent activity should appear in any context where objects scatter waves. As a concrete example, we combine analysis, experiment and simulation to demonstrate how nonreciprocity and emergent activity arise in collections of acoustically levitated spheres.

The spontaneous breaking of a $U(1)$ symmetry via an intermediate discrete symmetry may yield a hybrid topological defect of \emph{domain walls bounded by cosmic strings}. The decay of this defect network leads to a unique gravitational wave signal spanning many orders in observable frequencies, that can be distinguished from signals generated by other sources. We investigate the production of gravitational waves from this mechanism in the context of the type-I two-Higgs-doublet model extended by a $U(1)_R$ symmetry, that simultaneously accommodates the seesaw mechanism, anomaly cancellation, and eliminates flavour-changing neutral currents. The gravitational wave spectrum produced by the string-bounded-wall network can be detected for $U(1)_R$ breaking scale from $10^{12}$ to $10^{15}$ GeV in forthcoming interferometers including LISA and Einstein Telescope, with a distinctive $f^{3}$ slope and inflexion in the frequency range between microhertz and hertz., Comment: 19 pages + reference, 4 figures, extended discussion on gravitational waves with unchanged conclusions, matches version published in JHEP

Observations of high-redshift quasars frequently promote suggestions of large black hole masses, whose presence so early in cosmic time is not easily explicable. I consider the parallel with ultraluminous X-ray sources (ULXs) -- now known to be stellar-mass black hole (and neutron star) binaries apparently radiating far above their Eddington luminosities $L_{\rm Edd}$. The true luminosity in ULXs is actually only of order $L_{\rm Edd}$, for {\it stellar-mass} accretors, but has a very anisotropic (`beamed') component, plus a near-isotropic component of similar luminosity but much lower specific intensity. Observers viewing ULXs from within the beam but assuming spherical symmetry deduce a luminosity $\gg L_{\rm Edd}$. These features appear because the accretors are fed mass at highly super-Eddington rates, most of it expelled in high-speed ($v >0.2c$) outflows from the accretion disc. I show that in similarly-beamed AGN, emission-line properties would be essentially the same as in unbeamed sources, but standard virial mass indicators unusable because velocity widths are dominated by the outflows, not bound motions about the black holes. In an ensemble of this kind the apparently most luminous systems are always the most distant, but have the lowest black hole masses. Interpreting observations of this ensemble without knowing that they are beamed leads instead to very high black hole mass estimates. The analogy with ULXs therefore suggests that high-redshift quasars might actually have central black hole masses which could have grown from stellar values within the lookback time. I consider how one might test these ideas observationally., Comment: Accepted for MNRAS. Missing reference added

This paper focuses on the opportunities and the ethical and societal risks posed by advanced AI assistants. We define advanced AI assistants as artificial agents with natural language interfaces, whose function is to plan and execute sequences of actions on behalf of a user, across one or more domains, in line with the user's expectations. The paper starts by considering the technology itself, providing an overview of AI assistants, their technical foundations and potential range of applications. It then explores questions around AI value alignment, well-being, safety and malicious uses. Extending the circle of inquiry further, we next consider the relationship between advanced AI assistants and individual users in more detail, exploring topics such as manipulation and persuasion, anthropomorphism, appropriate relationships, trust and privacy. With this analysis in place, we consider the deployment of advanced assistants at a societal scale, focusing on cooperation, equity and access, misinformation, economic impact, the environment and how best to evaluate advanced AI assistants. Finally, we conclude by providing a range of recommendations for researchers, developers, policymakers and public stakeholders.

Inclusive electron scattering at carefully chosen kinematics can isolate scattering from short-range correlations (SRCs), produced through hard, short-distance interactions of nucleons in the nucleus. Because the two-nucleon (2N) SRCs arise from the same N-N interaction in all nuclei, the cross section in the SRC-dominated regime is identical up to an overall scaling factor, and the A/2H cross section ratio is constant in this region. This scaling behavior has been used to identify SRC dominance and to map out the contribution of SRCs for a wide range of nuclei. We examine this scaling behavior at lower momentum transfers using new data on $^2$H, $^3$H, and $^3$He which show that the scaling region is larger than in heavy nuclei. Based on the improved scaling, especially for $^3$H/$^3$He, we examine the ratios at kinematics where three-nucleon SRCs may play an important role. The data for the largest initial nucleon momenta are consistent with isolation of scattering from 3N-SRCs, and suggest that the very-highest momentum nucleons in $^3$He have a nearly isospin-independent momentum configuration, or a small enhancement of the proton distribution.

Blue Large-Amplitude Pulsators (BLAPs) are a recently discovered class of short-period pulsating variable stars. In this work, we present new information on these stars based on photometric and spectroscopic data obtained for known and new objects detected by the OGLE survey. BLAPs are evolved objects with pulsation periods in the range of 3--75 min, stretching between subdwarf B-type stars and upper main-sequence stars in the Hertzsprung-Russell diagram. In general, BLAPs are single-mode stars pulsating in the fundamental radial mode. Their phase-folded light curves are typically sawtooth shaped, but light curves of shorter-period objects are more rounded and symmetric, while many longer-period objects exhibit an additional bump. The long-term OGLE observations show that the period change rates of BLAPs are usually of the order of 10^-7 per year and in a quarter of the sample are negative. An exception is the triple-mode object OGLE-BLAP-030, which changes its dominant period much faster, at a rate of about +4.6 x 10^-6 per year. The spectroscopic data indicate that the BLAPs form a homogeneous group in the period, surface gravity, and effective temperature spaces. However, we observe a split into two groups in terms of helium-to-hydrogen content. The atmospheres of the He-enriched BLAPs are more abundant in metals (about five times) than the atmosphere of the Sun. We discover that the BLAPs obey a period--gravity relationship and we use the distance to OGLE-BLAP-009 to derive a period--luminosity relation. Most of the stars observed in the OGLE Galactic bulge fields seem to reside in the bulge, while the remaining objects likely are in the foreground Galactic disk., Comment: submitted to ApJ, 21 pages, 14 figures

Fluoridoargentates(II) represent a fascinating class of silver(II) compounds that exhibit structural and magnetic similarities to cuprate superconductors. However, their synthesis is challenging, leaving their properties largely unexplored. In this study, mechanochemistry is introduced as a new technique for the synthesis of fluoridoargentates(II), which avoids the use of anhydrous HF or heating, and employs simple equipment. Furthermore, by ball milling of commercially available precursors, AgF and AgF2, we obtained the first two examples of binary mixed-valent silver(I,II) phases, AgI2AgIIF4 and AgIAgIIF3. While the AgI2AgIIF4 phase was obtained at room temperature, the AgIAgIIF3 phase is metastable and required milling under cryogenic conditions. Using synchrotron powder X-ray diffraction, it was found that AgI2AgIIF4 crystallizes in the P21/c space group and is isostructural to beta-K2AgF4. Here, double-bridged [AgF6]4- octahedra form chains that propagate along the a-crystallographic direction, giving a quasi-1D canted antiferromagnetic character, as shown by magnetic susceptibility. The AgIAgIIF phase adopts the P-1 space group, is isostructural to AgCuF3 and shows features of a one-dimensional antiferromagnet. It is expected that this facile synthetic approach will enable the expansion of silver(II) chemistry and accelerate the search for a silver analogue to cuprate superconductors., Comment: 28 pages, 16 figures, 4 tables

Training task-oriented dialogue systems typically requires turn-level annotations for interacting with their APIs: e.g. a dialogue state and the system actions taken at each step. These annotations can be costly to produce, error-prone, and require both domain and annotation expertise. With advances in LLMs, we hypothesize that unlabeled data and a schema definition are sufficient for building a working task-oriented dialogue system, completely unsupervised. We consider a novel unsupervised setting of only (1) a well-defined API schema (2) a set of unlabeled dialogues between a user and agent. We propose an innovative approach using expectation-maximization (EM) that infers turn-level annotations as latent variables using a noisy channel model to build an end-to-end dialogue agent. Evaluating our approach on the MultiWOZ benchmark, our method more than doubles the dialogue success rate of a strong GPT-3.5 baseline., Comment: To be presented at Empirical Methods in Natural Language Processing (EMNLP 2024). 18 Pages, 8 Figures

Doping of quantum antiferromagnets is an established approach to investigate the robustness of their ground state against the competing phases. Predictions of doping effects on the ground state of the Shastry-Sutherland dimer model are here verified experimentally on Mg-doped SrCu2(BO3)2. A partial incorporation of Mg2+ on the Cu2+-site in the SrCu2(BO3)2 structure leads to a subtle but systematic lattice expansion with the increasing Mg-doping concentration, which is accompanied by a concomitant decrease in the spin gap, the Curie-Weiss temperature and the peak temperature of the susceptibility. These findings indicate a doping-induced breaking of Cu2+ spin-1/2 dimers which is also corroborated by X-band EPR spectroscopy that points to a systematic increase in intensity of free Cu2+ sites with increasing Mg-doping concentration. Extending the Mg-doping up to nominal x = 0.10 or SrCu1.9Mg0.1(BO3)2, in the magnetisation measurements taken up to 35 T, a suppression of the pseudo-1/8 plateau is found along with a clear presence of an anomaly at an onset critical field H'C0 ~ 9 T. The latter, absent in pure SrCu2(BO3)2, emerges due to the coupling of liberated Cu2+ spin-1/2 entities in the vicinity of Mg-doping induced impurities., Comment: 17 pages, 6 figures

The homogeneous coordinate ring $\mathbb{C}[\operatorname{Gr}(k,n)]$ of the Grassmannian is a cluster algebra, with an additive categorification $\operatorname{CM}C$. Thus every $M\in\operatorname{CM}C$ has a cluster character $\Psi_M\in\mathbb{C}[\operatorname{Gr}(k,n)]$. The aim is to use the categorification to enrich Rietsch-Williams' mirror symmetry result that the Newton-Okounkov (NO) body/cone, made from leading exponents of functions in $\mathbb{C}[\operatorname{Gr}(k,n)]$ in an $\mathbb{X}$-cluster chart, can also be described by tropicalisation of the Marsh-Reitsch superpotential~$W$. For any cluster tilting object $T$, with endomorphism algebra $A$, we define two new cluster characters, a generalised partition function $\mathcal{P}^T_M\in\mathbb{C}[K(\operatorname{CM}A)]$ and a generalised flow polynomial $\mathcal{F}^T_M\in\mathbb{C}[K(\operatorname{fd}A)]$, related by a `dehomogenising' map $\operatorname{wt}\colon K(\operatorname{CM}A)\to K(\operatorname{fd}A)$. In the $\mathbb{X}$-cluster chart corresponding to $T$, the function $\Psi_M$ becomes $\mathcal{F}^T_M$ and thus its leading exponent is $\boldsymbol{\kappa}(T,M)$, an invariant introduced in earlier paper (and the image of the $g$-vector of $M$ under $\operatorname{wt}$). When $T$ mutates, $\mathcal{F}^T_M$ undergoes $\mathbb{X}$-mutation and $\boldsymbol{\kappa}(T,M)$ undergoes tropical $\mathbb{A}$-mutation. We then show that the monoid of $g$-vectors is saturated, and that this cone can be identified with the NO-cone, so the NO-body of Rietsch--Williams can be described in terms of $\boldsymbol{\kappa}(T,M)$. Furthermore, we adapt Rietsch-Williams' mirror symmetry strategy to find module-theoretic inequalities that determine the cone of $g$-vectors. Some of the machinery we develop works in a greater generality, which is relevant to the positroid subvarieties of $\operatorname{Gr}(k,n)$., Comment: 78 pages

On 24 September 2023, the NASA OSIRIS-REx mission dropped a capsule to Earth containing approximately 120 g of pristine carbonaceous regolith from Bennu. We describe the delivery and initial allocation of this asteroid sample and introduce its bulk physical, chemical, and mineralogical properties from early analyses. The regolith is very dark overall, with higher-reflectance inclusions and particles interspersed. Particle sizes range from sub-micron dust to a stone about 3.5 cm long. Millimeter-scale and larger stones typically have hummocky or angular morphologies. A subset of the stones appears mottled by brighter material that occurs as veins and crusts. Hummocky stones have the lowest densities and mottled stones have the highest. Remote sensing of the surface of Bennu detected hydrated phyllosilicates, magnetite, organic compounds, carbonates, and scarce anhydrous silicates, all of which the sample confirms. We also find sulfides, presolar grains, and, less expectedly, Na-rich phosphates, as well as other trace phases. The sample composition and mineralogy indicate substantial aqueous alteration and resemble those of Ryugu and the most chemically primitive, low-petrologic-type carbonaceous chondrites. Nevertheless, we find distinct hydrogen, nitrogen, and oxygen isotopic compositions, and some of the material we analyzed is enriched in fluid-mobile elements. Our findings underscore the value of sample return, especially for low-density material that may not readily survive atmospheric entry, and lay the groundwork for more comprehensive analyses., Comment: 73 pages, 22 figures

The tri-hypercharge proposal introduces a separate gauged weak hypercharge assigned to each fermion family as the origin of flavour. This is arguably one of the simplest setups for building "gauge non-universal theories of flavour" or "flavour deconstructed theories". In this paper we propose and study two minimal but ultraviolet complete and renormalisable tri-hypercharge models. We show that both models, which differ only by the heavy messengers that complete the effective theory, are able to explain the observed patterns of fermion masses and mixings (including neutrinos) with all fundamental coefficients being of $\mathcal{O}(1)$. In fact, both models translate the complicated flavour structure of the Standard Model into three simple physical scales above electroweak symmetry breaking, completely correlated with each other, that carry meaningful phenomenology. In particular, the heavy messenger sector determines the origin and size of fermion mixing, which controls the size and nature of the flavour-violating currents mediated by the two heavy $Z'$ gauge bosons of the theory. The phenomenological implications of the two minimal models are compared. In both models the lightest $Z'$ remains discoverable in dilepton searches at the LHC Run 3., Comment: 23 pages + Appendix, 2 figures. v2: Minor discussions added, conclusions unchanged. Published in JHEP

One of the leading mechanisms invoked to explain the existence of the radius valley is atmospheric mass loss driven by X-ray and extreme-ultraviolet irradiation, with this process stripping the primordial envelopes of young, small planets to produce the observed bimodal distribution. We present an investigation into the TOI-431 and $\nu^2$ Lupi planetary systems, both of which host planets either side of the radius valley, to determine if their architectures are consistent with evolution by the XUV mechanism. With $\textit{XMM-Newton}$, we measure the current X-ray flux of each star, and see evidence for a stellar flare in the TOI-431 observations. We then simulate the evolution of all of the transiting planets across the two systems in response to the high-energy irradiation over their lifetimes. We use the measured X-ray fluxes as an anchor point for the XUV time evolution in our simulations, and employ several different models of estimating mass loss rates. While the simulations for TOI-431b encountered a problem with the initial calculated radii, we estimate a likely short ($\sim$ Myr) timespan for primordial envelope removal using reasonable assumptions for the initial planet. $\nu^2$ Lupi b is likely harder to strip, but is achieved in a moderate fraction of our simulations. None of our simulations stripped any of the lower density planets of their envelope, in line with prediction. We conclude that both systems are consistent with expectations for generation of the radius valley through XUV photoevaporation., Comment: 14 pages, 12 figures, 5 tables; accepted for publication in MNRAS

We consider online model selection with decentralized data over $M$ clients, and study the necessity of collaboration among clients. Previous work proposed various federated algorithms without demonstrating their necessity,while we answer the question from a novel perspective of computational constraints. We prove lower bounds on the regret, and propose a federated algorithm and analyze the upper bound.Our results show (i) collaboration is unnecessary in the absence of computational constraints on clients; (ii) collaboration is necessary if the computational cost on each client is limited to $o(K)$, where $K$ is the number of candidate hypothesis spaces. We clarify the unnecessary nature of collaboration in previous federated algorithms for distributed online multi-kernel learning,and improve the regret bounds at a smaller computational and communication cost. Our algorithm relies on three new techniques including an improved Bernstein's inequality for martingale, a federated online mirror descent framework, and decoupling model selection and prediction, which might be of independent interest.

As an organic semiconductor and a prototypical acceptor molecule in organic photovoltaics, C$_{60}$ has broad relevance to the world of organic thin film electronics. Although highly uniform C$_{60}$ thin films are necessary to conduct spectroscopic analysis of the electronic structure of these C$_{60}$-based materials, reported C$_{60}$ films show a relatively low degree of order beyond a monolayer. Here, we develop a generalizable two-stage growth technique that consistently produces single-domain C$_{60}$ films of controllable thicknesses, using Au(111) as an epitaxially well-matched substrate. We characterize the films using low-energy electron diffraction, low-energy electron microscopy, scanning tunneling microscopy, and angle-resolved photoemission spectroscopy (ARPES). We report highly oriented epitaxial film growth of C$_{60}$/Au(111) from 1 monolayer (ML) up to 20 ML films. The high-quality of the C$_{60}$ thin films enables the direct observation of the electronic dispersion of the HOMO and HOMO-1 bands via ARPES without need for small spot sizes. Our results indicate a path for the growth of organic films on metallic substrates with long-range ordering., Comment: 16 pages, 7 figures

Accurate explicit and implicit product identification in search queries is critical for enhancing user experiences, especially at a company like Adobe which has over 50 products and covers queries across hundreds of tools. In this work, we present a novel approach to training a product classifier from user behavioral data. Our semantic model led to >25% relative improvement in CTR (click through rate) across the deployed surfaces; a >50% decrease in null rate; a 2x increase in the app cards surfaced, which helps drive product visibility.

The Knowledge Graph Entity Typing (KGET) task aims to predict missing type annotations for entities in knowledge graphs. Recent works only utilize the \textit{\textbf{structural knowledge}} in the local neighborhood of entities, disregarding \textit{\textbf{semantic knowledge}} in the textual representations of entities, relations, and types that are also crucial for type inference. Additionally, we observe that the interaction between semantic and structural knowledge can be utilized to address the false-negative problem. In this paper, we propose a novel \textbf{\underline{S}}emantic and \textbf{\underline{S}}tructure-aware KG \textbf{\underline{E}}ntity \textbf{\underline{T}}yping~{(SSET)} framework, which is composed of three modules. First, the \textit{Semantic Knowledge Encoding} module encodes factual knowledge in the KG with a Masked Entity Typing task. Then, the \textit{Structural Knowledge Aggregation} module aggregates knowledge from the multi-hop neighborhood of entities to infer missing types. Finally, the \textit{Unsupervised Type Re-ranking} module utilizes the inference results from the two models above to generate type predictions that are robust to false-negative samples. Extensive experiments show that SSET significantly outperforms existing state-of-the-art methods., Comment: Accepted in NAACL2024 main

Knowledge graphs are useful tools to organize, recommend and sort data. Hierarchies in knowledge graphs provide significant benefit in improving understanding and compartmentalization of the data within a knowledge graph. This work leverages large language models to generate and augment hierarchies in an existing knowledge graph. For small (<100,000 node) domain-specific KGs, we find that a combination of few-shot prompting with one-shot generation works well, while larger KG may require cyclical generation. We present techniques for augmenting hierarchies, which led to coverage increase by 98% for intents and 99% for colors in our knowledge graph., Comment: European Conference on Information Retrieval 2024

Battery recycling is a critical process for minimizing environmental harm and resource waste for used batteries. However, it is challenging, largely because sorting batteries is costly and hardly automated to group batteries based on battery types. In this paper, we introduce a machine learning-based approach for battery-type classification and address the daunting problem of data scarcity for the application. We propose BatSort which applies transfer learning to utilize the existing knowledge optimized with large-scale datasets and customizes ResNet to be specialized for classifying battery types. We collected our in-house battery-type dataset of small-scale to guide the knowledge transfer as a case study and evaluate the system performance. We conducted an experimental study and the results show that BatSort can achieve outstanding accuracy of 92.1% on average and up to 96.2% and the performance is stable for battery-type classification. Our solution helps realize fast and automated battery sorting with minimized cost and can be transferred to related industry applications with insufficient data.

The flavor structure of quarks and leptons and quark-lepton unification are studied in the framework of Pati-Salam models with $A_4$ modular symmetry. The three generations of the left-handed and right-handed fermions are assigned to be triplet or singlets of $A_4$. The light neutrino masses are generated through the type-I seesaw mechanism. We perform a systematic classification of Pati-Salam models according to the transformations of matter fields under the $A_4$ modular symmetry, and the general form of the fermion mass matrix is given. We present four phenomenologically viable benchmark models which provide excellent descriptions of masses and flavor mixing of quarks and leptons, including neutrinos. In such models we find that the normal ordered neutrino mass spectrum is preferred over the inverted case, with neutrinoless double beta decay predicted to be too small to be observed by the next generation of experiments., Comment: 33 pages, 10 figures

Recent work has shown a variety of ways in which machine learning can be used to accelerate the solution of constrained optimization problems. Increasing demand for real-time decision-making capabilities in applications such as artificial intelligence and optimal control has led to a variety of approaches, based on distinct strategies. This work proposes a novel approach to learning optimization, in which the underlying metric space of a proximal operator splitting algorithm is learned so as to maximize its convergence rate. While prior works in optimization theory have derived optimal metrics for limited classes of problems, the results do not extend to many practical problem forms including general Quadratic Programming (QP). This paper shows how differentiable optimization can enable the end-to-end learning of proximal metrics, enhancing the convergence of proximal algorithms for QP problems beyond what is possible based on known theory. Additionally, the results illustrate a strong connection between the learned proximal metrics and active constraints at the optima, leading to an interpretation in which the learning of proximal metrics can be viewed as a form of active set learning.