Search

Despite seemingly performant web agents on the task-completion benchmarks, most existing methods evaluate the agents based on a presupposition: the web navigation task consists of linear sequence of actions with an end state that marks task completion. In contrast, our work focuses on web navigation for information aggregation, wherein the agent must explore different websites to gather information for a complex query. We consider web information aggregation from two different perspectives: (i) Direct API-driven Access relies on a text-only view of the Web, leveraging external tools such as Google Search API to navigate the web and a scraper to extract website contents. (ii) Interactive Visual Access uses screenshots of the webpages and requires interaction with the browser to navigate and access information. Motivated by these diverse information access settings, we introduce Infogent, a novel modular framework for web information aggregation involving three distinct components: Navigator, Extractor and Aggregator. Experiments on different information access settings demonstrate Infogent beats an existing SOTA multi-agent search framework by 7% under Direct API-Driven Access on FRAMES, and improves over an existing information-seeking web agent by 4.3% under Interactive Visual Access on AssistantBench., Comment: Preprint

Complex news events, such as natural disasters and socio-political conflicts, require swift responses from the government and society. Relying on historical events to project the future is insufficient as such events are sparse and do not cover all possible conditions and nuanced situations. Simulation of these complex events can help better prepare and reduce the negative impact. We develop a controllable complex news event simulator guided by both the event schema representing domain knowledge about the scenario and user-provided assumptions representing case-specific conditions. As event dynamics depend on the fine-grained social and cultural context, we further introduce a geo-diverse commonsense and cultural norm-aware knowledge enhancement component. To enhance the coherence of the simulation, apart from the global timeline of events, we take an agent-based approach to simulate the individual character states, plans, and actions. By incorporating the schema and cultural norms, our generated simulations achieve much higher coherence and appropriateness and are received favorably by participants from a humanitarian assistance organization., Comment: Accepted as EMNLP 2024 Demo

The XLZD collaboration is developing a two-phase xenon time projection chamber with an active mass of 60 to 80 t capable of probing the remaining WIMP-nucleon interaction parameter space down to the so-called neutrino fog. In this work we show that, based on the performance of currently operating detectors using the same technology and a realistic reduction of radioactivity in detector materials, such an experiment will also be able to competitively search for neutrinoless double beta decay in $^{136}$Xe using a natural-abundance xenon target. XLZD can reach a 3$\sigma$ discovery potential half-life of 5.7$\times$10$^{27}$ yr (and a 90% CL exclusion of 1.3$\times$10$^{28}$ yr) with 10 years of data taking, corresponding to a Majorana mass range of 7.3-31.3 meV (4.8-20.5 meV). XLZD will thus exclude the inverted neutrino mass ordering parameter space and will start to probe the normal ordering region for most of the nuclear matrix elements commonly considered by the community., Comment: 29 pages, 7 figures

This report describes the experimental strategy and technologies for a next-generation xenon observatory sensitive to dark matter and neutrino physics. The detector will have an active liquid xenon target mass of 60-80 tonnes and is proposed by the XENON-LUX-ZEPLIN-DARWIN (XLZD) collaboration. The design is based on the mature liquid xenon time projection chamber technology of the current-generation experiments, LZ and XENONnT. A baseline design and opportunities for further optimization of the individual detector components are discussed. The experiment envisaged here has the capability to explore parameter space for Weakly Interacting Massive Particle (WIMP) dark matter down to the neutrino fog, with a 3$\sigma$ evidence potential for the spin-independent WIMP-nucleon cross sections as low as $3\times10^{-49}\rm cm^2$ (at 40 GeV/c$^2$ WIMP mass). The observatory is also projected to have a 3$\sigma$ observation potential of neutrinoless double-beta decay of $^{136}$Xe at a half-life of up to $5.7\times 10^{27}$ years. Additionally, it is sensitive to astrophysical neutrinos from the atmosphere, sun, and galactic supernovae., Comment: 32 pages, 14 figures

We present a novel deep learning pipeline to perform a model-independent, likelihood-free search for anomalous (i.e., non-background) events in the proposed next generation multi-ton scale liquid Xenon-based direct detection experiment, DARWIN. We train an anomaly detector comprising a variational autoencoder and a classifier on extensive, high-dimensional simulated detector response data and construct a one-dimensional anomaly score optimised to reject the background only hypothesis in the presence of an excess of non-background-like events. We benchmark the procedure with a sensitivity study that determines its power to reject the background-only hypothesis in the presence of an injected WIMP dark matter signal, outperforming the classical, likelihood-based background rejection test. We show that our neural networks learn relevant energy features of the events from low-level, high-dimensional detector outputs, without the need to compress this data into lower-dimensional observables, thus reducing computational effort and information loss. For the future, our approach lays the foundation for an efficient end-to-end pipeline that eliminates the need for many of the corrections and cuts that are traditionally part of the analysis chain, with the potential of achieving higher accuracy and significant reduction of analysis time., Comment: 10 Figures, 3 Tables, 23 Pages (incl. references)

In this paper, we introduce SearchDet, a training-free long-tail object detection framework that significantly enhances open-vocabulary object detection performance. SearchDet retrieves a set of positive and negative images of an object to ground, embeds these images, and computes an input image-weighted query which is used to detect the desired concept in the image. Our proposed method is simple and training-free, yet achieves over 48.7% mAP improvement on ODinW and 59.1% mAP improvement on LVIS compared to state-of-the-art models such as GroundingDINO. We further show that our approach of basing object detection on a set of Web-retrieved exemplars is stable with respect to variations in the exemplars, suggesting a path towards eliminating costly data annotation and training procedures.

We search for dark matter (DM) with a mass [3,12] $\mathrm{GeV} / c^2$ using an exposure of 3.51 $\mathrm{t} \times \mathrm{y}$ with the XENONnT experiment. We consider spin-independent, spin-dependent, momentum-dependent, mirror DM, and self-interacting DM with a light mediator coupling to Standard Model particles. Using a lowered energy threshold compared to the previous WIMP search, a blind analysis of [0.5, 5.0] $\mathrm{keV}$ nuclear recoil events reveals no significant signal excess over the background. XENONnT excludes spin-independent DM-nucleon cross sections $>2.5 \times 10^{-45} \mathrm{~cm}^2$ at $90 \%$ confidence level for 6 $\mathrm{GeV} / c^2$ DM. The solar ${ }^8 \mathrm{B}$ neutrino coherent elastic neutrino-nucleus scattering background accounts for approximately half of the background in the signal region. In the considered mass range, the DM sensitivity approaches the 'neutrino fog', the limitation where neutrinos produce a signal that is indistinguishable from that of light DM-xenon nucleus scattering.

The XENONnT experiment, located at the INFN Laboratori Nazionali del Gran Sasso, Italy, features a 5.9 tonne liquid xenon time projection chamber surrounded by an instrumented neutron veto, all of which is housed within a muon veto water tank. Due to extensive shielding and advanced purification to mitigate natural radioactivity, an exceptionally low background level of (15.8 $\pm$ 1.3) events/(tonne$\cdot$year$\cdot$keV) in the (1, 30) keV region is reached in the inner part of the TPC. XENONnT is thus sensitive to a wide range of rare phenomena related to Dark Matter and Neutrino interactions, both within and beyond the Standard Model of particle physics, with a focus on the direct detection of Dark Matter in the form of weakly interacting massive particles (WIMPs). From May 2021 to December 2021, XENONnT accumulated data in rare-event search mode with a total exposure of one tonne $\cdot$ year. This paper provides a detailed description of the signal reconstruction methods, event selection procedure, and detector response calibration, as well as an overview of the detector performance in this time frame. This work establishes the foundational framework for the `blind analysis' methodology we are using when reporting XENONnT physics results., Comment: 27 pages, 23 figures

The aim of this work is to characterize the accretion process of the classical T Tauri Star RU Lup. We studied optical high-resolution spectroscopic observations from CHIRON and ESPRESSO, obtained simultaneously with photometric data from AAVSO and TESS. We detected a periodic modulation in the narrow component of the He I 5876 line with a period that is compatible with the stellar rotation period, indicating the presence of a compact region on the stellar surface that we identified as the footprint of the accretion shock. We show that this region is responsible for the veiling spectrum, which is made up of a continuum component plus narrow line emission. An analysis of the high-cadence TESS light curve reveals quasi-periodic oscillations on timescales shorter than the stellar rotation period, suggesting that the accretion disk in RU~Lup extends inward of the corotation radius, with a truncation radius at $\sim 2 ~ R_{\star}$. This is compatible with predictions from three-dimensional magnetohydrodynamic models of accretion through a magnetic boundary layer (MBL). In this scenario, the photometric variability of RU Lup is produced by a nonstationary hot spot on the stellar surface that rotates with the Keplerian period at the truncation radius. The analysis of the broad components of selected emission lines reveals the existence of a non-axisymmetric, temperature-stratified flow around the star, in which the gas leaves the accretion disk at the truncation radius and accretes onto the star channeled by the magnetic field lines. The unusually rich metallic emission line spectrum of RU Lup might be characteristic of the MBL regime of accretion. In conclusion, the behavior of RU Lup reveals many similarities to predictions from the MBL accretion scenario. Alternative explanations would require the existence of a hot spot with a complex shape, or a warped structure in the inner disk., Comment: 20 pages, 21 figures. Accepted for publication in A&A

Classical T Tauri Stars (CTTSs) are highly variable stars that possess gas- and dust-rich disks from which planets form. Much of their variability is driven by mass accretion from the surrounding disk, a process that is still not entirely understood. A multi-epoch optical spectral monitoring campaign of four CTTSs (TW Hya, RU Lup, BP Tau, and GM Aur) was conducted along with contemporaneous HST UV spectra and ground-based photometry in an effort to determine accretion characteristics and gauge variability in this sample. Using an accretion flow model, we find that the magnetospheric truncation radius varies between 2.5-5 R* across all of our observations. There is also significant variability in all emission lines studied, particularly Halpha, Hbeta, and Hgamma. Using previously established relationships between line luminosity and accretion, we find that, on average, most lines reproduce accretion rates consistent with accretion shock modeling of HST spectra to within 0.5 dex. Looking at individual contemporaneous observations, however, these relationships are less accurate, suggesting that variability trends differ from the trends of the population and that these empirical relationships should be used with caution in studies of variability., Comment: 39 pages, 10 figures, 12 tables

We present the first measurement of nuclear recoils from solar $^8$B neutrinos via coherent elastic neutrino-nucleus scattering with the XENONnT dark matter experiment. The central detector of XENONnT is a low-background, two-phase time projection chamber with a 5.9\,t sensitive liquid xenon target. A blind analysis with an exposure of 3.51\,t$\times$y resulted in 37 observed events above 0.5\,keV, with ($26.4^{+1.4}_{-1.3}$) events expected from backgrounds. The background-only hypothesis is rejected with a statistical significance of 2.73\,$\sigma$. The measured $^8$B solar neutrino flux of $(4.7_{-2.3}^{+3.6})\times 10^6\,\mathrm{cm}^{-2}\mathrm{s}^{-1}$ is consistent with results from dedicated solar neutrino experiments. The measured neutrino flux-weighted CE$\nu$NS cross-section on Xe of $(1.1^{+0.8}_{-0.5})\times10^{-39}\,\mathrm{cm}^2$ is consistent with the Standard Model prediction. This is the first direct measurement of nuclear recoils from solar neutrinos with a dark matter detector.

LLM-driven dialog systems are used in a diverse set of applications, ranging from healthcare to customer service. However, given their generalization capability, it is difficult to ensure that these chatbots stay within the boundaries of the specialized domains, potentially resulting in inaccurate information and irrelevant responses. This paper introduces an unsupervised approach for automatically inducing domain-specific dialog flows that can be used to constrain LLM-based chatbots. We introduce two variants of dialog flow based on the availability of in-domain conversation instances. Through human and automatic evaluation over various dialog domains, we demonstrate that our high-quality data-guided dialog flows achieve better domain coverage, thereby overcoming the need for extensive manual crafting of such flows., Comment: Accepted at SIGDIAL 2024

Seaports play a crucial role in the global economy, and researchers have sought to understand their significance through various studies. In this paper, we aim to explore the common characteristics shared by important ports by analyzing the network of connections formed by vessel movement among them. To accomplish this task, we adopt a bottom-up network construction approach that combines three years' worth of AIS (Automatic Identification System) data from around the world, constructing a Ports Network that represents the connections between different ports. Through such representation, we use machine learning to measure the relative significance of different port features. Our model examined such features and revealed that geographical characteristics and the depth of the port are indicators of a port's significance to the Ports Network. Accordingly, this study employs a data-driven approach and utilizes machine learning to provide a comprehensive understanding of the factors contributing to ports' importance. The outcomes of our work are aimed to inform decision-making processes related to port development, resource allocation, and infrastructure planning in the industry.

Photonic computing has emerged as a promising solution for accelerating computation-intensive artificial intelligence (AI) workloads. However, limited reconfigurability, high electrical-optical conversion cost, and thermal sensitivity limit the deployment of current optical analog computing engines to support power-restricted, performance-sensitive AI workloads at scale. Sparsity provides a great opportunity for hardware-efficient AI accelerators. However, current dense photonic accelerators fail to fully exploit the power-saving potential of algorithmic sparsity. It requires sparsity-aware hardware specialization with a fundamental re-design of photonic tensor core topology and cross-layer device-circuit-architecture-algorithm co-optimization aware of hardware non-ideality and power bottleneck. To trim down the redundant power consumption while maximizing robustness to thermal variations, we propose SCATTER, a novel algorithm-circuit co-sparse photonic accelerator featuring dynamically reconfigurable signal path via thermal-tolerant, power-efficient in-situ light redistribution and power gating. A power-optimized, crosstalk-aware dynamic sparse training framework is introduced to explore row-column structured sparsity and ensure marginal accuracy loss and maximum power efficiency. The extensive evaluation shows that our cross-stacked optimized accelerator SCATTER achieves a 511X area reduction and 12.4X power saving with superior crosstalk tolerance that enables unprecedented circuit layout compactness and on-chip power efficiency.

Large Language Models (LLMs) have significantly advanced the field of information retrieval, particularly for reranking. Listwise LLM rerankers have showcased superior performance and generalizability compared to existing supervised approaches. However, conventional listwise LLM reranking methods lack efficiency as they provide ranking output in the form of a generated ordered sequence of candidate passage identifiers. Further, they are trained with the typical language modeling objective, which treats all ranking errors uniformly--potentially at the cost of misranking highly relevant passages. Addressing these limitations, we introduce FIRST, a novel listwise LLM reranking approach leveraging the output logits of the first generated identifier to directly obtain a ranked ordering of the candidates. Further, we incorporate a learning-to-rank loss during training, prioritizing ranking accuracy for the more relevant passages. Empirical results demonstrate that FIRST accelerates inference by 50% while maintaining a robust ranking performance with gains across the BEIR benchmark. Finally, to illustrate the practical effectiveness of listwise LLM rerankers, we investigate their application in providing relevance feedback for retrievers during inference. Our results show that LLM rerankers can provide a stronger distillation signal compared to cross-encoders, yielding substantial improvements in retriever recall after relevance feedback., Comment: Preprint

The XENONnT experiment searches for weakly-interacting massive particle (WIMP) dark matter scattering off a xenon nucleus. In particular, XENONnT uses a dual-phase time projection chamber with a 5.9-tonne liquid xenon target, detecting both scintillation and ionization signals to reconstruct the energy, position, and type of recoil. A blind search for nuclear recoil WIMPs with an exposure of 1.1 tonne-years yielded no signal excess over background expectations, from which competitive exclusion limits were derived on WIMP-nucleon elastic scatter cross sections, for WIMP masses ranging from 6 GeV/$c^2$ up to the TeV/$c^2$ scale. This work details the modeling and statistical methods employed in this search. By means of calibration data, we model the detector response, which is then used to derive background and signal models. The construction and validation of these models is discussed, alongside additional purely data-driven backgrounds. We also describe the statistical inference framework, including the definition of the likelihood function and the construction of confidence intervals., Comment: 20 pages, 10 figures

Ranking is a fundamental and popular problem in search. However, existing ranking algorithms usually restrict the granularity of ranking to full passages or require a specific dense index for each desired level of granularity. Such lack of flexibility in granularity negatively affects many applications that can benefit from more granular ranking, such as sentence-level ranking for open-domain question-answering, or proposition-level ranking for attribution. In this work, we introduce the idea of any-granularity ranking, which leverages multi-vector embeddings to rank at varying levels of granularity while maintaining encoding at a single (coarser) level of granularity. We propose a multi-granular contrastive loss for training multi-vector approaches, and validate its utility with both sentences and propositions as ranking units. Finally, we demonstrate the application of proposition-level ranking to post-hoc citation addition in retrieval-augmented generation, surpassing the performance of prompt-driven citation generation.

In this research, we developed a low-power silicon photonics foundry-fabricated slow-light thermal phase shifter (SLTPS) where the slow-light (SL) effect is achieved using an integrated Bragg grating (BG) waveguide. Heating the grating induces a red shift in the transmission spectrum, leading to an increased group index $n_g$ during operation, which facilitates a further reduction in the voltage needed for a $\pi$ phase shift, i.e. $V_{\pi}$. Additionally, we investigated a compact Mach-Zehnder Interferometer (MZI) that incorporates the SLTPS in both arms with a phase shifter length of 50 $\mu$m. A detailed theoretical analysis was conducted to address the non-idealities of the SL-MZI due to uneven optical power splitting and unbalanced loss in the two MZI arms. The $V_{\pi}$ and power consumption for a $\pi$ phase shift $(P_{\pi})$ of the SL-MZI were quantified for operation in the slow light regime, demonstrating a $V_{\pi}$ of 1.1 V and a $P_{\pi}$ of 3.63 mW at an operational wavelength near the photonic band edge. The figure of merit (FOM) $P_{\pi} \times \tau$ is commonly used to assess the performance of thermal optical switches. The SL-MZI in this work has achieved a low $P_{\pi} \times \tau$ of 5.1 mW $\mu$s. Insertion loss of the SL-MZI ranges from 1.3 dB to 4.4 dB depending on the operation wavelength, indicating a trade-off with the $V_\pi$ reduction., Comment: 14 pages, 9 figures, 1 table

On March 6 2023, the Gaia telescope has alerted a 2-magnitude burst from Gaia23bab, a Young Stellar Object in the Galactic plane. We observed Gaia23bab with the Large Binocular Telescope obtaining optical and near-infrared spectra close in time to the peak of the burst, and collected all public multi-band photometry to reconstruct the historical light curve. This latter shows three bursts in ten years (2013, 2017 and 2023), whose duration and amplitude are typical of EXor variables. We estimate that, due to the bursts, the mass accumulated on the star is about twice greater than if the source had remained quiescent for the same period of time. Photometric analysis indicates that Gaia23bab is a Class,II source with age < 1 Myr, spectral type G3-K0, stellar luminosity 4.0 L_sun, and mass 1.6 M_sun. The optical/near infrared spectrum is rich in emission lines. From the analysis of these lines we measured the accretion luminosity and the mass accretion rate L_acc(burst)=3.7 L_sun, M_acc(burst) 2.0 10 $^(-7) M_sun/yr, consistent with those of EXors. More generally, we derive the relationships between accretion and stellar parameters in a sample of EXors. We find that, when in burst, the accretion parameters become almost independent of the stellar parameters and that EXors, even in quiescence, are more efficient than classical T Tauri stars in assembling mass., Comment: Accepted by The Astrophysical Journal

Open-domain Question Answering (OpenQA) aims at answering factual questions with an external large-scale knowledge corpus. However, real-world knowledge is not static; it updates and evolves continually. Such a dynamic characteristic of knowledge poses a vital challenge for these models, as the trained models need to constantly adapt to the latest information to make sure that the answers remain accurate. In addition, it is still unclear how well an OpenQA model can transfer to completely new knowledge domains. In this paper, we investigate the generalization performance of a retrieval-augmented QA model in two specific scenarios: 1) adapting to updated versions of the same knowledge corpus; 2) switching to completely different knowledge domains. We observe that the generalization challenges of OpenQA models stem from the reader's over-reliance on memorizing the knowledge from the external corpus, which hinders the model from generalizing to a new knowledge corpus. We introduce Corpus-Invariant Tuning (CIT), a simple but effective training strategy, to mitigate the knowledge over-memorization by controlling the likelihood of retrieved contexts during training. Extensive experimental results on multiple OpenQA benchmarks show that CIT achieves significantly better generalizability without compromising the model's performance in its original corpus and domain., Comment: Accepted to NAACL 2024 Findings

This paper details the first application of a software tagging algorithm to reduce radon-induced backgrounds in liquid noble element time projection chambers, such as XENON1T and XENONnT. The convection velocity field in XENON1T was mapped out using $^{222}\text{Rn}$ and $^{218}\text{Po}$ events, and the root-mean-square convection speed was measured to be $0.30 \pm 0.01$ cm/s. Given this velocity field, $^{214}\text{Pb}$ background events can be tagged when they are followed by $^{214}\text{Bi}$ and $^{214}\text{Po}$ decays, or preceded by $^{218}\text{Po}$ decays. This was achieved by evolving a point cloud in the direction of a measured convection velocity field, and searching for $^{214}\text{Bi}$ and $^{214}\text{Po}$ decays or $^{218}\text{Po}$ decays within a volume defined by the point cloud. In XENON1T, this tagging system achieved a $^{214}\text{Pb}$ background reduction of $6.2^{+0.4}_{-0.9}\%$ with an exposure loss of $1.8\pm 0.2 \%$, despite the timescales of convection being smaller than the relevant decay times. We show that the performance can be improved in XENONnT, and that the performance of such a software-tagging approach can be expected to be further improved in a diffusion-limited scenario. Finally, a similar method might be useful to tag the cosmogenic $^{137}\text{Xe}$ background, which is relevant to the search for neutrinoless double-beta decay., Comment: 17 pages, 19 figures

Optimization of product and system characteristics is required in many fields, including design and control. Bayesian optimization (BO) is often used when there are high observing costs, because BO theoretically guarantees an upper bound on regret. However, computational costs increase exponentially with the number of parameters to be optimized, decreasing search efficiency. We propose a BO that limits the search region to lower dimensions and utilizes local Gaussian process regression (LGPR) to scale the BO to higher dimensions. LGPR treats the low-dimensional search region as "local," improving prediction accuracies there. The LGPR model is trained on a local subset of data specific to that region. This improves prediction accuracy and search efficiency and reduces the time complexity of matrix inversion in the Gaussian process regression. In evaluations with 20D Ackley and Rosenbrock functions, search efficiencies are equal to or higher than those of the compared methods, improved by about 69% and 40% from the case without LGPR. We apply our method to an automatic design task for a power semiconductor device. We successfully reduce the specific on-resistance to 25% better than a conventional method and 3.4% better than without LGPR., Comment: 8 pages, 13 figures, 22nd International Conference on Machine Learning and Applications (ICMLA2023)

The increasing demand for personalized interactions with large language models (LLMs) calls for methodologies capable of accurately and efficiently identifying user opinions and preferences. Retrieval augmentation emerges as an effective strategy, as it can accommodate a vast number of users without the costs from fine-tuning. Existing research, however, has largely focused on enhancing the retrieval stage and devoted limited exploration toward optimizing the representation of the database, a crucial aspect for tasks such as personalization. In this work, we examine the problem from a novel angle, focusing on how data can be better represented for more data-efficient retrieval in the context of LLM customization. To tackle this challenge, we introduce Persona-DB, a simple yet effective framework consisting of a hierarchical construction process to improve generalization across task contexts and collaborative refinement to effectively bridge knowledge gaps among users. In the evaluation of response prediction, Persona-DB demonstrates superior context efficiency in maintaining accuracy with a significantly reduced retrieval size, a critical advantage in scenarios with extensive histories or limited context windows. Our experiments also indicate a marked improvement of over 10% under cold-start scenarios, when users have extremely sparse data. Furthermore, our analysis reveals the increasing importance of collaborative knowledge as the retrieval capacity expands.

The multi-staged XENON program at INFN Laboratori Nazionali del Gran Sasso aims to detect dark matter with two-phase liquid xenon time projection chambers of increasing size and sensitivity. The XENONnT experiment is the latest detector in the program, planned to be an upgrade of its predecessor XENON1T. It features an active target of 5.9 tonnes of cryogenic liquid xenon (8.5 tonnes total mass in cryostat). The experiment is expected to extend the sensitivity to WIMP dark matter by more than an order of magnitude compared to XENON1T, thanks to the larger active mass and the significantly reduced background, improved by novel systems such as a radon removal plant and a neutron veto. This article describes the XENONnT experiment and its sub-systems in detail and reports on the detector performance during the first science run., Comment: 32 pages, 19 figures

Electronic-photonic computing systems offer immense potential in energy-efficient artificial intelligence (AI) acceleration tasks due to the superior computing speed and efficiency of optics, especially for real-time, low-energy deep neural network (DNN) inference tasks on resource-restricted edge platforms. However, current optical neural accelerators based on foundry-available devices and conventional system architecture still encounter a performance gap compared to highly customized electronic counterparts. To bridge the performance gap due to lack of domain specialization, we present a time-multiplexed dynamic photonic tensor accelerator, dubbed TeMPO, with cross-layer device/circuit/architecture customization. At the device level, we present foundry-compatible, customized photonic devices, including a slow-light electro-optic modulator with experimental demonstration, optical splitters, and phase shifters that significantly reduce the footprint and power in input encoding and dot-product calculation. At the circuit level, partial products are hierarchically accumulated via parallel photocurrent aggregation, lightweight capacitive temporal integration, and sequential digital summation, considerably relieving the analog-to-digital conversion bottleneck. We also employ a multi-tile, multi-core architecture to maximize hardware sharing for higher efficiency. Across diverse edge AI workloads, TeMPO delivers digital-comparable task accuracy with superior quantization/noise tolerance. We achieve a 368.6 TOPS peak performance, 22.3 TOPS/W energy efficiency, and 1.2 TOPS/mm$^2$ compute density, pushing the Pareto frontier in edge AI hardware. This work signifies the power of cross-layer co-design and domain-specific customization, paving the way for future electronic-photonic accelerators with even greater performance and efficiency., Comment: 17 pages, 19 figures

This paper aims at revisit the physical and dynamical properties of the warm atomic gas in the inner disk region of classical T Tauri stars (CTTs) and relate them to the properties of the outer dusty disk. We used the high resolution (R=115,000) spectra of 36 CTTs observed as part of the GHOsT project and analysed the profile and luminosity of the brightest optical forbidden lines, namely [OI]630 and 557nm, [SII]406 and 673nm, and [NII]658nm. We find that in about 40% of sources the so-called narrow low-velocity component (NLVC) display a peak velocity compatible with the stellar velocity. In these sources, that typically show lower mass accretion rates and the absence of a jet, the [OI]630nm profiles are well fitted by a simple Keplerian disk model, indicating that the emission from the disk is dominant with respect to the wind contribution. For transitional disks (TD), no correlation is found between $R_{kep}$, derived from the line HWHM, and the size of the dust cavity. We also see an anti-correlation between the [OI] 557/630 nm ratio and $R_{kep}$, which suggests that the [OI] emitting region expands as the gas cools and becomes less dense. We confirmed previous findings on the density and temperature ranges implied by the line ratios, and additionally constrained the ionisation fraction in the NLVC to be < 0.1. We however discuss the limits of applying this diagnostic to winds that are not spatially resolved. For the outflow component, we estimated the mass-loss for both the disk winds and jets and compared the results with X-ray photoevaporative models. We conclude that without better knowledge of the wind geometry, and given the limitation of the diagnostics, the mass-loss in the wind traced by the LVC cannot be constrained better than a factor of 100, with a mass-loss/mass-accretion ratio spanning between ~ 0.01 and more than 1., Comment: Accepted for publication in Astronomy and Astrophysics

The increased use of large language models (LLMs) across a variety of real-world applications calls for mechanisms to verify the factual accuracy of their outputs. In this work, we present a holistic end-to-end solution for annotating the factuality of LLM-generated responses, which encompasses a multi-stage annotation scheme designed to yield detailed labels concerning the verifiability and factual inconsistencies found in LLM outputs. We further construct an open-domain document-level factuality benchmark in three-level granularity: claim, sentence and document, aiming to facilitate the evaluation of automatic fact-checking systems. Preliminary experiments show that FacTool, FactScore and Perplexity.ai are struggling to identify false claims, with the best F1=0.63 by this annotation solution based on GPT-4. Annotation tool, benchmark and code are available at https://github.com/yuxiaw/Factcheck-GPT., Comment: 30 pages, 13 figures

Open-domain dialog involves generating search queries that help obtain relevant knowledge for holding informative conversations. However, it can be challenging to determine what information to retrieve when the user is passive and does not express a clear need or request. To tackle this issue, we present a novel approach that focuses on generating internet search queries that are guided by social commonsense. Specifically, we leverage a commonsense dialog system to establish connections related to the conversation topic, which subsequently guides our query generation. Our proposed framework addresses passive user interactions by integrating topic tracking, commonsense response generation and instruction-driven query generation. Through extensive evaluations, we show that our approach overcomes limitations of existing query generation techniques that rely solely on explicit dialog information, and produces search queries that are more relevant, specific, and compelling, ultimately resulting in more engaging responses., Comment: Accepted in EMNLP 2023 Findings

The precision in reconstructing events detected in a dual-phase time projection chamber depends on an homogeneous and well understood electric field within the liquid target. In the XENONnT TPC the field homogeneity is achieved through a double-array field cage, consisting of two nested arrays of field shaping rings connected by an easily accessible resistor chain. Rather than being connected to the gate electrode, the topmost field shaping ring is independently biased, adding a degree of freedom to tune the electric field during operation. Two-dimensional finite element simulations were used to optimize the field cage, as well as its operation. Simulation results were compared to ${}^{83m}\mathrm{Kr}$ calibration data. This comparison indicates an accumulation of charge on the panels of the TPC which is constant over time, as no evolution of the reconstructed position distribution of events is observed. The simulated electric field was then used to correct the charge signal for the field dependence of the charge yield. This correction resolves the inconsistent measurement of the drift electron lifetime when using different calibrations sources and different field cage tuning voltages.

HM Lup is a young M-type star that accretes material from a circumstellar disk through a magnetosphere. Our aim is to study the inner disk structure of HM Lup and to characterize its variability. We used spectroscopic data from HST/STIS, X-Shooter, and ESPRESSO taken in the framework of the ULLYSES and PENELLOPE programs, together with photometric data from TESS and AAVSO. The 2021 TESS light curve shows variability typical for young stellar objects of the "accretion burster" type. The spectra cover the temporal evolution of the main burst in the 2021 TESS light curve. We compared the strength and morphology of emission lines from different species and ionization stages. We determined the mass accretion rate from selected emission lines and from the UV continuum excess emission at different epochs, and we examined its relation to the photometric light curves. The emission lines in the optical spectrum of HM Lup delineate a temperature stratification along the accretion flow. While the wings of the H I and He I lines originate near the star, the lines of species such as Na I, Mg I, Ca I, Ca II, Fe I, and Fe II are formed in an outer and colder region. The shape and periodicity of the 2019 and 2021 TESS light curves, when qualitatively compared to predictions from magnetohydrodynamic models, suggest that HM Lup was in a regime of unstable ordered accretion during the 2021 TESS observation due to an increase in the accretion rate. Although HM Lup is not an extreme accretor, it shows enhanced emission in the metallic species during this high accretion state that is produced by a density enhancement in the outer part of the accretion flow., Comment: 15 pages, 14 figures. Accepted for publication in A&A

Xenon dual-phase time projections chambers (TPCs) have proven to be a successful technology in studying physical phenomena that require low-background conditions. With 40t of liquid xenon (LXe) in the TPC baseline design, DARWIN will have a high sensitivity for the detection of particle dark matter, neutrinoless double beta decay ($0\nu\beta\beta$), and axion-like particles (ALPs). Although cosmic muons are a source of background that cannot be entirely eliminated, they may be greatly diminished by placing the detector deep underground. In this study, we used Monte Carlo simulations to model the cosmogenic background expected for the DARWIN observatory at four underground laboratories: Laboratori Nazionali del Gran Sasso (LNGS), Sanford Underground Research Facility (SURF), Laboratoire Souterrain de Modane (LSM) and SNOLAB. We determine the production rates of unstable xenon isotopes and tritium due to muon-included neutron fluxes and muon-induced spallation. These are expected to represent the dominant contributions to cosmogenic backgrounds and thus the most relevant for site selection.

Existing reference-free turn-level evaluation metrics for chatbots inadequately capture the interaction between the user and the system. Consequently, they often correlate poorly with human evaluations. To address this issue, we propose a novel model-agnostic approach that leverages Conditional Pointwise Mutual Information (C-PMI) to measure the turn-level interaction between the system and the user based on a given evaluation dimension. Experimental results on the widely used FED dialogue evaluation dataset demonstrate that our approach significantly improves the correlation with human judgment compared with existing evaluation systems. By replacing the negative log-likelihood-based scorer with our proposed C-PMI scorer, we achieve a relative 62.6% higher Spearman correlation on average for the FED evaluation metric. Our code is publicly available at https://github.com/renll/C-PMI., Comment: Published at ACL2023 DialDoc Workshop; Updated Results

We perform a blind search for particle signals in the XENON1T dark matter detector that occur close in time to gravitational wave signals in the LIGO and Virgo observatories. No particle signal is observed in the nuclear recoil, electronic recoil, CE$\nu$NS, and S2-only channels within $\pm$ 500 seconds of observations of the gravitational wave signals GW170104, GW170729, GW170817, GW170818, and GW170823. We use this null result to constrain mono-energetic neutrinos and Beyond Standard Model particles emitted in the closest coalescence GW170817, a binary neutron star merger. We set new upper limits on the fluence (time-integrated flux) of coincident neutrinos down to 17 keV at 90% confidence level. Furthermore, we constrain the product of coincident fluence and cross section of Beyond Standard Model particles to be less than $10^{-29}$ cm$^2$/cm$^2$ in the [5.5-210] keV energy range at 90% confidence level.

Automatic Arabic diacritization is useful in many applications, ranging from reading support for language learners to accurate pronunciation predictor for downstream tasks like speech synthesis. While most of the previous works focused on models that operate on raw non-diacritized text, production systems can gain accuracy by first letting humans partly annotate ambiguous words. In this paper, we propose 2SDiac, a multi-source model that can effectively support optional diacritics in input to inform all predictions. We also introduce Guided Learning, a training scheme to leverage given diacritics in input with different levels of random masking. We show that the provided hints during test affect more output positions than those annotated. Moreover, experiments on two common benchmarks show that our approach i) greatly outperforms the baseline also when evaluated on non-diacritized text; and ii) achieves state-of-the-art results while reducing the parameter count by over 60%., Comment: Keywords: Arabic text diacritization, partially-diacritized text, Arabic natural language processing, Accepted at INTERSPEECH 2023