Search

Large-volume neutrino experiments are ideal for testing boosted dark matter (BDM) scenarios. We propose, for the first time, an approach to utilize knockout neutrons by detecting de-excitation $\gamma$ rays and coincident captured neutrons from dark-matter interactions with bound neutrons in oxygen, while previous studies have focused on knockout-protons and electrons. This method is especially crucial for water \v{C}erenkov detectors, where high proton \v{C}erenkov threshold ($\sim$1 GeV) suppresses signal acceptance. Recently, Super-Kamiokande (SK) was doped with 0.03\% gadolinium (SK-Gd) to enhance neutron tagging efficiency. Using SK-Gd as a target experiment, we demonstrate that this method increases sensitivity to BDM models by an order of magnitude compared to proton-based analysis, and it allows exploration of a wider range of light dark-matter models previously inaccessible with proton-based analysis. We also present the projected sensitivity for the upcoming Hyper-Kamiokande detector.

Gloss-free Sign Language Translation (SLT) converts sign videos directly into spoken language sentences without relying on glosses. Recently, Large Language Models (LLMs) have shown remarkable translation performance in gloss-free methods by harnessing their powerful natural language generation capabilities. However, these methods often rely on domain-specific fine-tuning of visual encoders to achieve optimal results. By contrast, this paper emphasizes the importance of capturing the spatial configurations and motion dynamics inherent in sign language. With this in mind, we introduce Spatial and Motion-based Sign Language Translation (SpaMo), a novel LLM-based SLT framework. The core idea of SpaMo is simple yet effective. We first extract spatial and motion features using off-the-shelf visual encoders and then input these features into an LLM with a language prompt. Additionally, we employ a visual-text alignment process as a warm-up before the SLT supervision. Our experiments demonstrate that SpaMo achieves state-of-the-art performance on two popular datasets, PHOENIX14T and How2Sign., Comment: Under Review

The growing demand for air travel requires technological advancements in air traffic management as well as mechanisms for monitoring and ensuring safe and efficient operations. In terminal airspaces, predictive models of future movements and traffic flows can help with proactive planning and efficient coordination; however, varying airport topologies, and interactions with other agents, among other factors, make accurate predictions challenging. Data-driven predictive models have shown promise for handling numerous variables to enable various downstream tasks, including collision risk assessment, taxi-out time prediction, departure metering, and emission estimations. While data-driven methods have shown improvements in these tasks, prior works lack large-scale curated surface movement datasets within the public domain and the development of generalizable trajectory forecasting models. In response to this, we propose two contributions: (1) Amelia-48, a large surface movement dataset collected using the System Wide Information Management (SWIM) Surface Movement Event Service (SMES). With data collection beginning in Dec 2022, the dataset provides more than a year's worth of SMES data (~30TB) and covers 48 airports within the US National Airspace System. In addition to releasing this data in the public domain, we also provide post-processing scripts and associated airport maps to enable research in the forecasting domain and beyond. (2) Amelia-TF model, a transformer-based next-token-prediction large multi-agent multi-airport trajectory forecasting model trained on 292 days or 9.4 billion tokens of position data encompassing 10 different airports with varying topology. The open-sourced model is validated on unseen airports with experiments showcasing the different prediction horizon lengths, ego-agent selection strategies, and training recipes to demonstrate the generalization capabilities., Comment: 24 pages, 9 figures, 8 tables

Stochastic resetting has recently emerged as an efficient target-searching strategy in various physical and biological systems. The efficiency of this strategy depends on the type of environmental noise, whether it is thermal or telegraphic (active). While the impact of each noise type on a search process has been investigated separately, their combined effects have not been explored. In this work, we explore the effects of stochastic resetting on an active system, namely a self-propelled run-and-tumble particle immersed in a thermal bath. In particular, we assume that the position of the particle is reset at a fixed rate with or without reversing the direction of self-propelled velocity. Using standard renewal techniques, we compute the mean search time of this active particle to a fixed target and investigate the interplay between active and thermal fluctuations. We find that the active search can outperform the Brownian search when the magnitude and flipping rate of self-propelled velocity are large and the strength of environmental noise is small. Notably, we find that the presence of thermal noise in the environment helps reduce the mean first passage time of the run-and-tumble particle compared to the absence of thermal noise. Finally, we observe that reversing the direction of self-propelled velocity while resetting can also reduce the overall search time., Comment: 10 pages, 4 figures

Recent advancements in Large Language Models (LLMs) have significantly improved their performance across various Natural Language Processing (NLP) tasks. However, LLMs still struggle with generating non-factual responses due to limitations in their parametric memory. Retrieval-Augmented Generation (RAG) systems address this issue by incorporating external knowledge with a retrieval module. Despite their successes, however, current RAG systems face challenges with retrieval failures and the limited ability of LLMs to filter out irrelevant information. Therefore, in this work, we propose DSLR (Document Refinement with Sentence-Level Re-ranking and Reconstruction), an unsupervised framework that decomposes retrieved documents into sentences, filters out irrelevant sentences, and reconstructs them again into coherent passages. We experimentally validate DSLR on multiple open-domain QA datasets and the results demonstrate that DSLR significantly enhances the RAG performance over conventional fixed-size passage. Furthermore, our DSLR enhances performance in specific, yet realistic scenarios without the need for additional training, providing an effective and efficient solution for refining retrieved documents in RAG systems., Comment: 20 pages

Sign language, essential for the deaf and hard-of-hearing, presents unique challenges in translation and production due to its multimodal nature and the inherent ambiguity in mapping sign language motion to spoken language words. Previous methods often rely on gloss annotations, requiring time-intensive labor and specialized expertise in sign language. Gloss-free methods have emerged to address these limitations, but they often depend on external sign language data or dictionaries, failing to completely eliminate the need for gloss annotations. There is a clear demand for a comprehensive approach that can supplant gloss annotations and be utilized for both Sign Language Translation (SLT) and Sign Language Production (SLP). We introduce Universal Gloss-level Representation (UniGloR), a unified and self-supervised solution for both SLT and SLP, trained on multiple datasets including PHOENIX14T, How2Sign, and NIASL2021. Our results demonstrate UniGloR's effectiveness in the translation and production tasks. We further report an encouraging result for the Sign Language Recognition (SLR) on previously unseen data. Our study suggests that self-supervised learning can be made in a unified manner, paving the way for innovative and practical applications in future research., Comment: 14 pages, 5 figures

As a fundamental thermodynamic principle, speed limits reveal the lower bound of entropy production (EP) required for a system to transition from a given initial state to a final state. While various speed limits have been developed for continuous-time Markov processes, their application to discrete-time Markov chains remains unexplored. In this study, we investigate the speed limits in discrete-time Markov chains, focusing on two types of EP commonly used to measure the irreversibility of a discrete-time process: time-reversed EP and time-backward EP. We find that time-reversed EP satisfies the speed limit for the continuous-time Markov processes, whereas time-backward EP does not. Additionally, for time-reversed EP, we derive practical speed limits applicable to systems driven by cyclic protocols or with unidirectional transitions, where conventional speed limits become meaningless or invalid. We show that these relations also hold for continuous-time Markov processes by taking the time-continuum limit of our results. Finally, we validate our findings through several examples., Comment: 8 pages, 3 figures

Information retrieval models that aim to search for the documents relevant to the given query have shown many successes, which have been applied to diverse tasks. However, the query provided by the user is oftentimes very short, which challenges the retrievers to correctly fetch relevant documents. To tackle this, existing studies have proposed expanding the query with a couple of additional (user-related) features related to the query. Yet, they may be suboptimal to effectively augment the query, though there is plenty of information available to augment it in a relational database. Motivated by this, we present a novel retrieval framework called Database-Augmented Query representation (DAQu), which augments the original query with various (query-related) metadata across multiple tables. In addition, as the number of features in the metadata can be very large and there is no order among them, we encode them with our graph-based set encoding strategy, which considers hierarchies of features in the database without order. We validate DAQu in diverse retrieval scenarios that can incorporate metadata from the relational database, demonstrating that ours significantly enhances overall retrieval performance, compared to existing query augmentation methods.

We report on investigations of mode coupling between rotating spokes during the onset of the breathing oscillation. Demonstrating the existence of nonlinear coupling between the sporadic spokes and the breathing oscillation, we suggest the oscillating azimuthal electric field as the energy source for additional ionization within the plasma. Our results indicate that intermittent three-wave coupling is a possible mechanism for triggering low-frequency breathing oscillations in partially magnetized cross-field plasma.

Recent language models have shown remarkable performance on natural language understanding (NLU) tasks. However, they are often sub-optimal when faced with ambiguous samples that can be interpreted in multiple ways, over-confidently predicting a single label without consideration for its correctness. To address this issue, we propose a novel self-knowledge distillation method that enables models to learn label distributions more accurately by leveraging knowledge distilled from their lower layers. This approach also includes a learning phase that re-calibrates the unnecessarily strengthened confidence for training samples judged as extremely ambiguous based on the distilled distribution knowledge. We validate our method on diverse NLU benchmark datasets and the experimental results demonstrate its effectiveness in producing better label distributions. Particularly, through the process of re-calibrating the confidence for highly ambiguous samples, the issue of over-confidence when predictions for unseen samples do not match with their ground-truth labels has been significantly alleviated. This has been shown to contribute to generating better distributions than the existing state-of-the-art method. Moreover, our method is more efficient in training the models compared to the existing method, as it does not involve additional training processes to refine label distributions., Comment: 9 pages, 5 figures

Traditional federated learning (FL) methods have limited support for clients with varying computational and communication abilities, leading to inefficiencies and potential inaccuracies in model training. This limitation hinders the widespread adoption of FL in diverse and resource-constrained environments, such as those with client devices ranging from powerful servers to mobile devices. To address this need, this paper introduces Federated Learning with Flexible Architectures (FedFA), an FL training algorithm that allows clients to train models of different widths and depths. Each client can select a network architecture suitable for its resources, with shallower and thinner networks requiring fewer computing resources for training. Unlike prior work in this area, FedFA incorporates the layer grafting technique to align clients' local architectures with the largest network architecture in the FL system during model aggregation. Layer grafting ensures that all client contributions are uniformly integrated into the global model, thereby minimizing the risk of any individual client's data skewing the model's parameters disproportionately and introducing security benefits. Moreover, FedFA introduces the scalable aggregation method to manage scale variations in weights among different network architectures. Experimentally, FedFA outperforms previous width and depth flexible aggregation strategies. Furthermore, FedFA demonstrates increased robustness against performance degradation in backdoor attack scenarios compared to earlier strategies.

Vertical take-off and landing (VTOL) aircraft do not require a prolonged runway, thus allowing them to land almost anywhere. In recent years, their flexibility has made them popular in development, research, and operation. When compared to traditional fixed-wing aircraft and rotorcraft, VTOLs bring unique challenges as they combine many maneuvers from both types of aircraft. Pilot workload is a critical factor for safe and efficient operation of VTOLs. In this work, we conduct a user study to collect multimodal data from 28 pilots while they perform a variety of VTOL flight tasks. We analyze and interpolate behavioral patterns related to their performance and perceived workload. Finally, we build machine learning models to estimate their workload from the collected data. Our results are promising, suggesting that quantitative and accurate VTOL pilot workload monitoring is viable. Such assistive tools would help the research field understand VTOL operations and serve as a stepping stone for the industry to ensure VTOL safe operations and further remote operations., Comment: 8 pages, 7 figures

Social bias is shaped by the accumulation of social perceptions towards targets across various demographic identities. To fully understand such social bias in large language models (LLMs), it is essential to consider the composite of social perceptions from diverse perspectives among identities. Previous studies have either evaluated biases in LLMs by indirectly assessing the presence of sentiments towards demographic identities in the generated text or measuring the degree of alignment with given stereotypes. These methods have limitations in directly quantifying social biases at the level of distinct perspectives among identities. In this paper, we aim to investigate how social perceptions from various viewpoints contribute to the development of social bias in LLMs. To this end, we propose a novel strategy to intuitively quantify these social perceptions and suggest metrics that can evaluate the social biases within LLMs by aggregating diverse social perceptions. The experimental results show the quantitative demonstration of the social attitude in LLMs by examining social perception. The analysis we conducted shows that our proposed metrics capture the multi-dimensional aspects of social bias, enabling a fine-grained and comprehensive investigation of bias in LLMs., Comment: Findings of ACL 2024

ITER coil tolerances are re-evaluated using the modern understanding of coupling to least-stable plasma modes and an updated center-line-traced model of ITER's coil windings. This reassessment finds the tolerances to be conservative through a statistical, linear study of $n=1$ error fields (EFs) due to tilted, shifted misplacements and nominal windings of central solenoid and poloidal field coils within tolerance. We also show that a model-based correction scheme remains effective even when metrology quality is sub-optimal, and compare this to projected empirical correction schemes. We begin with an analysis of the necessity of error field correction (EFC) for daily operation in ITER using scaling laws for the EF penetration threshold. We then consider the predictability of EF dominant mode overlap across early planned ITER scenarios and, as measuring EFs in high power scenarios can pose risks to the device, the potential for extrapolation to the ITER Baseline Scenario (IBS). We find that carefully designing a scenario matching currents proportionally to those of the IBS is far more important than plasma shape or profiles in accurately measuring an optimal correction current set., Comment: 11 pages, 13 Figures. Submitted to Nuclear Fusion

The path of tokamak fusion and ITER is maintaining high-performance plasma to produce sufficient fusion power. This effort is hindered by the transient energy burst arising from the instabilities at the boundary of high-confinement plasmas. The application of 3D magnetic perturbations is the method in ITER and possibly in future fusion power plants to suppress this instability and avoid energy busts damaging the device. Unfortunately, the conventional use of the 3D field in tokamaks typically leads to degraded fusion performance and an increased risk of other plasma instabilities, two severe issues for reactor implementation. In this work, we present an innovative 3D field optimization, exploiting machine learning, real-time adaptability, and multi-device capabilities to overcome these limitations. This integrated scheme is successfully deployed on DIII-D and KSTAR tokamaks, consistently achieving reactor-relevant core confinement and the highest fusion performance without triggering damaging instabilities or bursts while demonstrating ITER-relevant automated 3D optimization for the first time. This is enabled both by advances in the physics understanding of self-organized transport in the plasma edge and by advances in machine-learning technology, which is used to optimize the 3D field spectrum for automated management of a volatile and complex system. These findings establish real-time adaptive 3D field optimization as a crucial tool for ITER and future reactors to maximize fusion performance while simultaneously minimizing damage to machine components.

The robustness of recent Large Language Models (LLMs) has become increasingly crucial as their applicability expands across various domains and real-world applications. Retrieval-Augmented Generation (RAG) is a promising solution for addressing the limitations of LLMs, yet existing studies on the robustness of RAG often overlook the interconnected relationships between RAG components or the potential threats prevalent in real-world databases, such as minor textual errors. In this work, we investigate two underexplored aspects when assessing the robustness of RAG: 1) vulnerability to noisy documents through low-level perturbations and 2) a holistic evaluation of RAG robustness. Furthermore, we introduce a novel attack method, the Genetic Attack on RAG (\textit{GARAG}), which targets these aspects. Specifically, GARAG is designed to reveal vulnerabilities within each component and test the overall system functionality against noisy documents. We validate RAG robustness by applying our \textit{GARAG} to standard QA datasets, incorporating diverse retrievers and LLMs. The experimental results show that GARAG consistently achieves high attack success rates. Also, it significantly devastates the performance of each component and their synergy, highlighting the substantial risk that minor textual inaccuracies pose in disrupting RAG systems in the real world., Comment: Under Review

Federated learning (FL) enables collaborative learning across multiple clients. In most FL work, all clients train a single learning task. However, the recent proliferation of FL applications may increasingly require multiple FL tasks to be trained simultaneously, sharing clients' computing and communication resources, which we call Multiple-Model Federated Learning (MMFL). Current MMFL algorithms use naive average-based client-task allocation schemes that can lead to unfair performance when FL tasks have heterogeneous difficulty levels, e.g., tasks with larger models may need more rounds and data to train. Just as naively allocating resources to generic computing jobs with heterogeneous resource needs can lead to unfair outcomes, naive allocation of clients to FL tasks can lead to unfairness, with some tasks having excessively long training times, or lower converged accuracies. Furthermore, in the FL setting, since clients are typically not paid for their training effort, we face a further challenge that some clients may not even be willing to train some tasks, e.g., due to high computational costs, which may exacerbate unfairness in training outcomes across tasks. We address both challenges by firstly designing FedFairMMFL, a difficulty-aware algorithm that dynamically allocates clients to tasks in each training round. We provide guarantees on airness and FedFairMMFL's convergence rate. We then propose a novel auction design that incentivizes clients to train multiple tasks, so as to fairly distribute clients' training efforts across the tasks. We show how our fairness-based learning and incentive mechanisms impact training convergence and finally evaluate our algorithm with multiple sets of learning tasks on real world datasets.

Understanding the behaviors of ecological systems is challenging given their multi-faceted complexity. To proceed, theoretical models such as Lotka-Volterra dynamics with random interactions have been investigated by the dynamical mean-field theory to provide insights into underlying principles such as how biodiversity and stability depend on the randomness in interaction strength. Yet the fully-connected structure assumed in these previous studies is not realistic as revealed by a vast amount of empirical data. We derive a generic formula for the abundance distribution under an arbitrary distribution of degree, the number of interacting neighbors, which leads to degree-dependent abundance patterns of species. Notably, in contrast to the well-mixed system, the number of surviving species can be reduced as the community becomes cooperative in heterogeneous interaction structures. Our study, therefore, demonstrates that properly taking into account heterogeneity in the interspecific interaction structure is indispensable to understanding the diversity in large ecosystems, and our general theoretical framework can apply to a much wider range of interacting many-body systems.

Retrieval-Augmented Large Language Models (LLMs), which incorporate the non-parametric knowledge from external knowledge bases into LLMs, have emerged as a promising approach to enhancing response accuracy in several tasks, such as Question-Answering (QA). However, even though there are various approaches dealing with queries of different complexities, they either handle simple queries with unnecessary computational overhead or fail to adequately address complex multi-step queries; yet, not all user requests fall into only one of the simple or complex categories. In this work, we propose a novel adaptive QA framework, that can dynamically select the most suitable strategy for (retrieval-augmented) LLMs from the simplest to the most sophisticated ones based on the query complexity. Also, this selection process is operationalized with a classifier, which is a smaller LM trained to predict the complexity level of incoming queries with automatically collected labels, obtained from actual predicted outcomes of models and inherent inductive biases in datasets. This approach offers a balanced strategy, seamlessly adapting between the iterative and single-step retrieval-augmented LLMs, as well as the no-retrieval methods, in response to a range of query complexities. We validate our model on a set of open-domain QA datasets, covering multiple query complexities, and show that ours enhances the overall efficiency and accuracy of QA systems, compared to relevant baselines including the adaptive retrieval approaches. Code is available at: https://github.com/starsuzi/Adaptive-RAG., Comment: NAACL 2024

Radiation therapy plays a crucial role in cancer treatment, necessitating precise delivery of radiation to tumors while sparing healthy tissues over multiple days. Computed tomography (CT) is integral for treatment planning, offering electron density data crucial for accurate dose calculations. However, accurately representing patient anatomy is challenging, especially in adaptive radiotherapy, where CT is not acquired daily. Magnetic resonance imaging (MRI) provides superior soft-tissue contrast. Still, it lacks electron density information while cone beam CT (CBCT) lacks direct electron density calibration and is mainly used for patient positioning. Adopting MRI-only or CBCT-based adaptive radiotherapy eliminates the need for CT planning but presents challenges. Synthetic CT (sCT) generation techniques aim to address these challenges by using image synthesis to bridge the gap between MRI, CBCT, and CT. The SynthRAD2023 challenge was organized to compare synthetic CT generation methods using multi-center ground truth data from 1080 patients, divided into two tasks: 1) MRI-to-CT and 2) CBCT-to-CT. The evaluation included image similarity and dose-based metrics from proton and photon plans. The challenge attracted significant participation, with 617 registrations and 22/17 valid submissions for tasks 1/2. Top-performing teams achieved high structural similarity indices (>0.87/0.90) and gamma pass rates for photon (>98.1%/99.0%) and proton (>97.3%/97.0%) plans. However, no significant correlation was found between image similarity metrics and dose accuracy, emphasizing the need for dose evaluation when assessing the clinical applicability of sCT. SynthRAD2023 facilitated the investigation and benchmarking of sCT generation techniques, providing insights for developing MRI-only and CBCT-based adaptive radiotherapy., Comment: Preprint submitted to Medical Image Analysis

We propose a new approach to search for light dark matter (DM) in the range of keV-GeV via inelastic nucleus scattering at large-volume neutrino detectors such as Borexino, DUNE, JUNO, and Hyper-K. The approach uses inelastic nuclear scattering of cosmic-ray boosted DM, enabling a low-background search for DM in these experiments. The large neutrino detectors with higher threshold can be used since the nuclear deexcitation lines are $O(10)$ MeV. Using a hadrophilic dark-gauge-boson-portal model as a benchmark, we show that the nuclear inelastic channels generally provide better sensitivity than the elastic scattering for a large region of light DM parameter space., Comment: 7 pages, 4 figures

Instruction tuning benefits from large and diverse datasets; however, creating such datasets involves a high cost of human labeling. While synthetic datasets generated by large language models (LLMs) have partly solved this issue, they often contain low-quality data. One effective solution is selectively annotating unlabelled instructions, especially given the relative ease of acquiring unlabeled instructions or texts from various sources. However, how to select unlabelled instructions is not well-explored, especially in the context of LLMs. Therefore, we introduce SelectLLM, an alternative framework that leverages the capabilities of LLMs to select unlabeled instructions more effectively. Specifically, SelectLLM consists of two key steps: Coreset-based clustering of unlabelled instructions for enlarging diversity and prompting of LLM to identify the most beneficial instructions within each cluster. We evaluate SelectLLM on AlpacaEval2 and MT-Bench, demonstrating its ability to outperform state-of-the-art methods like Alpagasus. In addition, we compare the performance and compatibility of SelectLLM with various LLMs, such as ChatGPT, LLaMA-3.1-70B, and Gemma-2-27b. SelectLLM's adaptability and robustness are further evidenced by its ability to maintain high performance across both human and synthetic datasets. All code and data are publicly available (https://github.com/minnesotanlp/select-llm)., Comment: First Authors: Ritik Sachin Parkar and Jaehyung Kim | Second Author: Jong Inn Park | PI: Dongyeop Kang

Sub-MeV cold dark-matter particles are unable to produce electronic recoil in conventional dark-matter direct detection experiments such as XENONnT and LUX-ZEPLIN above the detector threshold. The mechanism of boosted dark matter comes into picture to constrain the parameter space of such low mass dark matter from direct detection experiments. We consider the effect of the leading components of cosmic rays to boost the cold dark matter, which results in significant improvements on the exclusion limits compared to the existing ones. To present concrete study results, we choose to work on models consisting of a dark-matter particle $\chi$ with an additional $U(1)'$ gauge symmetry including the secluded dark photon, $U(1)_{B-L}$, and $U(1)_{L_e-L_\mu}$. We find that the energy dependence of the scattering cross section plays a crucial role in improving the constraints. In addition, we systematically estimate the Earth shielding effect on boosted dark matter in losing energy while traveling to the underground detector through the Earth., Comment: Accepted for Publication in JCAP