Search

Domain wall (DW) dynamics in antiferromagnetic (AFM) systems offer the advantages over their ferromagnetic counterparts of having faster and more energy efficient manipulation due to the absence of net magnetization, leading to reduced magnetic crosstalk and improved performance in spintronic devices. A comprehensive analysis of DW dynamics across regimes such as creep, depinning, and flow is well established in ferromagnetic systems but remains lacking in AFM-coupled systems. In this study, we explore the nature of DW dynamics in synthetic ferrimagnetic multilayers composed of Pt|Co|Tb|Al for different Tb thickness, focusing on the underlying pinning parameters, and on the different regimes of DW dynamics driven by spin-orbit torques (SOTs). We find that due to the AFM coupling between Co and Tb, the magnetic moment of Tb increases with Tb thickness resulting in a reduced saturation magnetization and an enhanced depinning field. The DW disorder interaction is found to vary weakly with the AFM coupling between Co and Tb, while the complete withdrawal of the Tb layer strongly increases the anisotropy and the DW pinning. Furthermore, we propose a novel approach to measure effective SOTs by comparing depinning transitions in current and field-induced DW motion. This research reveals new insights into DW dynamics in coupled AFM systems, highlighting enhancements in mobility through optimized SOTs and pinning landscapes., Comment: 12 pages, 9 figures

Benchmarks are important tools for tracking the rapid advancements in large language model (LLM) capabilities. However, benchmarks are not keeping pace in difficulty: LLMs now achieve over 90\% accuracy on popular benchmarks like MMLU, limiting informed measurement of state-of-the-art LLM capabilities. In response, we introduce Humanity's Last Exam (HLE), a multi-modal benchmark at the frontier of human knowledge, designed to be the final closed-ended academic benchmark of its kind with broad subject coverage. HLE consists of 2,700 questions across dozens of subjects, including mathematics, humanities, and the natural sciences. HLE is developed globally by subject-matter experts and consists of multiple-choice and short-answer questions suitable for automated grading. Each question has a known solution that is unambiguous and easily verifiable, but cannot be quickly answered via internet retrieval. State-of-the-art LLMs demonstrate low accuracy and calibration on HLE, highlighting a significant gap between current LLM capabilities and the expert human frontier on closed-ended academic questions. To inform research and policymaking upon a clear understanding of model capabilities, we publicly release HLE at https://lastexam.ai., Comment: 27 pages, 6 figures

Many active galaxies harbor powerful relativistic jets, however, the detailed mechanisms of their formation and acceleration remain poorly understood. To investigate the area of jet acceleration and collimation with the highest available angular resolution, we study the innermost region of the bipolar jet in the nearby low-ionization nuclear emission-line region (LINER) galaxy NGC 1052. We combined observations of NGC 1052 taken with VLBA, GMVA, and EHT over one week in the spring of 2017. For the first time, NGC 1052 was detected with the EHT, providing a size of the central region in-between both jet bases of 250 RS (Schwarzschild radii) perpendicular to the jet axes. This size estimate supports previous studies of the jets expansion profile which suggest two breaks of the profile at around 300 RS and 10000 RS distances to the core. Furthermore, we estimated the magnetic field to be 1.25 Gauss at a distance of 22 {\mu}as from the central engine by fitting a synchrotron-self absorption spectrum to the innermost emission feature, which shows a spectral turn-over at about 130 GHz. Assuming a purely poloidal magnetic field, this implies an upper limit on the magnetic field strength at the event horizon of 26000 Gauss, which is consistent with previous measurements. The complex, low-brightness, double-sided jet structure in NGC 1052 makes it a challenge to detect the source at millimeter (mm) wavelengths. However, our first EHT observations have demonstrated that detection is possible up to at least 230 GHz. This study offers a glimpse through the dense surrounding torus and into the innermost central region, where the jets are formed. This has enabled us to finally resolve this region and provide improved constraints on its expansion and magnetic field strength., Comment: 22 pages, 10 figures, published in A&A

The 2017 observing campaign of the Event Horizon Telescope (EHT) delivered the first very long baseline interferometry (VLBI) images at the observing frequency of 230 GHz, leading to a number of unique studies on black holes and relativistic jets from active galactic nuclei (AGN). In total, eighteen sources were observed: the main science targets, Sgr A* and M87 along with various calibrators. We investigated the morphology of the sixteen AGN in the EHT 2017 data set, focusing on the properties of the VLBI cores: size, flux density, and brightness temperature. We studied their dependence on the observing frequency in order to compare it with the Blandford-K\"onigl (BK) jet model. We modeled the source structure of seven AGN in the EHT 2017 data set using linearly polarized circular Gaussian components and collected results for the other nine AGN from dedicated EHT publications, complemented by lower frequency data in the 2-86 GHz range. Then, we studied the dependences of the VLBI core flux density, size, and brightness temperature on the frequency measured in the AGN host frame. We compared the observations with the BK jet model and estimated the magnetic field strength dependence on the distance from the central black hole. Our results indicate a deviation from the standard BK model, particularly in the decrease of the brightness temperature with the observing frequency. Either bulk acceleration of the jet material, energy transfer from the magnetic field to the particles, or both are required to explain the observations.

We report ring resonators on a silicon germanium on silicon platform operating in the mid-infrared wavelength range around 3.5 - 4.6 {\mu}m with quality factors reaching up to one million. Advances in fabrication technology enable us to demonstrate such high Q-factors, which put silicon germanium at the forefront of mid-infrared integrated photonic platforms. The achievement of high Q is attested by the observation of degeneracy lifting between clockwise (CW) and counter-clockwise (CCW) resonances, as well as optical bistability due to an efficient power buildup in the rings.

A major challenge for reinforcement learning is automatically generating curricula to reduce training time or improve performance in some target task. We introduce SEBNs (Skill-Environment Bayesian Networks) which model a probabilistic relationship between a set of skills, a set of goals that relate to the reward structure, and a set of environment features to predict policy performance on (possibly unseen) tasks. We develop an algorithm that uses the inferred estimates of agent success from SEBN to weigh the possible next tasks by expected improvement. We evaluate the benefit of the resulting curriculum on three environments: a discrete gridworld, continuous control, and simulated robotics. The results show that curricula constructed using SEBN frequently outperform other baselines.

Large language models have demonstrated remarkable progress in mathematical reasoning, leveraging chain-of-thought and test-time compute scaling. However, many open questions remain regarding the interplay between reasoning token usage and accuracy gains. In particular, when comparing models across generations, it is unclear whether improved performance results from longer reasoning chains or more efficient reasoning. We systematically analyze chain-of-thought length across o1-mini and o3-mini variants on the Omni-MATH benchmark, finding that o3-mini (m) achieves superior accuracy without requiring longer reasoning chains than o1-mini. Moreover, we show that accuracy generally declines as reasoning chains grow across all models and compute settings, even when controlling for difficulty of the questions. This accuracy drop is significantly smaller in more proficient models, suggesting that new generations of reasoning models use test-time compute more effectively. Finally, we highlight that while o3-mini (h) achieves a marginal accuracy gain over o3-mini (m), it does so by allocating substantially more reasoning tokens across all problems, even the ones that o3-mini (m) can already solve. These findings provide new insights into the relationship between model capability and reasoning length, with implications for efficiency, scaling, and evaluation methodologies., Comment: 19 pages, 11 figures

Amorphous Al$_{2}$O$_{3}$ is a fundamental component of modern superconducting qubits. While amphorphous oxides offer distinct advantages, such as directional isotropy and a consistent bulk electronic gap, in realistic systems these compounds support two-level systems (TLSs) which couple to the qubit, expediting decoherence. In this work, we perform a first-principles study of amorphous Al$_{2}$O$_{3}$ and identify low-energy modes in the electronic and phonon spectra as a possible origin for TLSs., Comment: 5+1 Pages, 7 + 3 Figures

Extremely large-scale multiple-input multiple-output (XL-MIMO) is expected to be an important technology in future sixth generation (6G) networks. Compared with conventional single-polarized XL-MIMO, where signals are transmitted and received in only one polarization direction, dual-polarized XL-MIMO systems achieve higher data rate by improving multiplexing performances, and thus are the focus of this paper. Due to enlarged aperture, near-field regions become non-negligible in XL-MIMO communications, necessitating accurate near-far field boundary characterizations. However, existing boundaries developed for single-polarized systems only consider phase or power differences across array elements while irrespective of cross-polarization discrimination (XPD) variances in dual-polarized XL-MIMO systems, deteriorating transmit covariance optimization performances. In this paper, we revisit near-far field boundaries for dual-polarized XL-MIMO systems by taking XPD differences into account, which faces the following challenge. Unlike existing near-far field boundaries, which only need to consider co-polarized channel components, deriving boundaries for dual-polarized XL-MIMO systems requires modeling joint effects of co-polarized and cross-polarized components. To address this issue, we model XPD variations across antennas and introduce a non-uniform XPD distance to complement existing near-far field boundaries. Based on the new distance criterion, we propose an efficient scheme to optimize transmit covariance. Numerical results validate our analysis and demonstrate the proposed algorithm's effectiveness., Comment: 16 pages, 8 figures

We introduce Meta MLGym and MLGym-Bench, a new framework and benchmark for evaluating and developing LLM agents on AI research tasks. This is the first Gym environment for machine learning (ML) tasks, enabling research on reinforcement learning (RL) algorithms for training such agents. MLGym-bench consists of 13 diverse and open-ended AI research tasks from diverse domains such as computer vision, natural language processing, reinforcement learning, and game theory. Solving these tasks requires real-world AI research skills such as generating new ideas and hypotheses, creating and processing data, implementing ML methods, training models, running experiments, analyzing the results, and iterating through this process to improve on a given task. We evaluate a number of frontier large language models (LLMs) on our benchmarks such as Claude-3.5-Sonnet, Llama-3.1 405B, GPT-4o, o1-preview, and Gemini-1.5 Pro. Our MLGym framework makes it easy to add new tasks, integrate and evaluate models or agents, generate synthetic data at scale, as well as develop new learning algorithms for training agents on AI research tasks. We find that current frontier models can improve on the given baselines, usually by finding better hyperparameters, but do not generate novel hypotheses, algorithms, architectures, or substantial improvements. We open-source our framework and benchmark to facilitate future research in advancing the AI research capabilities of LLM agents., Comment: 35 pages, 12 figures, 10 tables

Extracting tables from documents is a critical task across various industries, especially on business documents like invoices and reports. Existing systems based on DEtection TRansformer (DETR) such as TAble TRansformer (TATR), offer solutions for Table Detection (TD) and Table Structure Recognition (TSR) but face challenges with diverse table formats and common errors like incorrect area detection and overlapping columns. This research introduces RAPTOR, a modular post-processing system designed to enhance state-of-the-art models for improved table extraction, particularly for product tables. RAPTOR addresses recurrent TD and TSR issues, improving both precision and structural predictions. For TD, we use DETR (trained on ICDAR 2019) and TATR (trained on PubTables-1M and FinTabNet), while TSR only relies on TATR. A Genetic Algorithm is incorporated to optimize RAPTOR's module parameters, using a private dataset of product tables to align with industrial needs. We evaluate our method on two private datasets of product tables, the public DOCILE dataset (which contains tables similar to our target product tables), and the ICDAR 2013 and ICDAR 2019 datasets. The results demonstrate that while our approach excels at product tables, it also maintains reasonable performance across diverse table formats. An ablation study further validates the contribution of each module in our system., Comment: Accepted for WACVW 2025 (VisionDocs)

Vehicle-to-everything (V2X) collaborative perception has emerged as a promising solution to address the limitations of single-vehicle perception systems. However, existing V2X datasets are limited in scope, diversity, and quality. To address these gaps, we present Mixed Signals, a comprehensive V2X dataset featuring 45.1k point clouds and 240.6k bounding boxes collected from three connected autonomous vehicles (CAVs) equipped with two different types of LiDAR sensors, plus a roadside unit with dual LiDARs. Our dataset provides precisely aligned point clouds and bounding box annotations across 10 classes, ensuring reliable data for perception training. We provide detailed statistical analysis on the quality of our dataset and extensively benchmark existing V2X methods on it. Mixed Signals V2X Dataset is one of the highest quality, large-scale datasets publicly available for V2X perception research. Details on the website https://mixedsignalsdataset.cs.cornell.edu/.

The rapid development of advanced AI agents and the imminent deployment of many instances of these agents will give rise to multi-agent systems of unprecedented complexity. These systems pose novel and under-explored risks. In this report, we provide a structured taxonomy of these risks by identifying three key failure modes (miscoordination, conflict, and collusion) based on agents' incentives, as well as seven key risk factors (information asymmetries, network effects, selection pressures, destabilising dynamics, commitment problems, emergent agency, and multi-agent security) that can underpin them. We highlight several important instances of each risk, as well as promising directions to help mitigate them. By anchoring our analysis in a range of real-world examples and experimental evidence, we illustrate the distinct challenges posed by multi-agent systems and their implications for the safety, governance, and ethics of advanced AI., Comment: Cooperative AI Foundation, Technical Report #1

Citations are a key indicator of research impact but are shaped by factors beyond intrinsic research quality, including prestige, social networks, and thematic similarity. While the Matthew Effect explains how prestige accumulates and influences citation distributions, our study contextualizes this by showing that other mechanisms also play a crucial role. Analyzing a large dataset of disambiguated authors (N=43,467) and citation linkages (N=264,436) in U.S. economics, we find that close ties in the collaboration network are the strongest predictor of citation, closely followed by thematic similarity between papers. This reinforces the idea that citations are not only a matter of prestige but mostly of social networks and intellectual proximity. Prestige remains important for understanding highly cited papers, but for the majority of citations, proximity--both social and semantic--plays a more significant role. These findings shift attention from extreme cases of highly cited research toward the broader distribution of citations, which shapes career trajectories and the production of knowledge. Recognizing the diverse factors influencing citations is critical for science policy, as this work highlights inequalities that are not based on preferential attachment, but on the role of self-citations, collaborations, and mainstream versus no mainstream research subjects.

The field of visual localization has been researched for several decades and has meanwhile found many practical applications. Despite the strong progress in this field, there are still challenging situations in which established methods fail. We present an approach to significantly improve the accuracy and reliability of established visual localization methods by adding rendered images. In detail, we first use a modern visual SLAM approach that provides a 3D Gaussian Splatting (3DGS) based map to create reference data. We demonstrate that enriching reference data with images rendered from 3DGS at randomly sampled poses significantly improves the performance of both geometry-based visual localization and Scene Coordinate Regression (SCR) methods. Through comprehensive evaluation in a large industrial environment, we analyze the performance impact of incorporating these additional rendered views.

Large Language Models (LLMs), when used for conditional text generation, often produce hallucinations, i.e., information that is unfaithful or not grounded in the input context. This issue arises in typical conditional text generation tasks, such as text summarization and data-to-text generation, where the goal is to produce fluent text based on contextual input. When fine-tuned on specific domains, LLMs struggle to provide faithful answers to a given context, often adding information or generating errors. One underlying cause of this issue is that LLMs rely on statistical patterns learned from their training data. This reliance can interfere with the model's ability to stay faithful to a provided context, leading to the generation of ungrounded information. We build upon this observation and introduce a novel self-supervised method for generating a training set of unfaithful samples. We then refine the model using a training process that encourages the generation of grounded outputs over unfaithful ones, drawing on preference-based training. Our approach leads to significantly more grounded text generation, outperforming existing self-supervised techniques in faithfulness, as evaluated through automatic metrics, LLM-based assessments, and human evaluations., Comment: 10 pages, ICLR 2025 conference

Ultra-hot Jupiters, an extreme class of planets not found in our solar system, provide a unique window into atmospheric processes. The extreme temperature contrasts between their day- and night-sides pose a fundamental climate puzzle: how is energy distributed? To address this, we must observe the 3D structure of these atmospheres, particularly their vertical circulation patterns, which can serve as a testbed for advanced Global Circulation Models (GCM) [e.g. 1]. Here, we show a dramatic shift in atmospheric circulation in an ultra-hot Jupiter: a unilateral flow from the hot star-facing side to the cooler space-facing side of the planet sits below an equatorial super-rotational jet stream. By resolving the vertical structure of atmospheric dynamics, we move beyond integrated global snapshots of the atmosphere, enabling more accurate identification of flow patterns and allowing for a more nuanced comparison to models. Global circulation models based on first principles struggle to replicate the observed circulation pattern [3], underscoring a critical gap between theoretical understanding of atmospheric flows and observational evidence. This work serves as a testbed to develop more comprehensive models applicable beyond our Solar System as we prepare for the next generation of giant telescopes., Comment: Accepted for publication in Nature on 16th January 2025, published with DOI 10.1038/s41586-025-08664-1, 5 main figures, 12 main pages plus methods. This work has a companion paper on the same dataset: Prinoth et al. 2025, A&A, DOI: 10.1051/0004-6361/202452405

Neural signals related to movement can be measured from intracranial recordings and used in brain-machine interface devices (BMI) to restore physical function in impaired patients. In this study, we explore the use of more abstract neural signals related to economic value in a BMI context. Using data collected from the orbitofrontal cortex in non-human primates, we develop deep learning-based neural decoders that can predict the monkey's choice in a value-based decision-making task. Out-of-sample performance was improved by augmenting the training set with synthesized data, showing the feasibility of using limited training data. We further demonstrate that we can predict the monkey's choice sooner using a neural forecasting module that is equipped with task-related information. These findings support the feasibility of user preference-informed neuroengineering devices that leverage abstract cognitive signals., Comment: 23 pages, 5 figures

Generating long, high-quality videos remains a challenge due to the complex interplay of spatial and temporal dynamics and hardware limitations. In this work, we introduce \textbf{MaskFlow}, a unified video generation framework that combines discrete representations with flow-matching to enable efficient generation of high-quality long videos. By leveraging a frame-level masking strategy during training, MaskFlow conditions on previously generated unmasked frames to generate videos with lengths ten times beyond that of the training sequences. MaskFlow does so very efficiently by enabling the use of fast Masked Generative Model (MGM)-style sampling and can be deployed in both fully autoregressive as well as full-sequence generation modes. We validate the quality of our method on the FaceForensics (FFS) and Deepmind Lab (DMLab) datasets and report Fr\'echet Video Distance (FVD) competitive with state-of-the-art approaches. We also provide a detailed analysis on the sampling efficiency of our method and demonstrate that MaskFlow can be applied to both timestep-dependent and timestep-independent models in a training-free manner.

Mechanism design has long been a cornerstone of economic theory, with traditional approaches relying on mathematical derivations. Recently, automated approaches, including differentiable economics with neural networks, have emerged for designing payments and allocations. While both analytical and automated methods have advanced the field, they each face significant weaknesses: mathematical derivations are not automated and often struggle to scale to complex problems, while automated and especially neural-network-based approaches suffer from limited interpretability. To address these challenges, we introduce a novel framework that reformulates mechanism design as a code generation task. Using large language models (LLMs), we generate heuristic mechanisms described in code and evolve them to optimize over some evaluation metrics while ensuring key design criteria (e.g., strategy-proofness) through a problem-specific fixing process. This fixing process ensures any mechanism violating the design criteria is adjusted to satisfy them, albeit with some trade-offs in performance metrics. These trade-offs are factored in during the LLM-based evolution process. The code generation capabilities of LLMs enable the discovery of novel and interpretable solutions, bridging the symbolic logic of mechanism design and the generative power of modern AI. Through rigorous experimentation, we demonstrate that LLM-generated mechanisms achieve competitive performance while offering greater interpretability compared to previous approaches. Notably, our framework can rediscover existing manually designed mechanisms and provide insights into neural-network based solutions through Programming-by-Example. These results highlight the potential of LLMs to not only automate but also enhance the transparency and scalability of mechanism design, ensuring safe deployment of the mechanisms in society.

The Athena mission entered a redefinition phase in July 2022, driven by the imperative to reduce the mission cost at completion for the European Space Agency below an acceptable target, while maintaining the flagship nature of its science return. This notably called for a complete redesign of the X-ray Integral Field Unit (X-IFU) cryogenic architecture towards a simpler active cooling chain. Passive cooling via successive radiative panels at spacecraft level is now used to provide a 50 K thermal environment to an X-IFU owned cryostat. 4.5 K cooling is achieved via a single remote active cryocooler unit, while a multi-stage Adiabatic Demagnetization Refrigerator ensures heat lift down to the 50 mK required by the detectors. Amidst these changes, the core concept of the readout chain remains robust, employing Transition Edge Sensor microcalorimeters and a SQUID-based Time-Division Multiplexing scheme. Noteworthy is the introduction of a slower pixel. This enables an increase in the multiplexing factor (from 34 to 48) without compromising the instrument energy resolution, hence keeping significant system margins to the new 4 eV resolution requirement. This allows reducing the number of channels by more than a factor two, and thus the resource demands on the system, while keeping a 4' field of view (compared to 5' before). In this article, we will give an overview of this new architecture, before detailing its anticipated performances. Finally, we will present the new X-IFU schedule, with its short term focus on demonstration activities towards a mission adoption in early 2027., Comment: 44 pages, 14 figures, accepted for publication in Experimental Astronomy

In this article, we are interested in planning problems where the agent is aware of the presence of an observer, and where this observer is in a partial observability situation. The agent has to choose its strategy so as to optimize the information transmitted by observations. Building on observer-aware Markov decision processes (OAMDPs), we propose a framework to handle this type of problems and thus formalize properties such as legibility, explicability and predictability. This extension of OAMDPs to partial observability can not only handle more realistic problems, but also permits considering dynamic hidden variables of interest. These dynamic target variables allow, for instance, working with predictability, or with legibility problems where the goal might change during execution. We discuss theoretical properties of PO-OAMDPs and, experimenting with benchmark problems, we analyze HSVI's convergence behavior with dedicated initializations and study the resulting strategies., Comment: 23 pages, 13 figures. Complete version of AAMAS 2025 extended abstract

An angular analysis of $B^0\rightarrow K^{*0}e^{+}e^{-}$ decays is presented using proton-proton collision data collected by the LHCb experiment at centre-of-mass energies of 7, 8 and 13 TeV, corresponding to an integrated luminosity of 9 fb$^{-1}$. The analysis is performed in the region of the dilepton invariant mass squared of 1.1-6.0 GeV$^{2}/c^{4}$. In addition, a test of lepton flavour universality is performed by comparing the obtained angular observables with those measured in $B^0\rightarrow K^{*0}\mu^{+}\mu^{-}$ decays. In general, the angular observables are found to be consistent with the Standard Model expectations as well as with global analyses of other $b \rightarrow s \ell^{+} \ell^{-}$ processes, where $\ell$ is either a muon or an electron. No sign of lepton-flavour-violating effects is observed., Comment: All figures and tables, along with machine-readable versions and any supplementary material and additional information, are available at https://lbfence.cern.ch/alcm/public/analysis/full-details/1628/ (LHCb public pages)

The stochastic Airy and sine operators, which are respectively a random Sturm-Liouville operator and a random Dirac operator, characterize the soft edge and bulk scaling limits of $\beta$-ensembles. Dirac and Sturm-Liouville operators are distinct operator classes which can both be represented as canonical systems, which gives a unified framework for defining important properties, such as their spectral data. Seeing both as canonical systems, we prove that in a suitable high-energy scaling limit, the Airy operator converges to the sine operator. We prove this convergence in the vague topology of canonical systems' coefficient matrices, and deduce the convergence of the associated Weyl-Titchmarsh functions and spectral measures. Our proof relies on a coupling between the Brownian paths that drive the two operators, under which the convergence holds in probability. This extends the corresponding result at the level of the eigenvalue point processes, proven by Valk\'o and Vir\'ag (2009) by comparison to the Gaussian $\beta$-ensemble., Comment: 62 pages

In this paper, we propose a way to solve partial differential equations (PDEs) by combining machine learning techniques and the finite element method called Phi-FEM. For that, we use the Fourier Neural Operator (FNO), a learning mapping operator. The purpose of this paper is to provide numerical evidence to show the effectiveness of this technique. We will focus here on the resolution of two equations: the Poisson-Dirichlet equation and the non-linear elasticity equations. The key idea of our method is to address the challenging scenario of varying domains, where each problem is solved on a different geometry. The considered domains are defined by level-set functions due to the use of the Phi-FEM approach. We will first recall the idea of $\varphi$-FEM and of the Fourier Neural Operator. Then, we will explain how to combine these two methods. We will finally illustrate the efficiency of this combination with some numerical results on three test cases. In addition, in the last test case, we propose a new numerical scheme for hyperelastic materials following the Phi-FEM paradigm.

We prove that the minimal size $M(\pi_n)$ of a maximal matching in the permutahedron $\pi_n$ is asymptotically $n!/3$. On the one hand, we obtain a lower bound $M(\pi_n) \ge n! (n-1) / (3n-2)$ by considering $4$-cycles in the permutahedron. On the other hand, we obtain an asymptotical upper bound $M(\pi_n) \le n!(1/3+o(1))$ by multiple applications of Hall's theorem (similar to the approach of Forcade (1973) for the hypercube) and an exact upper bound $M(\pi_n) \le n!/3$ by an explicit construction. We also derive bounds on minimum maximal matchings in products of permutahedra., Comment: 11 pages, 2 figures

Representation learning plays a central role in structuring internal embeddings to capture the statistical properties of language, influencing the coherence and contextual consistency of generated text. Statistical Coherence Alignment is introduced as a method to enforce structured token representations through tensor field convergence, guiding embeddings to reflect statistical dependencies inherent in linguistic data. A mathematical framework is established to quantify coherence alignment, integrating a loss function that optimizes representational consistency across training iterations. Empirical evaluations demonstrate that applying coherence constraints improves perplexity, enhances classification accuracy, and refines rare word embeddings, contributing to a more stable representation space. Comparative analyses with baseline models reveal that the proposed method fosters a more interpretable internal structure, ensuring that embeddings retain contextual dependencies while mitigating representation collapse. The impact on coherence score distributions suggests that the alignment mechanism strengthens semantic integrity across diverse linguistic constructs, leading to a more balanced organization of learned embeddings. Computational assessments indicate that while the method introduces additional memory and training costs, the structured optimization process justifies the trade-offs in applications requiring heightened contextual fidelity. Experimental results validate the effectiveness of coherence alignment in optimizing token representations, providing insights into how statistical dependencies can be leveraged to improve language model training.

The differential diagnosis of neurodegenerative diseases, characterized by overlapping symptoms, may be challenging. Brain imaging coupled with artificial intelligence has been previously proposed for diagnostic support, but most of these methods have been trained to discriminate only isolated diseases from controls. Here, we develop a novel machine learning framework, named lifespan tree of brain anatomy, dedicated to the differential diagnosis between multiple diseases simultaneously. It integrates the modeling of volume changes for 124 brain structures during the lifespan with non-linear dimensionality reduction and synthetic sampling techniques to create easily interpretable representations of brain anatomy over the course of disease progression. As clinically relevant proof-of-concept applications, we constructed a cognitive lifespan tree of brain anatomy for the differential diagnosis of six causes of neurodegenerative dementia and a motor lifespan tree of brain anatomy for the differential diagnosis of four causes of parkinsonism using 37594 MRI as a training dataset. This original approach enhanced significantly the efficiency of differential diagnosis in the external validation cohort of 1754 cases, outperforming existing state-of-the art machine learning techniques. Lifespan tree holds promise as a valuable tool for differential diagnostic in relevant clinical conditions, especially for diseases still lacking effective biological markers.

Opto-physiological monitoring is a non-contact technique for measuring cardiac signals, i.e., photoplethysmography (PPG). Quality PPG signals directly lead to reliable physiological readings. However, PPG signal acquisition procedures are often accompanied by spurious motion artefacts (MAs), especially during low-to-high-intensity physical activity. This study proposes a practical adversarial learning approach for opto-physiological monitoring by using a generative adversarial network with an attention mechanism (AM-GAN) to model motion noise and to allow MA removal. The AM-GAN learns an MA-resistant mapping from raw and noisy signals to clear PPG signals in an adversarial manner, guided by an attention mechanism to directly translate the motion reference of triaxial acceleration to the MAs appearing in the raw signal. The AM-GAN was experimented with three various protocols engaged with 39 subjects in various physical activities. The average absolute error for heart rate (HR) derived from the MA-free PPG signal via the AM-GAN, is 1.81 beats/min for the IEEE-SPC dataset and 3.86 beats/min for the PPGDalia dataset. The same procedure applied to an in-house LU dataset resulted in average absolute errors for HR and respiratory rate (RR) of less than 1.37 beats/min and 2.49 breaths/min, respectively. The study demonstrates the robustness and resilience of AM-GAN, particularly during low-to-high-intensity physical activities.

We present a comprehensive computational model to simulate the coupled dynamics of aqueous humor flow and heat transfer in the human eye. To manage the complexity of the model, we make significant efforts in meshing and efficient solution of the discrete problem using high-performance resources. The model accurately describes the dynamics of the aqueous humor in the anterior and posterior chambers and accounts for convective effects due to temperature variations. Results for fluid velocity, pressure, and temperature distribution are in good agreement with existing numerical results in the literature. Furthermore, the effects of postural changes and wall shear stress behavior are analyzed, providing new insights into the mechanical forces acting on ocular tissues. Overall, the present contribution provides a detailed three-dimensional simulation that enhances the understanding of ocular physiology and may contribute to further progress in clinical research and treatment optimization in ophthalmology.

In the differentially private partition selection problem (a.k.a. private set union, private key discovery), users hold subsets of items from an unbounded universe. The goal is to output as many items as possible from the union of the users' sets while maintaining user-level differential privacy. Solutions to this problem are a core building block for many privacy-preserving ML applications including vocabulary extraction in a private corpus, computing statistics over categorical data, and learning embeddings over user-provided items. We propose an algorithm for this problem, MaximumAdaptiveDegree (MAD), which adaptively reroutes weight from items with weight far above the threshold needed for privacy to items with smaller weight, thereby increasing the probability that less frequent items are output. Our algorithm can be efficiently implemented in massively parallel computation systems allowing scalability to very large datasets. We prove that our algorithm stochastically dominates the standard parallel algorithm for this problem. We also develop a two-round version of our algorithm where results of the computation in the first round are used to bias the weighting in the second round to maximize the number of items output. In experiments, our algorithms provide the best results across the board among parallel algorithms and scale to datasets with hundreds of billions of items, up to three orders of magnitude larger than those analyzed by prior sequential algorithms.

Near-infrared spectroscopic surveys target high-redshift emission-line galaxies (ELGs) to probe cosmological scenarios. Understanding the clustering properties of ELGs is essential to derive optimal constraints. We present a simple radiative transfer model for spatially resolved galactic H$\alpha$ emission, which includes emission from the warm-hot diffuse interstellar medium. The atomic level populations are in steady-state and computed in the coronal approximation. The model is applied to multiple IllustrisTNG simulations in the redshift range $1\leq z \leq 2$ to produce the luminosity function (LF) and the halo occupation distribution (HOD). Collisional processes account for a significant fraction of $\approx 40\%$ of the total ${\rm H}\alpha$ luminosity ($L_{{\rm H}\alpha}$). Our LFs are in reasonable agreement with measurements from H$\alpha$ surveys if a uniform extinction of $0.3

Faithfulness evaluators based on large language models (LLMs) are often fooled by the fluency of the text and struggle with identifying errors in the summaries. We propose an approach to summary faithfulness evaluation in which multiple LLM-based agents are assigned initial stances (regardless of what their belief might be) and forced to come up with a reason to justify the imposed belief, thus engaging in a multi-round debate to reach an agreement. The uniformly distributed initial assignments result in a greater diversity of stances leading to more meaningful debates and ultimately more errors identified. Furthermore, by analyzing the recent faithfulness evaluation datasets, we observe that naturally, it is not always the case for a summary to be either faithful to the source document or not. We therefore introduce a new dimension, ambiguity, and a detailed taxonomy to identify such special cases. Experiments demonstrate our approach can help identify ambiguities, and have even a stronger performance on non-ambiguous summaries.

We develop an algorithm to semantically parse linear ordering problems, which require a model to arrange entities using deductive reasoning. Our method takes as input a number of premises and candidate statements, parsing them to a first-order logic of an ordering domain, and then utilizes constraint logic programming to infer the truth of proposed statements about the ordering. Our semantic parser transforms Heim and Kratzer's syntax-based compositional formal semantic rules to a computational algorithm. This transformation involves introducing abstract types and templates based on their rules, and introduces a dynamic component to interpret entities within a contextual framework. Our symbolic system, the Formal Semantic Logic Inferer (FSLI), is applied to answer multiple choice questions in BIG-bench's logical_deduction multiple choice problems, achieving perfect accuracy, compared to 67.06% for the best-performing LLM (GPT-4) and 87.63% for the hybrid system Logic-LM. These promising results demonstrate the benefit of developing a semantic parsing algorithm driven by first-order logic constructs., Comment: 3 figures, 9 pages main paper and 6 pages references and appendix

Simulation-based inference (SBI) is a method to perform inference on a variety of complex scientific models with challenging inference (inverse) problems. Bayesian Optimal Experimental Design (BOED) aims to efficiently use experimental resources to make better inferences. Various stochastic gradient-based BOED methods have been proposed as an alternative to Bayesian optimization and other experimental design heuristics to maximize information gain from an experiment. We demonstrate a link via mutual information bounds between SBI and stochastic gradient-based variational inference methods that permits BOED to be used in SBI applications as SBI-BOED. This link allows simultaneous optimization of experimental designs and optimization of amortized inference functions. We evaluate the pitfalls of naive design optimization using this method in a standard SBI task and demonstrate the utility of a well-chosen design distribution in BOED. We compare this approach on SBI-based models in real-world simulators in epidemiology and biology, showing notable improvements in inference., Comment: Preprint. Under Review

Web browsing agents powered by large language models (LLMs) have shown tremendous potential in automating complex web-based tasks. Existing approaches typically rely on large LLMs (e.g., GPT-4o) to explore web environments and generate trajectory data, which is then used either for demonstration retrieval (for large LLMs) or to distill small LLMs (e.g., Llama3) in a process that remains decoupled from the exploration. In this paper, we propose AgentSymbiotic, an iterative framework that couples data synthesis with task-performance, yielding a "symbiotic improvement" for both large and small LLMs. Our study uncovers a complementary dynamic between LLM types: while large LLMs excel at generating high-quality trajectories for distillation, the distilled small LLMs-owing to their distinct reasoning capabilities-often choose actions that diverge from those of their larger counterparts. This divergence drives the exploration of novel trajectories, thereby enriching the synthesized data. However, we also observe that the performance of small LLMs becomes a bottleneck in this iterative enhancement process. To address this, we propose two innovations in LLM distillation: a speculative data synthesis strategy that mitigates off-policy bias, and a multi-task learning approach designed to boost the reasoning capabilities of the student LLM. Furthermore, we introduce a Hybrid Mode for Privacy Preservation to address user privacy concerns. Evaluated on the WEBARENA benchmark, AgentSymbiotic achieves SOTA performance with both LLM types. Our best Large LLM agent reaches 52%, surpassing the previous best of 45%, while our 8B distilled model demonstrates a competitive 49%, exceeding the prior best of 28%. Code will be released upon acceptance.

Therapeutic advancements in oncology have shifted towards targeted therapy based on genomic aberrations. This necessitates innovative statistical approaches in clinical trials, particularly in master protocol studies. Basket trials, a type of master protocol, evaluate a single treatment across cohorts sharing a genomic aberration but differing in tumor histology. While offering operational advantages, basket trial analysis presents statistical inference challenges. These trials help determine for which tumor histology the treatment is promising enough to advance to confirmatory evaluation and often use Bayesian designs to support decisions. Beyond decision-making, estimating cohort-specific response rates is crucial for designing subsequent trials. This study compares seven Bayesian estimation methods for basket trials with binary outcomes against the (frequentist) sample proportion estimate through simulations. The goal is to estimate cohort-specific response rates, focusing on bias, mean squared error, and information borrowing. Various scenarios are examined, including homogeneous, heterogeneous, and clustered response rates across cohorts. Results show trade-offs in bias and precision, highlighting the importance of method selection. Berry's method performs best with limited heterogeneity. No clear winner emerges in general cases, with performance affected by shrinkage, bias, and the choice of priors and tuning parameters. Challenges include computational complexity, parameter tuning, and the lack of clear guidance on selection. Researchers should consider these factors when designing and analyzing basket trials.

We present a novel extension of the Weak Neural Variational Inference (WNVI) framework for probabilistic material property estimation that explicitly quantifies model errors in PDE-based inverse problems. Traditional approaches assume the correctness of all governing equations, including potentially unreliable constitutive laws, which can lead to biased estimates and misinterpretations. Our proposed framework addresses this limitation by distinguishing between reliable governing equations, such as conservation laws, and uncertain constitutive relationships. By treating all state variables as latent random variables, we enforce these equations through separate sets of residuals, leveraging a virtual likelihood approach with weighted residuals. This formulation not only identifies regions where constitutive laws break down but also improves robustness against model uncertainties without relying on a fully trustworthy forward model. We demonstrate the effectiveness of our approach in the context of elastography, showing that it provides a structured, interpretable, and computationally efficient alternative to traditional model error correction techniques. Our findings suggest that the proposed framework enhances the accuracy and reliability of material property estimation by offering a principled way to incorporate uncertainty in constitutive modeling., Comment: 15 pages, 6 figures

Rayleigh-Bloch waves are guided acoustic waves propagating along a periodic line of inclusions placed inside an open, infinite medium. Below the sound cone, they are transversely evanescent on both sides of the line of inclusions. Guidance is then achieved without any cladding surrounding the segmented core. Inclusions usually impose definite boundary conditions, resulting in a single guided band. We consider instead the case of permeable, slow inclusions inside a fast medium. Introducing the concept of guided quasi-normal modes, we obtain the complex dispersion relation taking into account radiation at infinity. We thus show that multiple bands of leaky Rayleigh-Bloch waves appear and that guided bound states in the continuum arise as a result of the combination of symmetry and periodicity.

Arrays of optical tweezers form the backbone of neutral atoms analog and digital quantum processors. However, the inter-trap distance remains generally much larger than the size of the tweezers to avoid interference-induced trap distortions, limiting the trap density. Here, we report single-atom trapping in four super-resolved tweezers, meaning with a separation below the Sparrow diffraction limit. The optical pattern is generated using superoscillatory phenomenon leading to subwavelength traps with full control of the trap relative phases. We investigate two sets of relative phases that impede or allow the hopping and the reshuffling of atoms. We envision that superoscillatory light structuring will bridge the gap between large-distance traps generated by tweezer arrays and short-distance traps formed with optical lattices., Comment: 6 pages, 4 figures. SM is available upon request

Composite Higgs models are a class of Beyond the Standard Model (BSM) models proposed to address the hierarchy and naturalness problems associated with the Standard Model (SM) Higgs. A new QCD-like strongly interacting sector based on $SU(2)$ with two fundamental flavours can be used to build a composite Higgs model which is not yet ruled out by experiment. The role of the singlet scalar resonance will affect Higgs phenomenology at the LHC. In this project our goal is to understand the properties of the singlet scalar state in the new strongly interacting sector in isolation as a first step to understanding the role of this state in composite Higgs models. We present here the first lattice results for the mass of the $\sigma$ in $SU(2)$ with two fundamental flavours using exponential clover Wilson fermions., Comment: Proceedings of The 41st International Symposium on Lattice Field Theory (Lattice 2024)

We present Cryoscope -- a new 50 sq. deg field-of-view, 1.2 m aperture, K-dark survey telescope to be located at Dome C, Antarctica. Cryoscope has an innovative optical-thermal design wherein the entire telescope is cryogenically cooled. Cryoscope also explores new detector technology to cost-effectively tile the full focal plane. Leveraging the dark Antarctic sky and minimizing telescope thermal emission, Cryoscope achieves unprecedented deep, wide, fast and red observations, matching and exceeding volumetric survey speeds from the Ultraviolet Explorer, Vera Rubin Observatory, and Nancy Grace Roman Space Telescope. By providing coverage beyond wavelengths of 2 $\mu$m, we aim to create the most comprehensive dynamic movie of the most obscured reaches of the Universe. Cryoscope will be a dedicated discovery engine for electromagnetic emission from coalescing compact binaries, Earth-like exoplanets orbiting cold stars, and multiple facets of time-domain, stellar and solar system science. In this paper, we describe the scientific drivers and technical innovations for this new discovery engine operating in the K-dark passband, why we choose to deploy it in Antarctica, and the status of a fifth-scale prototype designed as a Pathfinder to retire technological risks prior to full-scale implementation., Comment: 36 pages, 20 figures, 4 tables; submitted to PASP on 2025-02-09

Classifier-free guidance (CFG) is a key technique for improving conditional generation in diffusion models, enabling more accurate control while enhancing sample quality. It is natural to extend this technique to video diffusion, which generates video conditioned on a variable number of context frames, collectively referred to as history. However, we find two key challenges to guiding with variable-length history: architectures that only support fixed-size conditioning, and the empirical observation that CFG-style history dropout performs poorly. To address this, we propose the Diffusion Forcing Transformer (DFoT), a video diffusion architecture and theoretically grounded training objective that jointly enable conditioning on a flexible number of history frames. We then introduce History Guidance, a family of guidance methods uniquely enabled by DFoT. We show that its simplest form, vanilla history guidance, already significantly improves video generation quality and temporal consistency. A more advanced method, history guidance across time and frequency further enhances motion dynamics, enables compositional generalization to out-of-distribution history, and can stably roll out extremely long videos. Website: https://boyuan.space/history-guidance, Comment: Project Website: https://boyuan.space/history-guidance

Early Classification of Time Series (ECTS) has been recognized as an important problem in many areas where decisions have to be taken as soon as possible, before the full data availability, while time pressure increases. Numerous ECTS approaches have been proposed, based on different triggering functions, each taking into account various pieces of information related to the incoming time series and/or the output of a classifier. Although their performances have been empirically compared in the literature, no studies have been carried out on the optimality of these triggering functions that involve ``man-tailored'' decision rules. Based on the same information, could there be better triggering functions? This paper presents one way to investigate this question by showing first how to translate ECTS problems into Reinforcement Learning (RL) ones, where the very same information is used in the state space. A thorough comparison of the performance obtained by ``handmade'' approaches and their ``RL-based'' counterparts has been carried out. A second question investigated in this paper is whether a different combination of information, defining the state space in RL systems, can achieve even better performance. Experiments show that the system we describe, called \textsc{Alert}, significantly outperforms its state-of-the-art competitors on a large number of datasets.

Kolmogorov--Arnold networks (KANs) have demonstrated their potential as an alternative to multi-layer perceptions (MLPs) in various domains, especially for science-related tasks. However, transfer learning of KANs remains a relatively unexplored area. In this paper, inspired by Tucker decomposition of tensors and evidence on the low tensor-rank structure in KAN parameter updates, we develop low tensor-rank adaptation (LoTRA) for fine-tuning KANs. We study the expressiveness of LoTRA based on Tucker decomposition approximations. Furthermore, we provide a theoretical analysis to select the learning rates for each LoTRA component to enable efficient training. Our analysis also shows that using identical learning rates across all components leads to inefficient training, highlighting the need for an adaptive learning rate strategy. Beyond theoretical insights, we explore the application of LoTRA for efficiently solving various partial differential equations (PDEs) by fine-tuning KANs. Additionally, we propose Slim KANs that incorporate the inherent low-tensor-rank properties of KAN parameter tensors to reduce model size while maintaining superior performance. Experimental results validate the efficacy of the proposed learning rate selection strategy and demonstrate the effectiveness of LoTRA for transfer learning of KANs in solving PDEs. Further evaluations on Slim KANs for function representation and image classification tasks highlight the expressiveness of LoTRA and the potential for parameter reduction through low tensor-rank decomposition.

This paper considers guessing-based decoders with abandonment for discrete memoryless channels in which all codewords have the same composition. This class of decoders rank-orders all input sequences in the codebook's composition class from ``closest'' to ``farthest'' from the channel output and then queries them sequentially in that order for codebook membership. Decoding terminates when a codeword is encountered or when a predetermined number of guesses is reached, and decoding is abandoned. We derive ensemble-tight first-order asymptotics for the code rate and abandonment rate, which shows that guessing-based decoding is more efficient than conventional testing-based decoding whenever the capacity of the channel exceeds half the entropy of the capacity-achieving input distribution. The main focus of this paper is on refined asymptotics, specifically, second-order asymptotics, error exponents, and strong converse exponents. The optimal second-order region is characterized in terms of the minimum of the second-order code and abandonment rates. The error (resp.\ strong converse) exponent is characterized in terms of the minimum (resp.\ maximum) of the usual channel coding exponent and an abandonment exponent, which turns out to be a special case of the exponent of conditional almost-lossless source coding., Comment: 12 pages; Submitted in part to ISIT 2025

After a successful kick-off meeting in 2021. two workshops in 2022 and 2023 on the future Global Cosmic-Ray Observatory (GCOS) focused mainly on a straw man design of the detector and science possibilities for astro- and particle physics. About 100 participants gathered for in-person and hybrid panel discussions. In this report, we summarize these discussions, present a preliminary straw-man design for GCOS and collect short write-ups of the flash talks given during the focus sessions., Comment: 48 pages, 27 figures

With the goal of ubiquitous global connectivity, multibeam low Earth orbit (LEO) satellite communication (SATCOM) has attracted significant attention in recent years. The traffic demands of users are heterogeneous within the broad coverage of SATCOM due to different geological conditions and user distributions. Motivated by this, this paper proposes a novel rate-matching (RM) framework based on rate-splitting multiple access (RSMA) that minimizes the difference between the traffic demands and offered rates while simultaneously minimizing transmit power for power-hungry satellite payloads. Moreover, channel phase perturbations arising from channel estimation and feedback errors are considered to capture realistic multibeam LEO SATCOM scenarios. To tackle the non-convexity of the RSMA-based RM problem under phase perturbations, we convert it into a tractable convex form via the successive convex approximation method and present an efficient algorithm to solve the RM problem. Through the extensive numerical analysis across various traffic demand distribution and channel state information accuracy at LEO satellites, we demonstrate that RSMA flexibly allocates the power between common and private streams according to different traffic patterns across beams, thereby efficiently satisfying users non-uniform traffic demands. In particular, the use of common messages plays a vital role in overcoming the limited spatial dimension available at LEO satellites, enabling it to manage inter- and intra-beam interference effectively in the presence of phase perturbation., Comment: 42 pages, 15 figures, 1 table, accepted by IEEE Transactions on Signal Processing