Search

Vertical Federated Learning (VFL) aims to enable collaborative training of deep learning models while maintaining privacy protection. However, the VFL procedure still has components that are vulnerable to attacks by malicious parties. In our work, we consider feature reconstruction attacks, a common risk targeting input data compromise. We theoretically claim that feature reconstruction attacks cannot succeed without knowledge of the prior distribution on data. Consequently, we demonstrate that even simple model architecture transformations can significantly impact the protection of input data during VFL. Confirming these findings with experimental results, we show that MLP-based models are resistant to state-of-the-art feature reconstruction attacks., Comment: 29 pages, 12 figures, 3 tables

The standard model for planet formation is a bottom-up process in which the origin of rocky and gaseous planets can be traced back to the collision of micron-sized dust grains within the gas-rich environment of protoplanetary disks. Key milestones along the way include disk formation, grain growth, planetesimal formation, core growth, gas accretion, and planetary system evolution. I provide an introductory overview of planet formation, emphasizing the main ideas and reviewing current theoretical understanding. Many of the phases of planet formation have a well-developed physical understanding, though the complexity of the problem means that few can be quantitatively modeled with complete confidence. Transformative advances in disk imaging and exoplanet detection provide the first direct information on the initial conditions for planet formation, and have motivated new formation models that are faster, more efficient, and lead to a more diverse set of architectures than their Solar System inspired forebears. Much remains to be learned, and I close with a personal, incomplete list, of open problems., Comment: To be published in: Handbook of Exoplanets, 2nd Edition, Hans Deeg and Juan Antonio Belmonte (Eds. in Chief), Springer International Publishing AG, part of Springer Nature

The vast, complex, and dynamic nature of social message data has posed challenges to social event detection (SED). Despite considerable effort, these challenges persist, often resulting in inadequately expressive message representations (ineffective) and prolonged learning durations (inefficient). In response to the challenges, this work introduces an unsupervised framework, HyperSED (Hyperbolic SED). Specifically, the proposed framework first models social messages into semantic-based message anchors, and then leverages the structure of the anchor graph and the expressiveness of the hyperbolic space to acquire structure- and geometry-aware anchor representations. Finally, HyperSED builds the partitioning tree of the anchor message graph by incorporating differentiable structural information as the reflection of the detected events. Extensive experiments on public datasets demonstrate HyperSED's competitive performance, along with a substantial improvement in efficiency compared to the current state-of-the-art unsupervised paradigm. Statistically, HyperSED boosts incremental SED by an average of 2%, 2%, and 25% in NMI, AMI, and ARI, respectively; enhancing efficiency by up to 37.41 times and at least 12.10 times, illustrating the advancement of the proposed framework. Our code is publicly available at https://github.com/XiaoyanWork/HyperSED., Comment: Accepted to AAAI 2025

Text-to-image (T2I) models have become widespread, but their limited safety guardrails expose end users to harmful content and potentially allow for model misuse. Current safety measures are typically limited to text-based filtering or concept removal strategies, able to remove just a few concepts from the model's generative capabilities. In this work, we introduce SafetyDPO, a method for safety alignment of T2I models through Direct Preference Optimization (DPO). We enable the application of DPO for safety purposes in T2I models by synthetically generating a dataset of harmful and safe image-text pairs, which we call CoProV2. Using a custom DPO strategy and this dataset, we train safety experts, in the form of low-rank adaptation (LoRA) matrices, able to guide the generation process away from specific safety-related concepts. Then, we merge the experts into a single LoRA using a novel merging strategy for optimal scaling performance. This expert-based approach enables scalability, allowing us to remove 7 times more harmful concepts from T2I models compared to baselines. SafetyDPO consistently outperforms the state-of-the-art on many benchmarks and establishes new practices for safety alignment in T2I networks. Code and data will be shared at https://safetydpo.github.io/.

Large language models (LLMs) frequently generate confident yet inaccurate responses, introducing significant risks for deployment in safety-critical domains. We present a novel approach to detecting model hallucination through systematic analysis of information flow across model layers when processing inputs with insufficient or ambiguous context. Our investigation reveals that hallucination manifests as usable information deficiencies in inter-layer transmissions. While existing approaches primarily focus on final-layer output analysis, we demonstrate that tracking cross-layer information dynamics ($\mathcal{L}$I) provides robust indicators of model reliability, accounting for both information gain and loss during computation. $\mathcal{L}$I improves model reliability by immediately integrating with universal LLMs without additional training or architectural modifications.

This study seeks to automate camera movement control for filming existing subjects into attractive videos, contrasting with the creation of non-existent content by directly generating the pixels. We select drone videos as our test case due to their rich and challenging motion patterns, distinctive viewing angles, and precise controls. Existing AI videography methods struggle with limited appearance diversity in simulation training, high costs of recording expert operations, and difficulties in designing heuristic-based goals to cover all scenarios. To avoid these issues, we propose a scalable method that involves collecting real-world training data to improve diversity, extracting camera trajectories automatically to minimize annotation costs, and training an effective architecture that does not rely on heuristics. Specifically, we collect 99k high-quality trajectories by running 3D reconstruction on online videos, connecting camera poses from consecutive frames to formulate 3D camera paths, and using Kalman filter to identify and remove low-quality data. Moreover, we introduce DVGFormer, an auto-regressive transformer that leverages the camera path and images from all past frames to predict camera movement in the next frame. We evaluate our system across 38 synthetic natural scenes and 7 real city 3D scans. We show that our system effectively learns to perform challenging camera movements such as navigating through obstacles, maintaining low altitude to increase perceived speed, and orbiting towers and buildings, which are very useful for recording high-quality videos. Data and code are available at dvgformer.github.io.

We introduce Olympus, a new approach that transforms Multimodal Large Language Models (MLLMs) into a unified framework capable of handling a wide array of computer vision tasks. Utilizing a controller MLLM, Olympus delegates over 20 specialized tasks across images, videos, and 3D objects to dedicated modules. This instruction-based routing enables complex workflows through chained actions without the need for training heavy generative models. Olympus easily integrates with existing MLLMs, expanding their capabilities with comparable performance. Experimental results demonstrate that Olympus achieves an average routing accuracy of 94.75% across 20 tasks and precision of 91.82% in chained action scenarios, showcasing its effectiveness as a universal task router that can solve a diverse range of computer vision tasks. Project page: http://yuanze-lin.me/Olympus_page/, Comment: Technical Report

Probabilistic embeddings have several advantages over deterministic embeddings as they map each data point to a distribution, which better describes the uncertainty and complexity of data. Many works focus on adjusting the distribution constraint under the Information Bottleneck (IB) principle to enhance representation learning. However, these proposed regularization terms only consider the constraint of each latent variable, omitting the structural information between latent variables. In this paper, we propose a novel structural entropy-guided probabilistic coding model, named SEPC. Specifically, we incorporate the relationship between latent variables into the optimization by proposing a structural entropy regularization loss. Besides, as traditional structural information theory is not well-suited for regression tasks, we propose a probabilistic encoding tree, transferring regression tasks to classification tasks while diminishing the influence of the transformation. Experimental results across 12 natural language understanding tasks, including both classification and regression tasks, demonstrate the superior performance of SEPC compared to other state-of-the-art models in terms of effectiveness, generalization capability, and robustness to label noise. The codes and datasets are available at https://github.com/SELGroup/SEPC., Comment: This paper is accepted by AAAI 2025

We propose a fully Bayesian approach to wideband, or broadband, direction-of-arrival (DoA) estimation and signal detection. Unlike previous works in wideband DoA estimation and detection, where the signals were modeled in the time-frequency domain, we directly model the time-domain representation and treat the non-causal part of the source signal as latent variables. Furthermore, our Bayesian model allows for closed-form marginalization of the latent source signals by leveraging conjugacy. To further speed up computation, we exploit the sparse ``stripe matrix structure'' of the considered system, which stems from the circulant matrix representation of linear time-invariant (LTI) systems. This drastically reduces the time complexity of computing the likelihood from $\mathcal{O}(N^3 k^3)$ to $\mathcal{O}(N k^3)$, where $N$ is the number of samples received by the array and $k$ is the number of sources. These computational improvements allow for efficient posterior inference through reversible jump Markov chain Monte Carlo (RJMCMC). We use the non-reversible extension of RJMCMC (NRJMCMC), which often achieves lower autocorrelation and faster convergence than the conventional reversible variant. Detection, estimation, and reconstruction of the latent source signals can then all be performed in a fully Bayesian manner through the samples drawn using NRJMCMC. We evaluate the detection performance of the procedure by comparing against generalized likelihood ratio testing (GLRT) and information criteria.

Clinical prognostic models help inform decision-making by estimating a patient's risk of experiencing an outcome in the future. The net benefit is increasingly being used to assess the clinical utility of models. By calculating an appropriately weighted average of the true and false positives of a model, the net benefit assesses the value added by a binary decision policy obtained when thresholding a model. Although such 'treat or not' decisions are common, prognostic models are also often used to tailor and personalise the care of patients, which implicitly involves the consideration of multiple interventions at different risk thresholds. We extend the net benefit to consider multiple decision thresholds simultaneously, by taking a weighted area under a rescaled version of the net benefit curve, deriving the continuous net benefit. In addition to the consideration of a continuum of interventions, we also show how the continuous net benefit can be used for populations with a range of optimal thresholds for a single treatment, due to individual variations in expected treatment benefit or harm, highlighting limitations of current proposed methods that calculate the area under the decision curve. We showcase the continuous net benefit through two examples of cardiovascular preventive care, comparing two modelling choices using the continuous net benefit. The continuous net benefit informs researchers of the clinical utility of models during selection, development, and validation, and helps decision makers understand their usefulness, improving their viability towards implementation., Comment: Main text is 17 pages, 3 figures. Supplementary is 9 pages. Submitted to Epidemiology

We propose DiTFlow, a method for transferring the motion of a reference video to a newly synthesized one, designed specifically for Diffusion Transformers (DiT). We first process the reference video with a pre-trained DiT to analyze cross-frame attention maps and extract a patch-wise motion signal called the Attention Motion Flow (AMF). We guide the latent denoising process in an optimization-based, training-free, manner by optimizing latents with our AMF loss to generate videos reproducing the motion of the reference one. We also apply our optimization strategy to transformer positional embeddings, granting us a boost in zero-shot motion transfer capabilities. We evaluate DiTFlow against recently published methods, outperforming all across multiple metrics and human evaluation., Comment: Project page: https://ditflow.github.io/

In recent years, large language models (LLMs) have been widely adopted in political science tasks such as election prediction, sentiment analysis, policy impact assessment, and misinformation detection. Meanwhile, the need to systematically understand how LLMs can further revolutionize the field also becomes urgent. In this work, we--a multidisciplinary team of researchers spanning computer science and political science--present the first principled framework termed Political-LLM to advance the comprehensive understanding of integrating LLMs into computational political science. Specifically, we first introduce a fundamental taxonomy classifying the existing explorations into two perspectives: political science and computational methodologies. In particular, from the political science perspective, we highlight the role of LLMs in automating predictive and generative tasks, simulating behavior dynamics, and improving causal inference through tools like counterfactual generation; from a computational perspective, we introduce advancements in data preparation, fine-tuning, and evaluation methods for LLMs that are tailored to political contexts. We identify key challenges and future directions, emphasizing the development of domain-specific datasets, addressing issues of bias and fairness, incorporating human expertise, and redefining evaluation criteria to align with the unique requirements of computational political science. Political-LLM seeks to serve as a guidebook for researchers to foster an informed, ethical, and impactful use of Artificial Intelligence in political science. Our online resource is available at: http://political-llm.org/., Comment: 54 Pages, 9 Figures

Physical data are representations of the fundamental laws governing the Universe, hiding complex compositional structures often well captured by hierarchical graphs. Hyperbolic spaces are endowed with a non-Euclidean geometry that naturally embeds those structures. To leverage the benefits of non-Euclidean geometries in representing natural data we develop machine learning on $\mathcal P \mathcal M$ spaces, Cartesian products of constant curvature Riemannian manifolds. As a use case we consider the classification of "jets", sprays of hadrons and other subatomic particles produced by the hadronization of quarks and gluons in collider experiments. We compare the performance of $\mathcal P \mathcal M$-MLP and $\mathcal P \mathcal M$-Transformer models across several possible representations. Our experiments show that $\mathcal P \mathcal M$ representations generally perform equal or better to fully Euclidean models of similar size, with the most significant gains found for highly hierarchical jets and small models. We discover significant correlation between the degree of hierarchical structure at a per-jet level and classification performance with the $\mathcal P \mathcal M$-Transformer in top tagging benchmarks. This is a promising result highlighting a potential direction for further improving machine learning model performance through tailoring geometric representation at a per-sample level in hierarchical datasets. These results reinforce the view of geometric representation as a key parameter in maximizing both performance and efficiency of machine learning on natural data.

Precision radial velocity (RV) spectrographs that use adaptive optics (AO) show promise to advance telescope observing capabilities beyond those of seeing-limited designs. We are building a spectrograph for the Large Binocular Telescope (LBT) named iLocater that uses AO to inject starlight directly into single mode fibers (SMF). iLocater's first acquisition camera system (the `SX' camera), which receives light from one of the 8.4m diameter primary mirrors of the LBT, was initially installed in summer 2019 and has since been used for several commissioning runs. We present results from first-light observations that include on-sky measurements as part of commissioning activities. Imaging measurements of the bright B3IV star 2 Cygni ($V=4.98$) resulted in the direct detection of a candidate companion star at an angular separation of only $\theta = 70$ mas. Follow-up AO measurements using Keck/NIRC2 recover the candidate companion in multiple filters. An $R\approx1500$ miniature spectrograph recently installed at the LBT named ``Lili'' provides spatially resolved spectra of each binary component, indicating similar spectral types and strengthening the case for companionship. Studying the multiplicity of young runaway star systems like 2 Cygni ($36.6 \pm 0.5$ Myr) can help to understand formation mechanisms for stars that exhibit anomalous velocities through the galaxy. This on-sky demonstration illustrates the spatial resolution of the iLocater SX acquisition camera working in tandem with the LBT AO system; it further derisks a number of technical hurdles involved in combining AO with Doppler spectroscopy., Comment: Accepted to AJ

The transcorrelated (TC) method performs a similarity transformation on the electronic Schr\"odinger equation via Jastrow factorization of the wave function. This has demonstrated significant advancements in computational electronic structure theory by improving basis set convergence and compactifying the description of the wave function. In this work, we introduce a new approach that incorporates pseudopotentials (PPs) into the TC framework, significantly accelerating Jastrow factor optimization and reducing computational costs. Our results for ionization potentials, atomization energies, and dissociation curves of first-row atoms and molecules show that PPs provide chemically accurate descriptions across a range of systems and give guidelines for future theory and applications. The new pseudopotential-based TC method opens possibilities for applying TC to more complex and larger systems, such as transition metals and solid-state systems.

We present multi-frequency radio data for a sample of narrow-line Seyfert 1 galaxies. We first focus on the sub-class of gamma-ray emitting narrow-line Seyfert 1 galaxies, studying the long-term radio variability of five sources and comparing it to their gamma-ray state. We then extend the observations of the southern narrow-line Seyfert 1 galaxy sample of Chen et al. by observing several candidate narrow-line Seyfert 1 sources for the first time, and re-observing several other gamma-ray quiet sources to obtain a first indication of their radio variability. We find that the gamma-ray emitting narrow-line Seyfert 1 galaxies are highly variable radio emitters and that there are instances of contemporaneous flaring activity between the radio and gamma-ray band (PKS 0440$-$00, PMN J0948+0022 and PKS 1244$-$255). However, there are also cases of significant radio outbursts without gamma-ray counterparts (PMN J0948+0022 and PKS 2004$-$447). The five gamma-ray NLS1s favour flat or inverted radio spectra, although the spectral indices vary significantly over time. For the gamma-ray quiet sample, the difference between the previous observations at 5.5 GHz and new ATCA observations indicates that over half of the 14 sources exhibit apparent variability. In contrast to gamma-ray loud sources, gamma-ray quiet objects tend to have steep spectra especially in the lower radio band (887.5$-$1367.5 MHz), with a number of the variable sources having flatter spectra at higher radio frequencies., Comment: Accepted for publication in MNRAS (13 pages, 7 figures)

Time series anomaly detection (TSAD) has been a research hotspot in both academia and industry in recent years. Deep learning methods have become the mainstream research direction due to their excellent performance. However, new viewpoints have emerged in recent TSAD research. Deep learning is not required for TSAD due to limitations such as slow deep learning speed. The Broad Learning System (BLS) is a shallow network framework that benefits from its ease of optimization and speed. It has been shown to outperform machine learning approaches while remaining competitive with deep learning. Based on the current situation of TSAD, we propose the Contrastive Patch-based Broad Learning System (CPatchBLS). This is a new exploration of patching technique and BLS, providing a new perspective for TSAD. We construct Dual-PatchBLS as a base through patching and Simple Kernel Perturbation (SKP) and utilize contrastive learning to capture the differences between normal and abnormal data under different representations. To compensate for the temporal semantic loss caused by various patching, we propose CPatchBLS with model level integration, which takes advantage of BLS's fast feature to build model-level integration and improve model detection. Using five real-world series anomaly detection datasets, we confirmed the method's efficacy, outperforming previous deep learning and machine learning methods while retaining a high level of computing efficiency., Comment: 13 pages, 7 figures, 3 tables, Under review

Side channels have become an essential component of many modern information-theoretic schemes. The emerging field of cross technology communications (CTC) provides practical methods for creating intentional side channels between existing communications technologies. This paper describes a theoretical foundation for one such, recently proposed, CTC scheme: Ghost Modulation (GM). Designed to modulate a low-data-rate message atop an existing network stream, GM is particularly suited for transmitting identification or covert information. The implementation only requires firmware updates to existing hardware, making it a cost-effective solution. However, GM provides an interesting technical challenge due to a highly asymmetric binary crossover erasure channel (BCEC) that results from packet drops and network delays. In this work, we provide a mathematical description of the signal and channel models for GM. A heuristic decision rule based on maximum-likelihood principles for simplified channel models is proposed. We describe an algorithm for GM packet acquisition and timing synchronization, supported by simulation results. Several well known short block codes are applied, and bit error rate (BER) results are presented., Comment: 7 Pages, 5 Figures

Large Audio-Language Models (LALMs) have unclocked audio dialogue capabilities, where audio dialogues are a direct exchange of spoken language between LALMs and humans. Recent advances, such as GPT-4o, have enabled LALMs in back-and-forth audio dialogues with humans. This progression not only underscores the potential of LALMs but also broadens their applicability across a wide range of practical scenarios supported by audio dialogues. However, given these advancements, a comprehensive benchmark to evaluate the performance of LALMs in the open-ended audio dialogue understanding remains absent currently. To address this gap, we propose an Audio Dialogue Understanding Benchmark (ADU-Bench), which consists of 4 benchmark datasets. They assess the open-ended audio dialogue ability for LALMs in 3 general scenarios, 12 skills, 9 multilingual languages, and 4 categories of ambiguity handling. Notably, we firstly propose the evaluation of ambiguity handling in audio dialogues that expresses different intentions beyond the same literal meaning of sentences, e.g., "Really!?" with different intonations. In summary, ADU-Bench includes over 20,000 open-ended audio dialogues for the assessment of LALMs. Through extensive experiments conducted on 13 LALMs, our analysis reveals that there is still considerable room for improvement in the audio dialogue understanding abilities of existing LALMs. In particular, they struggle with mathematical symbols and formulas, understanding human behavior such as roleplay, comprehending multiple languages, and handling audio dialogue ambiguities from different phonetic elements, such as intonations, pause positions, and homophones.

We consider succinct data structures for representing a set of $n$ horizontal line segments in the plane given in rank space to support \emph{segment access}, \emph{segment selection}, and \emph{segment rank} queries. A segment access query finds the segment $(x_1, x_2, y)$ given its $y$-coordinate ($y$-coordinates of the segments are distinct), a segment selection query finds the $j$th smallest segment (the segment with the $j$th smallest $y$-coordinate) among the segments crossing the vertical line for a given $x$-coordinate, and a segment rank query finds the number of segments crossing the vertical line through $x$-coordinate $i$ with $y$-coordinate at most $y$, for a given $x$ and $y$. This problem is a central component in compressed data structures for persistent strings supporting random access. Our main result is data structure using $2n\lg{n} + O(n\lg{n}/\lg{\lg{n}})$ bits of space and $O(\lg{n}/\lg{\lg{n}})$ query time for all operations. We show that this space bound is optimal up to lower-order terms. We will also show that the query time for segment rank is optimal. The query time for segment selection is also optimal by a previous bound. To obtain our results, we present a novel segment wavelet tree data structure of independent interest. This structure is inspired by and extends the classic wavelet tree for sequences. This leads to a simple, succinct solution with $O(\log n)$ query times. We then extend this solution to obtain optimal query time. Our space lower bound follows from a simple counting argument, and our lower bound for segment rank is obtained by a reduction from 2-dimensional counting.

Stroke is a major global health problem that causes mortality and morbidity. Predicting the outcomes of stroke intervention can facilitate clinical decision-making and improve patient care. Engaging and developing deep learning techniques can help to analyse large and diverse medical data, including brain scans, medical reports and other sensor information, such as EEG, ECG, EMG and so on. Despite the common data standardisation challenge within medical image analysis domain, the future of deep learning in stroke outcome prediction lie in using multimodal information, including final infarct data, to achieve better prediction of long-term functional outcomes. This article provides a broad review of recent advances and applications of deep learning in the prediction of stroke outcomes, including (i) the data and models used, (ii) the prediction tasks and measures of success, (iii) the current challenges and limitations, and (iv) future directions and potential benefits. This comprehensive review aims to provide researchers, clinicians, and policy makers with an up-to-date understanding of this rapidly evolving and promising field., Comment: The paper is under consideration at Biomedical Engineering Letters (Springer)

We prove two results concerning percolation on general graphs. - We establish the converse of the classical Peierls argument: if the critical parameter for (uniform) percolation satisfies $p_c<1$, then the number of minimal cutsets of size $n$ separating a given vertex from infinity is bounded above exponentially in $n$. This resolves a conjecture of Babson and Benjamini from 1999. - We prove that $p_c<1$ for every uniformly transient graph. This solves a problem raised by Duminil-Copin, Goswami, Raoufi, Severo and Yadin, and provides a new proof that $p_c<1$ for every transitive graph of superlinear growth., Comment: 12 pages

Graph-based and sequential methods are two popular recommendation paradigms, each excelling in its domain but lacking the ability to leverage signals from the other. To address this, we propose a novel method that integrates both approaches for enhanced performance. Our framework uses Graph Neural Network (GNN)-based and sequential recommenders as separate submodules while sharing a unified embedding space optimized jointly. To enable positive knowledge transfer, we design a loss function that enforces alignment and uniformity both within and across submodules. Experiments on three real-world datasets demonstrate that the proposed method significantly outperforms using either approach alone and achieves state-of-the-art results. Our implementations are publicly available at https://github.com/YuweiCao-UIC/GSAU.git., Comment: Accepted to The Web Conference 2025

The accurate computational study of wavepacket nuclear dynamics is considered to be a classically intractable problem, particularly with increasing dimensions. Here we present two algorithms that, in conjunction with other methods developed by us, will form the basis for performing quantum nuclear dynamics in arbitrary dimensions. For one algorithm, we present a direct map between the Born-Oppenheimer Hamiltonian describing the wavepacket time-evolution and the control parameters of a spin-lattice Hamiltonian that describes the dynamics of qubit states in an ion-trap quantum computer. This map is exact for three qubits, and when implemented, the dynamics of the spin states emulate those of the nuclear wavepacket. However, this map becomes approximate as the number of qubits grow. In a second algorithm we present a general quantum circuit decomposition formalism for such problems using a method called the Quantum Shannon Decomposition. This algorithm is more robust and is exact for any number of qubits, at the cost of increased circuit complexity. The resultant circuit is implemented on IBM's quantum simulator (QASM) for 3-7 qubits. In both cases the wavepacket dynamics is found to be in good agreement with the classical result and the corresponding vibrational frequencies obtained from the wavepacket density time-evolution, are in agreement to within a few tenths of a wavenumbers.

Pb on Si(111)-(7x7) shows surprising nucleation and mass transport dynamics at odds with standard theories. To create local imbalances on stable Pb islands we use the tip of a scanning force microscope. We enforce a short, local contact between the island and our tip. The subsequent island height growth and the local contact potential difference are studied via scanning force microscopy and Kelvin probe force microscopy. Though the island has a large volume increase after the contact, we observe that its surrounding wetting layer shows the same Pb density decrease as the global wetting layer. This indicates a collective density thinning of the wetting layer., Comment: 9 pages, 2 figures in main text and 5 figures in the Appendix; Paper accepted by Physical Review B: https://journals.aps.org/prb/accepted/5e078Yf5E3015a8088ba499754fdeefbd71da71aa

We present the first multi-band centimeter detection of POX 52, a nearby dwarf galaxy believed to habor a robust intermediate mass black hole (IMBH). We conducted the deep observations using the Australia Telescope Compact Array (ATCA), spanning frequencies from 4.5 to 10 GHz, as well as the sensitive observations from the Karl G. Jansky Very Large Array (VLA) operating in its most extended A-configuration at S band (2--4 GHz) and C band (4--8 GHz). In the ATCA observations, the source shows a compact morphology, with only one direction marginally resolved. The higher resolution of the VLA allowed us to slightly resolve the source, fitting it well with a two-dimensional Gaussian model. The detected radio emission confirms the presence of Active Galactic Nucleus (AGN) activity, indicating either a low-power jet or AGN-driven winds/outflows. Our dual-epoch observations with ATCA and VLA, together with previous non-detection flux density upper limits, reveal radio emission variability spanning two decades. In addition, we find that POX 52 aligns well with the low-mass extension of the fundamental plane for high-accretion, radio-quiet massive AGNs., Comment: 11 pages, 3 figures, 1 table. Accepted for publication in ApJ

Most atomistic machine learning (ML) models rely on a locality ansatz, and decompose the energy into a sum of short-ranged, atom-centered contributions. This leads to clear limitations when trying to describe problems that are dominated by long-range physical effects - most notably electrostatics. Many approaches have been proposed to overcome these limitations, but efforts to make them efficient and widely available are hampered by the need to incorporate an ad hoc implementation of methods to treat long-range interactions. We develop a framework aiming to bring some of the established algorithms to evaluate non-bonded interactions - including Ewald summation, classical particle-mesh Ewald (PME), and particle-particle/particle-mesh (P3M) Ewald - into atomistic ML. We provide a reference implementation for PyTorch as well as an experimental one for JAX. Beyond Coulomb and more general long-range potentials, we introduce purified descriptors which disregard the immediate neighborhood of each atom, and are more suitable for general long-ranged ML applications. Our implementations are fast, feature-rich, and modular: They provide an accurate evaluation of physical long-range forces that can be used in the construction of (semi)empirical baseline potentials; they exploit the availability of automatic differentiation to seamlessly combine long-range models with conventional, local ML schemes; and they are sufficiently flexible to implement more complex architectures that use physical interactions as building blocks. We benchmark and demonstrate our torch-pme and jax-pme libraries to perform molecular dynamics simulations, to train range-separated ML potentials, and to evaluate long-range equivariant descriptors of atomic structures.

In this paper, we introduce the first diffusion model designed to generate complete synthetic human genotypes, which, by standard protocols, one can straightforwardly expand into full-length, DNA-level genomes. The synthetic genotypes mimic real human genotypes without just reproducing known genotypes, in terms of approved metrics. When training biomedically relevant classifiers with synthetic genotypes, accuracy is near-identical to the accuracy achieved when training classifiers with real data. We further demonstrate that augmenting small amounts of real with synthetically generated genotypes drastically improves performance rates. This addresses a significant challenge in translational human genetics: real human genotypes, although emerging in large volumes from genome wide association studies, are sensitive private data, which limits their public availability. Therefore, the integration of additional, insensitive data when striving for rapid sharing of biomedical knowledge of public interest appears imperative.

We classify the shift operators for the symmetric Askey-Wilson polynomials and construct shift operators for the non-symmetric Askey-Wilson polynomials using two decompositions of non-symmetric Askey-Wilson polynomials in terms of symmetric ones. These shift operators are difference-reflection operators, and we discuss the conditions under which they restrict to shift operators for the symmetric Askey-Wilson polynomials. We use them to compute the norms of the non-symmetric Askey-Wilson polynomials and compute their specialisations for $q\to1$. These turn out to be shift operators for the non-symmetric Heckman-Opdam polynomials of type $BC_1$ that have recently been found., Comment: 57 pages

I present a novel formulation of substitution, where facts about substitution that previously required tens or hundreds of lines to justify in a proof assistant now follow immediately - they can be justified by writing the four letters "refl". The paper is an executable literate Agda script, and source of the paper is available as an artifact in the file Weaken.lagda.md. Not all consequences of the pandemic have been awful. For the last three years, I've had the great pleasure of meeting with Peter Thiemann and Jeremy Siek for a couple of hours every week, via Zoom, exploring topics including core calculi, gradual typing, and formalisation in Agda. The work reported here arose from those discussions, and is dedicated to Peter on the occasion of his 60th birthday., Comment: In Proceedings PT 2024, arXiv:2412.01856

Diffusion models have achieved impressive advancements in various vision tasks. However, these gains often rely on increasing model size, which escalates computational complexity and memory demands, complicating deployment, raising inference costs, and causing environmental impact. While some studies have explored pruning techniques to improve the memory efficiency of diffusion models, most existing methods require extensive retraining to retain the model performance. Retraining a modern large diffusion model is extremely costly and resource-intensive, which limits the practicality of these methods. In this work, we achieve low-cost diffusion pruning without retraining by proposing a model-agnostic structural pruning framework for diffusion models that learns a differentiable mask to sparsify the model. To ensure effective pruning that preserves the quality of the final denoised latent, we design a novel end-to-end pruning objective that spans the entire diffusion process. As end-to-end pruning is memory-intensive, we further propose time step gradient checkpointing, a technique that significantly reduces memory usage during optimization, enabling end-to-end pruning within a limited memory budget. Results on state-of-the-art U-Net diffusion models SDXL and diffusion transformers (FLUX) demonstrate that our method can effectively prune up to 20% parameters with minimal perceptible performance degradation, and notably, without the need for model retraining. We also showcase that our method can still prune on top of time step distilled diffusion models., Comment: Project page: https://yangzhang-v5.github.io/EcoDiff

The two-qubit controlled-not (C-NOT) gate is an essential component in the construction of a gate-based quantum computer. In fact, its operation, combined with single qubit rotations allows to realise any quantum circuit. Several strategies have been adopted in order to build quantum gates, among them the photonic one offers the dual advantage of excellent isolation from the external environment and ease of manipulation at the single qubit level. Here we adopt a scalable time-multiplexed approach in order to build a fully reconfigurable architecture capable of implementing a post-selected interferometric scheme that implements the C-NOT operation with a fidelity of $(93.8\pm1.4)\%$. We use our experimental platform to generate the four Bell states., Comment: 9 pages, 4 figures

The volume of data generated and stored in contemporary global data centers is experiencing exponential growth. This rapid data growth necessitates efficient processing and analysis to extract valuable business insights. In distributed data processing systems, data undergoes exchanges between the compute servers that contribute significantly to the total data processing duration in adequately large clusters, necessitating efficient data transport protocols. Traditionally, data transport frameworks such as JDBC and ODBC have used TCP/IP-over-Ethernet as their underlying network protocol. Such frameworks require serializing the data into a single contiguous buffer before handing it off to the network card, primarily due to the requirement of contiguous data in TCP/IP. In OLAP use cases, this serialization process is costly for columnar data batches as it involves numerous memory copies that hurt data transport duration and overall data processing performance. We study the serialization overhead in the context of a widely-used columnar data format, Apache Arrow, and propose leveraging RDMA to transport Arrow data over Infiniband in a zero-copy manner. We design and implement Thallus, an RDMA-based columnar data transport protocol for Apache Arrow based on the Thallium framework from the Mochi ecosystem, compare it with a purely Thallium RPC-based implementation, and show substantial performance improvements can be achieved by using RDMA for columnar data transport.

Inverse Reinforcement Learning (IRL) has demonstrated effectiveness in a variety of imitation tasks. In this paper, we introduce an IRL framework designed to extract rewarding features from expert trajectories affected by delayed disturbances. Instead of relying on direct observations, our approach employs an efficient off-policy adversarial training framework to derive expert features and recover optimal policies from augmented delayed observations. Empirical evaluations in the MuJoCo environment under diverse delay settings validate the effectiveness of our method. Furthermore, we provide a theoretical analysis showing that recovering expert policies from augmented delayed observations outperforms using direct delayed observations.

Code summarization facilitates program comprehension and software maintenance by converting code snippets into natural-language descriptions. Over the years, numerous methods have been developed for this task, but a key challenge remains: effectively evaluating the quality of generated summaries. While human evaluation is effective for assessing code summary quality, it is labor-intensive and difficult to scale. Commonly used automatic metrics, such as BLEU, ROUGE-L, METEOR, and BERTScore, often fail to align closely with human judgments. In this paper, we explore the potential of Large Language Models (LLMs) for evaluating code summarization. We propose CODERPE (Role-Player for Code Summarization Evaluation), a novel method that leverages role-player prompting to assess the quality of generated summaries. Specifically, we prompt an LLM agent to play diverse roles, such as code reviewer, code author, code editor, and system analyst. Each role evaluates the quality of code summaries across key dimensions, including coherence, consistency, fluency, and relevance. We further explore the robustness of LLMs as evaluators by employing various prompting strategies, including chain-of-thought reasoning, in-context learning, and tailored rating form designs. The results demonstrate that LLMs serve as effective evaluators for code summarization methods. Notably, our LLM-based evaluator, CODERPE , achieves an 81.59% Spearman correlation with human evaluations, outperforming the existing BERTScore metric by 17.27%.

In invasive breast cancer, cell clusters of varying sizes and shapes are embedded in the fibrous extracellular matrix (ECM). Although the prevailing view attributes this structure to increasing disorder resulting from loss of function and dedifferentiation, our findings reveal that it arises through a process of active self-organization driven by cancer cell motility. Simulations and histological analyses of tumours from over 2,000 breast cancer patients reveal that motile, aligned cancer cells within clusters move as active nematic aggregates through the surrounding highly aligned ECM fibres, which form a confining, passive nematic phase. Cellular motion leads to cluster splitting and coalescence. The degree of cluster activity, combined with heterogeneity in cell motility, is reflected in specific scaling behaviours for cluster shape, size distribution, and the distance between cluster boundaries and nematic defects in ECM alignment. Increased activity estimates correlate with tumour progression and are associated with a poorer prognosis for patients.

Enabling effective collaboration among LLMs is a crucial step toward developing autonomous systems capable of solving complex problems. While LLMs are typically used as single-model generators, where humans critique and refine their outputs, the potential for jointly-trained collaborative models remains largely unexplored. Despite promising results in multi-agent communication and debate settings, little progress has been made in training models to work together on tasks. In this paper, we present a first step toward "Multi-agent LLM training" (MALT) on reasoning problems. Our approach employs a sequential multi-agent setup with heterogeneous LLMs assigned specialized roles: a generator, verifier, and refinement model iteratively solving problems. We propose a trajectory-expansion-based synthetic data generation process and a credit assignment strategy driven by joint outcome based rewards. This enables our post-training setup to utilize both positive and negative trajectories to autonomously improve each model's specialized capabilities as part of a joint sequential system. We evaluate our approach across MATH, GSM8k, and CQA, where MALT on Llama 3.1 8B models achieves relative improvements of 14.14%, 7.12%, and 9.40% respectively over the same baseline model. This demonstrates an early advance in multi-agent cooperative capabilities for performance on mathematical and common sense reasoning questions. More generally, our work provides a concrete direction for research around multi-agent LLM training approaches., Comment: Preliminary work

This paper presents a unifying framework for Trefftz-like methods, which allows the analysis and construction of discretization methods based on the decomposition into, and coupling of, local and global problems. We apply the framework to provide a comprehensive error analysis for the Embedded Trefftz discontinuous Galerkin method, for a wide range of second-order scalar elliptic partial differential equations and a scalar reaction-advection problem. We also analyze quasi-Trefftz methods with our framework and build bridges to other methods that are similar in virtue., Comment: 43 pages, 6 figures

Multimodal sarcasm detection (MSD) is essential for various downstream tasks. Existing MSD methods tend to rely on spurious correlations. These methods often mistakenly prioritize non-essential features yet still make correct predictions, demonstrating poor generalizability beyond training environments. Regarding this phenomenon, this paper undertakes several initiatives. Firstly, we identify two primary causes that lead to the reliance of spurious correlations. Secondly, we address these challenges by proposing a novel method that integrate Multimodal Incongruities via Contrastive Learning (MICL) for multimodal sarcasm detection. Specifically, we first leverage incongruity to drive multi-view learning from three views: token-patch, entity-object, and sentiment. Then, we introduce extensive data augmentation to mitigate the biased learning of the textual modality. Additionally, we construct a test set, SPMSD, which consists potential spurious correlations to evaluate the the model's generalizability. Experimental results demonstrate the superiority of MICL on benchmark datasets, along with the analyses showcasing MICL's advancement in mitigating the effect of spurious correlation., Comment: Accepted to COLING 2025

While Transformers have revolutionized machine learning on various data, existing Transformers for temporal graphs face limitations in (1) restricted receptive fields, (2) overhead of subgraph extraction, and (3) suboptimal generalization capability beyond link prediction. In this paper, we rethink temporal graph Transformers and propose TGTOD, a novel end-to-end Temporal Graph Transformer for Outlier Detection. TGTOD employs global attention to model both structural and temporal dependencies within temporal graphs. To tackle scalability, our approach divides large temporal graphs into spatiotemporal patches, which are then processed by a hierarchical Transformer architecture comprising Patch Transformer, Cluster Transformer, and Temporal Transformer. We evaluate TGTOD on three public datasets under two settings, comparing with a wide range of baselines. Our experimental results demonstrate the effectiveness of TGTOD, achieving AP improvement of 61% on Elliptic. Furthermore, our efficiency evaluation shows that TGTOD reduces training time by 44x compared to existing Transformers for temporal graphs. To foster reproducibility, we make our implementation publicly available at https://github.com/kayzliu/tgtod., Comment: Preprint. Under review. Code available at https://github.com/kayzliu/tgtod

Aluminum scandium nitride (AlScN) has emerged as a highly promising material for high-temperature applications due to its robust piezoelectric, ferroelectric, and dielectric properties. This study investigates the behavior of Al0.7Sc0.3N thin films in extreme thermal environments, demonstrating functional stability up to 1000{\deg}C, making it suitable for use in aerospace, hypersonics, deep-well, and nuclear reactor systems. Tantalum silicide (TaSi2)/Al0.7Sc0.3N/TaSi2 capacitors were fabricated and characterized across a wide temperature range, revealing robust ferroelectric and dielectric properties, along with significant enhancement in piezoelectric performance. At 1000{\deg}C, the ferroelectric hysteresis loops showed a substantial reduction in coercive field from 4.3 MV/cm to 1.2 MV/cm, while the longitudinal piezoelectric coefficient increased nearly tenfold, reaching 75.1 pm/V at 800{\deg}C. Structural analysis via scanning and transmission electron microscopy confirmed the integrity of the TaSi2/Al0.7Sc0.3N interfaces, even after exposure to extreme temperatures. Furthermore, the electromechanical coupling coefficient was calculated to increase by over 500%, from 12.9% at room temperature to 82% at 700{\deg}C. These findings establish AlScN as a versatile material for high-temperature ferroelectric, piezoelectric, and dielectric applications, offering unprecedented thermal stability and functional enhancement.