October 2025 arXiv papers — page 178

James W. Johnson, Miqaela K. Weller, Ryan J. Cooke

Star formation drives changes in the compositions of galaxies, fusing H and He into heavier nuclei. This paper investigates the differences in abundance evolution between metal and non-metal isotopes using recent models of Galactic chemical evolution appropriate for the thin disk epoch. A strong degeneracy arises between metal yields from stellar populations

James W. Johnson, Diane K. Feuillet, Ana Bonaca, Danielle de Brito Silva

The Milky Way (MW) is thought to have experienced a $\sim$3:1 mass-ratio merger event near redshift $z\sim2$ with a significantly retrograde trajectory. This now-disrupted dwarf galaxy is commonly known as the Gaia-Sausage-Enceladus (GSE). In this paper, we investigate the impact of the GSE merger trajectory on metal abundances in the MW disk. We construct n

Romina Ballesteros, Marcela Cárdenas, Eric Lescano

We develop a covariant differential-form framework to define scalar charges for stationary, asymptotically flat black holes in $4$--dimensional Einstein-scalar-Gauss-Bonnet gravity with a general scalar coupling function. Contracting the scalar field equation of motion with the horizon generator $k$ yields a non-closed-form scalar charge, revealing a bulk co

Juan Pablo Elía, Lucas Cantarutti, Nahuel Mirón-Granese, Esteban Calzetta

We develop a causal hydrodynamic model that provides an effective macroscopic description of the field-theoretic dynamics during the early stages of reheating. The inflaton condensate is treated as a homogeneous background coupled to a relativistic fluid that represents its inhomogeneous fluctuations. Within the divergence-type theory framework derived from

Ling Zhu, Runsheng Cai, Xi Kang, Xiang-Xiang Xue

Unveiling the 3D shape of the Milky Way's dark-matter halo is critical to understanding its formation history. We created an innovative dynamical model with minimal assumptions on the internal dynamical structures and accommodates a highly flexible triaxial DM halo. By applying the method to 6D phase-space data of K-giant stars from LAMOST + Gaia, we robustl

Maria Charisi, Stephen Taylor, Jessie Runnoe, Caitlin Witt

Supermassive black hole binaries (SMBHBs) formed in galaxy mergers are promising multi-messenger sources. They can be identified as quasars with periodic variability in electromagnetic (EM) time-domain surveys. The most massive of those systems can be detected by Pulsar Timing Arrays (PTAs) in the nanohertz frequency gravitational-wave (GW) band. We present

Sumit Banik, Andreas Crivellin, Luca Naterop, Peter Stoffer

We compute the two-loop renormalization-group equations for the baryon-number-violating dimension-six operators in the SMEFT. This includes all three gauge interactions, the Yukawa, and Higgs self-interaction contributions. In addition, we present the one-loop matching of the $S_1$ scalar leptoquark on the SMEFT, which can generate the Wilson coefficients of

Ankan Sur, Roberto Tejada Arevalo, Adam Burrows, Yi-Xian Chen

Many evolutionary models of giant exoplanets still rely on simplifying assumptions that are no longer adequate given detailed constraints from Jupiter, Saturn, and modern exoplanet observations. Here, we identify the key physical improvements required for next-generation planetary evolution models using our code, \texttt{APPLE}, which enables systematic emul

Emilio Rui, Joachim Cohen, Alexandru Petrescu

We present a nonperturbative tensor-network approach to the excitation spectra of superconducting circuits based on Josephson junction arrays. These arrays provide the large lumped inductances required for qubit designs, yet their intrinsically many-body nature is typically reduced to effective single-mode descriptions. Perturbative treatments attempt to inc

T. G. Blackburn, B. King, M. Samuelsson

Several high power laser facilities are reaching field strengths where leading order strong-field quantum electrodynamical (QED) processes can be measured in the non-perturbative regime for the first time. At very high, as yet unobtainable in the laboratory, field strengths, the contribution of higher-order processes is predicted to dominate, implying a brea

Riek H. Rüstemeier, H. P. Ojeda Collado, Ludwig Mathey

The field of coherent electronics aims to advance electronic functionalities by utilizing quantum coherence. Here, we demonstrate a viable and versatile methodology for controlling electron dynamics optically in graphene nanoribbons. In particular, we propose to flatten the band structure of armchair graphene nanoribbons via control electrodes, arranged peri

Mattia Di Mauro, Bohan Xie

The particle-physics nature of dark matter (DM) remains one of the central open questions in modern physics. A widely used framework to investigate DM properties is provided by simplified models (DMSimps), which extend the Standard Model with a DM particle and a mediator that connects the visible and dark sectors. Much of the DMSimps parameter space is alrea

Érika Le Bourdais, Patrick Dufour, Carl Melis, Beth L. Klein

We report the detection of 13 heavy elements (Na, Mg, Al, Si, Ca, Ti, Cr, Mn, Fe, Co, Ni, Cu, and Sr) in the photosphere of LSPM J0207+3331, a ~3 Gyr old hydrogen-rich white dwarf with an effective temperature comparable to that of the Sun. Upper limits on carbon, obtained through the absence of molecular CH, suggest accretion from a carbon-volatile-depleted

Admir Greljo, Ajdin Palavrić, Mirsad Tunja, Jure Zupan

We chart new-physics models that produce exotic, high-multiplicity muon decays featuring prompt or displaced $e^+e^-$ pairs and/or photons, with or without missing energy, such as $\mu \to 5e$, $\mu \to 7e$, etc. Starting from an effective-field-theory perspective, we estimate the reach on the ultraviolet scale and identify conditions under which lower-multi

Haofei Xu, Daniel Barath, Andreas Geiger, Marc Pollefeys

While existing feed-forward Gaussian splatting models offer computational efficiency and can generalize to sparse view settings, their performance is fundamentally constrained by relying on a single forward pass for inference. We propose ReSplat, a feed-forward recurrent Gaussian splatting model that iteratively refines 3D Gaussians without explicitly comput

Thomas Pfeil, Philip J. Armitage, Yan-Fei Jiang

Super-thermal gas giant planets or their progenitor cores are known to open deep gaps in protoplanetary disks, which stop large, drifting dust particles on their way to the inner disk. The possible separation of the disk into distinct reservoirs and the resulting dust depletion interior to the gap have important implications for planetesimal formation and th

Baptiste Barthe-Gold, Nhat-Minh Nguyen, Leander Thiele

We construct a neural network to perform regression on the local dark-matter density field given line-of-sight peculiar velocities of dark-matter halos, biased tracers of the dark matter field. Our architecture combines a convolutional U-Net with a point-cloud DeepSets. This combination enables efficient use of small-scale information and improves reconstruc

Rocktim Jyoti Das, Harsh Singh, Diana Turmakhan, Muhammad Abdullah Sohail

Scaling data and models has played a pivotal role in the remarkable progress of computer vision and language. Inspired by these domains, recent efforts in robotics have similarly focused on scaling both data and model size to develop more generalizable and robust policies. However, unlike vision and language, robotics lacks access to internet-scale demonstra

Animikh Aich, Adwait Kulkarni, Eshed Ohn-Bar

Real-world evaluation of perception-based planning models for robotic systems, such as autonomous vehicles, can be safely and inexpensively conducted offline, i.e. by computing model prediction error over a pre-collected validation dataset with ground-truth annotations. However, extrapolating from offline model performance to online settings remains a challe

Nimrod Berman, Assaf Hallak, Assaf Shocher

Neural networks are famously nonlinear. However, linearity is defined relative to a pair of vector spaces, $f:X \to Y$. Leveraging the algebraic concept of transport of structure, we propose a method to explicitly identify non-standard vector spaces where a neural network acts as a linear operator. When sandwiching a linear operator $A$ between two invertibl

Qin Liu, Jacob Dineen, Yuxi Huang, Sheng Zhang

Benchmarks are central to measuring the capabilities of large language models and guiding model development, yet widespread data leakage from pretraining corpora undermines their validity. Models can match memorized content rather than demonstrate true generalization, which inflates scores, distorts cross-model comparisons, and misrepresents progress. We int

Hongyu Li, Lingfeng Sun, Yafei Hu, Duy Ta

Enabling robots to execute novel manipulation tasks zero-shot is a central goal in robotics. Most existing methods assume in-distribution tasks or rely on fine-tuning with embodiment-matched data, limiting transfer across platforms. We present NovaFlow, an autonomous manipulation framework that converts a task description into an actionable plan for a target

Tajamul Ashraf, Umair Nawaz, Abdelrahman M. Shaker, Rao Anwer

Vision language models (VLMs) are increasingly deployed as controllers with access to external tools for complex reasoning and decision-making, yet their effectiveness remains limited by the scarcity of high-quality multimodal trajectories and the cost of manual annotation. We address this challenge with a vision-centric agent tuning framework that automatic

Meixi Song, Xin Lin, Dizhe Zhang, Haodong Li

Recent advances in 3D Gaussian Splatting (3DGS) enable real-time, high-fidelity novel view synthesis (NVS) with explicit 3D representations. However, performance degradation and instability remain significant under sparse-view conditions. In this work, we identify two key failure modes under sparse-view conditions: overfitting in regions with excessive Gauss

Changyao Tian, Hao Li, Gen Luo, Xizhou Zhu

Compositional training has been the de-facto paradigm in existing Multimodal Large Language Models (MLLMs), where pre-trained vision encoders are connected with pre-trained LLMs through continuous multimodal pre-training. However, the multimodal scaling property of this paradigm remains difficult to explore due to the separated training. In this paper, we fo

Zhen Zhu, Yiming Gong, Yao Xiao, Yaoyao Liu

How can we teach large multimodal models (LMMs) new skills without erasing prior abilities? We study sequential fine-tuning on five target skills while monitoring general ability on eight held-out benchmarks across three model families. Surprisingly, we find that performance lost on held-out tasks after fine-tuning on one skill can partly recover when the mo

El Houcine Bergou, Soumia Boucherouite, Aritra Dutta, Xin Li

The randomized Kaczmarz (RK) algorithm is one of the most computationally and memory-efficient iterative algorithms for solving large-scale linear systems. However, practical applications often involve noisy and potentially inconsistent systems. While the convergence of RK is well understood for consistent systems, the study of RK on noisy, inconsistent line

Zhiyu Zheng, Shaoyu Chen, Haoran Yin, Xinbang Zhang

End-to-end autonomous driving (E2EAD) systems, which learn to predict future trajectories directly from sensor data, are fundamentally challenged by the inherent spatio-temporal imbalance of trajectory data. This imbalance creates a significant optimization burden, causing models to learn spurious correlations instead of robust driving logic, while also prio

Kang Liao, Size Wu, Zhonghua Wu, Linyi Jin

Camera-centric understanding and generation are two cornerstones of spatial intelligence, yet they are typically studied in isolation. We present Puffin, a unified camera-centric multimodal model that extends spatial awareness along the camera dimension. Puffin integrates language regression and diffusion-based generation to interpret and create scenes from

Maham Tanveer, Yang Zhou, Simon Niklaus, Ali Mahdavi Amiri

Video inbetweening creates smooth and natural transitions between two image frames, making it an indispensable tool for video editing and long-form video synthesis. Existing works in this domain are unable to generate large, complex, or intricate motions. In particular, they cannot accommodate the versatility of user intents and generally lack fine control o

Andong Deng, Taojiannan Yang, Shoubin Yu, Lincoln Spencer

Large Multimodal Models (LMMs) have achieved remarkable progress across various capabilities; however, complex video reasoning in the scientific domain remains a significant and challenging frontier. Current video benchmarks predominantly target general scenarios where perception/recognition is heavily relied on, while with relatively simple reasoning tasks,

Jaroslaw Pawlowski, Dickson Thian, Repaka Maheswar, Chai Jian Wei

The dielectric environment surrounding semiconductors plays a crucial role in determining device performance, a role that becomes especially pronounced in atomically thin semiconductors where charge carriers are confined within a few atomic layers and strongly interact with their surroundings. High-k dielectrics, such as hafnium oxide (HfO2), have been shown

Xueyi Liu, He Wang, Li Yi

Achieving generalized in-hand object rotation remains a significant challenge in robotics, largely due to the difficulty of transferring policies from simulation to the real world. The complex, contact-rich dynamics of dexterous manipulation create a "reality gap" that has limited prior work to constrained scenarios involving simple geometries, limited objec

Kevin Rojas, Jiahe Lin, Kashif Rasul, Anderson Schneider

Diffusion language models (DLMs) enable parallel, order-agnostic generation with iterative refinement, offering a flexible alternative to autoregressive large language models (LLMs). However, adapting reinforcement learning (RL) fine-tuning to DLMs remains an open challenge because of the intractable likelihood. Pioneering work such as diffu-GRPO estimated t

Yunzhe Xu, Yiyuan Pan, Zhe Liu

Vision-and-Language Navigation (VLN) requires agents to follow natural language instructions through environments, with memory-persistent variants demanding progressive improvement through accumulated experience. Existing approaches for memory-persistent VLN face critical limitations: they lack effective memory access mechanisms, instead relying on entire me

Shi Jie Samuel Tan, Yifan Hong, Ting-Chun Lin, Michael J. Gullans

Code-switching is a powerful technique in quantum error correction that allows one to leverage the complementary strengths of different codes to achieve fault-tolerant universal quantum computation. However, existing code-switching protocols that encapsulate recent generalized lattice surgery approaches often either require many rounds of measurements to ens

Guanghao Li, Kerui Ren, Linning Xu, Zhewen Zheng

On-the-fly 3D reconstruction from monocular image sequences is a long-standing challenge in computer vision, critical for applications such as real-to-sim, AR/VR, and robotics. Existing methods face a major tradeoff: per-scene optimization yields high fidelity but is computationally expensive, whereas feed-forward foundation models enable real-time inference

Tolga Altinoluk, Guillaume Beuf, Jani Penttala

We propose three possible rapidity regulators for higher-order calculations in low $x$ QCD with gluon saturation, as alternatives to the usual lower cut-off for the integrals over the light-cone momentum $k^+$. These rapidity regulators are closely related to the $\eta$ regulator and to the pure rapidity regulator, which have been used primarily in studies o

Zilin Kang, Chonghua Liao, Tingqiang Xu, Huazhe Xu

We propose ERA, a new paradigm that constrains the sampling entropy above given thresholds by applying specially designed activations to the outputs of models. Our approach demonstrates broad effectiveness across different domains: 1) for large language models(LLMs), boosting the AIME 2025 score for Qwen2.5-Math-7B by 37.4%; 2) for continuous control reinfor

Bo Yang, Elham Kashefi, Harold Ollivier

Cryptographic verification is essential for establishing trust in quantum-computing-as-a-service. However, a fundamental gap exists in the current verification landscape: existing efficient protocols are largely restricted to decision problems where correctness is boosted by classical majority voting. This excludes observable estimation, the statistical task

Xiuwei Xu, Angyuan Ma, Hankun Li, Bingyao Yu

Towards the aim of generalized robotic manipulation, spatial generalization is the most fundamental capability that requires the policy to work robustly under different spatial distribution of objects, environment and agent itself. To achieve this, substantial human demonstrations need to be collected to cover different spatial configurations for training a

Hengrui Zhang, Pratyush Patel, August Ning, David Wentzlaff

Large Language Models (LLMs) have gained popularity in recent years, driving up the demand for inference. LLM inference is composed of two phases with distinct characteristics: a compute-bound prefill phase followed by a memory-bound decode phase. To efficiently serve LLMs, prior work proposes prefill-decode disaggregation to run each phase on separate hardw

Ainesh Bakshi, Allen Liu, Ankur Moitra, Ewin Tang

A central challenge in quantum physics is to understand the structural properties of many-body systems, both in equilibrium and out of equilibrium. For classical systems, we have a unified perspective which connects structural properties of systems at thermal equilibrium to the Markov chain dynamics that mix to them. We lack such a perspective for quantum sy

Debsurya De, Dmitriy Kunisky

Recent work has generalized several results concerning the well-understood spiked Wigner matrix model of a low-rank signal matrix corrupted by additive i.i.d. Gaussian noise to the inhomogeneous case, where the noise has a variance profile. In particular, for the special case where the variance profile has a block structure, a series of results identified an

Xiangyu Zhao, Junming Lin, Tianhao Liang, Yifan Zhou

While current Multimodal Large Language Models (MLLMs) have demonstrated proficiency in reasoning tasks such as mathematics and logic, their capacity for long-chain reflective reasoning, a prerequisite for solving complex real-world problems, remains largely underexplored. In this work, we first conduct an extensive empirical investigation to evaluate this c

Joe Suk, Yaqi Duan

Reinforcement Learning with Verifiable Rewards (RLVR), which uses simple binary feedback to post-train large language models, has found significant empirical success. However, a principled understanding of why it works is lacking. This paper builds a theoretical foundation for RLVR by analyzing its training process at both the full-response (trajectory) and

Thiago Bergamaschi, Chi-Fang Chen, Umesh Vazirani

Statistical mechanics assumes that a quantum many-body system at low temperature can be effectively described by its Gibbs state. However, many complex quantum systems exist only as metastable states of dissipative open system dynamics, which appear stable and robust yet deviate substantially from true thermal equilibrium. In this work, we model metastable s

Nicholas Laracuente

Approximate unitary k-designs are ensembles of unitary matrices whose first k statistical moments approximate the Haar (uniform) measure. Shallow, random quantum circuits with temporally and spatially structured architectures are known to yield approximate k-designs in depth that is logarithmic in the number of qubits. It was open whether this structure is p

Gregor Olenik, Marcel Koch, Hartwig Anzt

Modern high-performance computing (HPC) increasingly relies on GPUs, but integrating GPU acceleration into complex scientific frameworks like OpenFOAM remains a challenge. Existing approaches either fully refactor the codebase or use plugin-based GPU solvers, each facing trade-offs between performance and development effort. In this work, we address the limi

Arianna Rizzieri, Clément Leloup, Josquin Errard, Davide Poletti

In the context of maximum-likelihood parametric component separation for next-generation full-sky CMB polarization experiments, we study the impact of fitting different spectral parameters of Galactic foregrounds in distinct subsets of pixels on the sky, with the goal of optimizing the search for primordial B modes. Using both simulations and analytical argu

Thiago Bergamaschi, Chi-Fang Chen

It is shown that every one-dimensional Hamiltonian with short-range interaction admits a quantum Gibbs sampler [CKG23] with a system-size independent spectral gap at all finite temperatures. Consequently, their Gibbs states can be prepared in polylogarithmic depth, and satisfy exponential clustering of correlations, generalizing [Ara69].

Hongxing Li, Dingming Li, Zixuan Wang, Yuchen Yan

Spatial reasoning remains a fundamental challenge for Vision-Language Models (VLMs), with current approaches struggling to achieve robust performance despite recent advances. We identify that this limitation stems from a critical gap: existing methods attempt to learn spatial reasoning directly without establishing the hierarchical foundations of perception

Zhitong Huang, Mohan Zhang, Renhan Wang, Rui Tang

We present X2Video, the first diffusion model for rendering photorealistic videos guided by intrinsic channels including albedo, normal, roughness, metallicity, and irradiance, while supporting intuitive multi-modal controls with reference images and text prompts for both global and local regions. The intrinsic guidance allows accurate manipulation of color,

Xiangyuan Xue, Yifan Zhou, Guibin Zhang, Zaibin Zhang

Self-evolution is a central research topic in enabling large language model (LLM)-based agents to continually improve their capabilities after pretraining. Recent research has witnessed a transition from reinforcement learning (RL)-free to RL-based methods. Current RL-based methods either rely on dense external reward signals or extract intrinsic reward sign

Thi Ha Kyaw, Guillermo Romero, Gaurav Saxena

A major drawback of adiabatic quantum computing (AQC) is fulfilling the energy gap constraint, which requires the total evolution time to scale inversely with the square of the minimum energy gap. Failure to satisfy this condition violates the adiabatic approximation, potentially undermining computational accuracy. Recently, several approaches have been prop

Zhiyuan Zhang, Can Wang, Dongdong Chen, Jing Liao

We present FlexTraj, a framework for image-to-video generation with flexible point trajectory control. FlexTraj introduces a unified point-based motion representation that encodes each point with a segmentation ID, a temporally consistent trajectory ID, and an optional color channel for appearance cues, enabling both dense and sparse trajectory control. Inst

Yash Jhaveri, Harley Wiltzer, Patrick Shafto, Marc G. Bellemare

In the pursuit of finding an optimal policy, reinforcement learning (RL) methods generally ignore the properties of learned policies apart from their expected return. Thus, even when successful, it is difficult to characterize which policies will be learned and what they will do. In this work, we present a theoretical framework for policy optimization that g

Hyunji Lee, Kevin Chenhao Li, Matthias Grabmair, Shanshan Xu

Prompt optimization aims to systematically refine prompts to enhance a language model's performance on specific tasks. Fairness detection in Terms of Service (ToS) clauses is a challenging legal NLP task that demands carefully crafted prompts to ensure reliable results. However, existing prompt optimization methods are often computationally expensive due to

Guo Zheng, Liang Jiang, Qian Xu

Scalable quantum computation requires not only quantum codes with low memory overhead but also encoded operations with low space-time overhead. High rate quantum low-density parity-check (qLDPC) codes address the former by achieving a high information-encoding rate, yet existing methods for implementing logical operations often suffer from a low information-

Shanshan Xu, Santosh T. Y. S. S, Barbara Plank

Human Label Variation (HLV) refers to legitimate disagreement in annotation that reflects the diversity of human perspectives rather than mere error. Long treated in NLP as noise to be eliminated, HLV has only recently been reframed as a signal for improving model robustness. With the rise of large language models (LLMs) and post-training methods such as hum

Yuanjun Dai, Keqiang He, An Wang

Existing batch size selection approaches in distributed machine learning rely on static allocation or simplistic heuristics that fail to adapt to heterogeneous, dynamic computing environments. We present DYNAMIX, a reinforcement learning framework that formulates batch size optimization as a sequential decision-making problem using Proximal Policy Optimizati

Yusong Hu, Runmin Ma, Yue Fan, Jinxin Shi

Deep research is an inherently challenging task that demands both breadth and depth of thinking. It involves navigating diverse knowledge spaces and reasoning over complex, multi-step dependencies, which presents substantial challenges for agentic systems. To address this, we propose FlowSearch, a multi-agent framework that actively constructs and evolves a

Sriram Akella

Holographic states satisfy several entropic inequalities owing to the Ryu-Takayangi formula. A drawback of these inequalities is that they only use bipartite entanglement in their formulation. We investigate a recently proposed "GHZ-forbidding" inequality, built out of the reflected entropy and the tripartite multi-entropy, that holds for holographic states.

Jake Belton, Ziwen Kong

In conformal field theory, the presence of a defect may break the global symmetry, giving rise to defect operators such as the tilts. In this work, we derive integral identities that relate correlation functions involving bulk and defect operators -- including tilts -- to lower-point bulk-defect correlators, based on a detailed analysis of the Lie algebra of

Aram W. Harrow, Angus Lowe, Freek Witteveen

A fundamental task in quantum information is to approximate a pure quantum state in terms of sparse states or, for a bipartite system, states of bounded Schmidt rank. The optimal deterministic approximation in each case is straightforward, and maximizes the fidelity: keep the largest entries or singular values. On the other hand, random mixtures of sparse st

Grace Liu, Yuxiao Qu, Jeff Schneider, Aarti Singh

Many tasks require learned models to strategically gather relevant information over multiple rounds of interaction before actually acting on a task. Strategic information gathering requires models to know not only how to effectively acquire information, but also when to stop gathering information and make a decision, in order to avoid overthinking or getting

Sajid Ullah

In this paper, by means of Birkhoff--Kellogg type Theorem in cones we address the existence of eigenvalues and the corresponding eigenvectors to a family of coupled system of thermostat type. The system is characterized by the presence of a real parameter that influences not only the differential equations but also the boundary conditions. Motivated by model

Mu-Chun Chen, Yan Cao

The final-parsec problem has long posed a central challenge in understanding the merger of supermassive black hole binaries. In this paper, we investigate a scenario in which a dark scalar or vector field is sourced by eccentric binaries, leading to accelerated mergers through additional dipole radiation, and thereby extending the range of masses for which t

Md Kowsher, Ali O. Polat, Ehsan Mohammady Ardehaly, Mehrdad Salehi

This paper presents a theoretical framework explaining why fine tuning small, randomly selected subnetworks (slices) within pre trained models can be sufficient for downstream adaptation. We prove that pretrained networks exhibit a universal winning slice property arising from two phenomena: (1) spectral balance the eigenspectra of different weight matrix sl

Nikolaos Stathoulopoulos, Christoforos Kanellakis, George Nikolakopoulos

Efficient transmission of 3D point cloud data is critical for advanced perception in centralized and decentralized multi-agent robotic systems, especially nowadays with the growing reliance on edge and cloud-based processing. However, the large and complex nature of point clouds creates challenges under bandwidth constraints and intermittent connectivity, of

Shangheng Du, Xiangchao Yan, Dengyang Jiang, Jiakang Yuan

Large language models (LLMs) have shown impressive performance in general programming tasks. However, in Machine Learning Engineering (MLE) scenarios such as AutoML and Kaggle competitions, achieving high performance depends heavily on expert intervention and repeated adjustments rather than simply generating correct code. When applied directly to these task

Jiayun Luo, Wan-Cyuan Fan, Lyuyang Wang, Xiangteng He

Large Vision Language Models (LVLMs) have recently emerged as powerful architectures capable of understanding and reasoning over both visual and textual information. These models typically rely on two key components: a Vision Transformer (ViT) and a Large Language Model (LLM). ViT encodes visual content into a sequence of image tokens and serves as the perce

Simon Apers, Arjan Cornelissen, Samson Wang

The complexity of matrix multiplication is a central topic in computer science. While the focus has traditionally been on exact algorithms, a long line of literature also considers randomized algorithms, which return an approximate solution in faster time. In this work, we adopt a unifying perspective that frames these randomized algorithms in terms of mean

Lu Liu, Chunlei Cai, Shaocheng Shen, Jianfeng Liang

Real-world videos often suffer from complex degradations, such as noise, compression artifacts, and low-light distortions, due to diverse acquisition and transmission conditions. Existing restoration methods typically require professional manual selection of specialized models or rely on monolithic architectures that fail to generalize across varying degrada

Xuanqiang Zhao, Benchi Zhao, Cyril Branciard, Giulio Chiribella

Quantum theory is in principle compatible with scenarios where physical processes occur in an indefinite order, potentially yielding advantages in a broad range of information processing tasks. However, advantages in communication, the most basic form of information processing, have so far remained controversial and hard to prove. Here we provide a framework

John Hewitt, Oyvind Tafjord, Robert Geirhos, Been Kim

Humans invent new words when there is a rising demand for a new useful concept (e.g., doomscrolling). We explore and validate a similar idea in our communication with LLMs: introducing new words to better understand and control the models, expanding on the recently introduced neologism learning. This method introduces a new word by adding a new word embeddin

Anne Inkenhaag, Andrew J. Levan, Andrew Mummery, Peter G. Jonker

The Luminous Fast Blue Optical Transient (LFBOT) AT 2018cow is the prototype of its class with an extensive set of multi-wavelength observations. Despite a rich data set there is, still, no consensus about the physical nature and origin of this event. AT 2018cow remained UV bright 2-4 years after the explosion, which points at an additional energy injection

Alex Meiburg, Leonardo A. Lessa, Rodolfo R. Soldati

The Generalized Quantum Stein's Lemma is a theorem in quantum hypothesis testing that provides an operational meaning to the relative entropy within the context of quantum resource theories. Its original proof was found to have a gap, which led to a search for a corrected proof. We formalize the proof presented in [Hayashi and Yamasaki (2024)] in the Lean in

Bayu Dirgantara, J. Julio

We examine electron and muon anomalous magnetic dipole moments within a radiative neutrino mass model featuring TeV-scale scalar leptoquarks $S(3,1,-1/3)$ and $R(3,2,1/6)$. We utilize textures with decoupling electron and muon sectors, so that both electron and muon anomalous magnetic dipole moments could receive internal chiral enhancements from different h

Thomas Schuster, Dominik Kufel, Norman Y. Yao, Hsin-Yuan Huang

We prove that recognizing the phase of matter of an unknown quantum state is quantum computationally hard. More specifically, we show that the quantum computational time of any phase recognition algorithm must grow exponentially in the range of correlations $\xi$ of the unknown state. This exponential growth renders the problem practically infeasible for eve

Denis S. Goldobin, Lyudmila S. Klimenko, Irina V. Tyulkina, Vasily A. Kostin

We study single-variable approaches for describing stochastic dynamics with small inertia. The basic models we deal with describe passive Brownian particles and phase elements (phase oscillators, rotators, superconducting Josephson junctions) with an effective inertia in the case of a linear dissipation term and active Brownian particles in the case of a non

Christopher Vairogs, Akanksha Chablani, Leo Lee, Hanyang Sha

In this work, we study the asymptotic behavior of protocols that localize entanglement in large multi-qubit states onto a subset of qubits by measuring the remaining qubits. We use the maximal average n-tangle that can be generated on a fixed subsystem by measuring its complement -- either with local or global measurements -- as our key figure of merit. Thes

Daniel Stilck França, Tim Möbus, Cambyse Rouzé, Albert H. Werner

Hamiltonian learning protocols are essential tools to benchmark quantum computers and simulators. Yet rigorous methods for time-dependent Hamiltonians and Lindbladians remain scarce despite their wide use. We close this gap by learning the time-dependent evolution of a locally interacting $n$-qubit system on a graph of effective dimension $D$ using only prep

Sitan Chen, Jordan Cotler, Hsin-Yuan Huang

Characterizing quantum many-body systems is a fundamental problem across physics, chemistry, and materials science. While significant progress has been made, many existing Hamiltonian learning protocols demand digital quantum control over the entire system, creating a disconnect from many real-world settings that provide access only through small, local prob

Riadh Bouslimi, Houda Trabelsi, Wahiba Ben Abdssalem Karaa, Hana Hedhli

Traumatic brain injuries present significant diagnostic challenges in emergency medicine, where the timely interpretation of medical images is crucial for patient outcomes. In this paper, we propose a novel AI-based approach for automatic radiology report generation tailored to cranial trauma cases. Our model integrates an AC-BiFPN with a Transformer archite

Alexander Zlokapa, Bobak T. Kiani, Eric R. Anschuetz

Glassiness -- a phenomenon in physics characterized by a rough free-energy landscape -- implies hardness for stable classical algorithms. For example, it can obstruct constant-time Langevin dynamics and message-passing in random $k$-SAT and max-cut instances. We provide an analogous framework for average-case quantum complexity showing that a natural family

Akid Abrar, Sagar Dasgupta, Mizanur Rahman, Ahmad Alsharif

Transportation Cyber-Physical Systems (TCPS) integrate physical elements, such as transportation infrastructure and vehicles, with cyber elements via advanced communication technologies, allowing them to interact seamlessly. This integration enhances the efficiency, safety, and sustainability of transportation systems. TCPS rely heavily on cryptographic secu

Andrew Huang, Yael Tauman Kalai

We present a generic compiler that converts any $\mathsf{MIP}^{*}$ protocol into a succinct interactive argument where the communication and the verifier are classical, and where post-quantum soundness relies on the post-quantum sub-exponential hardness of the Learning with Errors ($\mathsf{LWE}$) problem. Prior to this work, such a compiler for $\mathsf{MIP

Alexander Schmidhuber, Alexander Zlokapa

Community detection is a foundational problem in data science. Its natural extension to hypergraphs captures higher-order correlations beyond pairwise interactions. In this work, we develop a quantum algorithm for hypergraph community detection that achieves a quartic quantum speedup over the best known classical algorithm, along with superpolynomial savings

Sujay Kazi, Iman Marvian

We study coherence distillation under time-translation-invariant operations: given many copies of a quantum state containing coherence in the energy eigenbasis, the aim is to produce a purer coherent state while respecting the time-translation symmetry. This symmetry ensures that the output remains synchronized with the input and that the process can be real

Sharut Gupta, Shobhita Sundaram, Chenyu Wang, Stefanie Jegelka

Traditional multimodal learners find unified representations for tasks like visual question answering, but rely heavily on paired datasets. However, an overlooked yet potentially powerful question is: can one leverage auxiliary unpaired multimodal data to directly enhance representation learning in a target modality? We introduce UML: Unpaired Multimodal Lea

Milon Bhattacharya, Milan Kumar

Indias e-commerce market is projected to grow rapidly, with last-mile delivery accounting for nearly half of operational expenses. Although vehicle routing problem (VRP) based solvers are widely used for delivery planning, their effectiveness in real-world scenarios is limited due to unstructured addresses, incomplete maps, and computational constraints in d

Xilong Zhou, Bao-Huy Nguyen, Loïc Magne, Vladislav Golyanik

Radiance fields have emerged as a predominant representation for modeling 3D scene appearance. Neural formulations such as Neural Radiance Fields provide high expressivity but require costly ray marching for rendering, whereas primitive-based methods such as 3D Gaussian Splatting offer real-time efficiency through splatting, yet at the expense of representat

Zijian Wang, Ruihua Fan, Tianle Wang, Samuel J. Garratt

Fractional quantum Hall states are promising platforms for topological quantum computation due to their capacity to encode quantum information in topologically degenerate ground states and in the fusion space of non-abelian anyons. We investigate how the information encoded in two paradigmatic states, the Laughlin and Moore-Read states, is affected by densit

Immanuel Trummer

Semantic query processing engines often support semantic joins, enabling users to match rows that satisfy conditions specified in natural language. Such join conditions can be evaluated using large language models (LLMs) that solve novel tasks without task-specific training. Currently, many semantic query processing engines implement semantic joins via neste

Wencong Yang, Haoyu Ji, Leo Lonzarich, Yalan Song

Hourly predictions are critical for issuing flood warnings as the flood peaks on the hourly scale can be distinctly higher than the corresponding daily ones. Currently a popular hourly data-driven prediction scheme is multi-time-scale long short-term memory (MTS-LSTM), yet such models face challenges in probabilistic forecasts or integrating observations whe

Mohammadreza Bakhshizadeh Mohajer, Alex Dytso, Daniela Tuninetti, Luca Barletta

This paper addresses the fundamental performance limits of Integrated Sensing and Communication (ISAC) systems by introducing a novel converse bound based on rate-distortion theory. This rate-distortion bound (RDB) overcomes the restrictive regularity conditions of classical estimation theory, such as the Bayesian Cram\'er-Rao Bound (BCRB). The proposed fram

Lukas Novak

Let $p$ be a prime number, $K$ a number field that contains the $p$-th root of unity $\zeta_p$, $d$ a $p$-power-free integer and $L=K(\sqrt[p]{d})$. Let $E/K$ be an elliptic curve with full $p$-torsion and $S,T \in E(K)[p]$ be the generators. Define the cocycle $\xi_d : \operatorname{Gal}(\overline{K}/K) \to E$ by \[ \xi_d (\sigma)= \begin{cases} O, & \text{

Manuel Friedrich, Camille Labourie, Kerrek Stinson

We consider regularity of the crack set associated to a minimizer of the Griffith fracture energy, often used in modeling brittle materials. We show that the crack is uniformly rectifiable which in conjunction with our previous epsilon-regularity result allows us to prove that the singular set has dimension strictly less than $1$. This size estimate also app

Chong Mou, Qichao Sun, Yanze Wu, Pengze Zhang

With recent advances in Multimodal Large Language Models (MLLMs) showing strong visual understanding and reasoning, interest is growing in using them to improve the editing performance of diffusion models. Despite rapid progress, most studies lack an in-depth analysis of MLLM design choices. Moreover, the integration of MLLMs and diffusion models remains an