May 2024 arXiv papers — page 64

Henry Hengyuan Zhao, Pan Zhou, Difei Gao, Zechen Bai

Question answering, asking, and assessment are three innate human traits crucial for understanding the world and acquiring knowledge. By enhancing these capabilities, humans can more effectively utilize data, leading to better comprehension and learning outcomes. Current Multimodal Large Language Models (MLLMs) primarily focus on question answering, often ne

Andrew Rosemberg, Alexandre Street, Davi M. Valladão, Pascal Van Hentenryck

Constrained Markov Decision Processes (CMDPs) are critical in many high-stakes applications, where decisions must optimize cumulative rewards while strictly adhering to complex nonlinear constraints. In domains such as power systems, finance, supply chains, and precision robotics, violating these constraints can result in significant financial or societal co

Allison H. Moore, Nicola Tarasca

We study formal series which are invariants of plumbed 3-manifolds twisted by root lattices. These series extend the BPS $q$-series $\widehat{Z}(q)$ recently defined in Gukov-Pei-Putrov-Vafa, Gukov-Manolescu, Park, and further refined in Ri. We show that the series $\widehat{Z}(q)$ is unique in an appropriate sense and decomposes as the average of related se

Emil Albrychiewicz, Ori J. Ganor

We consider a model of a holographic 2+1d CFT interacting with an oscillating background gauge field. It is solved by an AdS-Vaidya metric describing Ohmic heating of the boundary field theory. However, we also show that if timelike singularities of Kasner type are permitted then a time independent solution that may be interpreted as a Floquet state of the s

Osama Bajaber, Bo Ji, Peng Gao

Modern targeted attacks such as Advanced Persistent Threats use multiple hosts as stepping stones and move laterally across them to gain deeper access to the network. However, existing defenses lack end-to-end information flow visibility across hosts and cannot block cross-host attack traffic in real time. In this paper, we propose P4Control, a network defen

Elian Conejo, Eduardo Calderón, Carlos Araya, Ralph Garcia

This paper describe a mechanical ventilator prototype with preclinical test performed on 10 bioporcine models, where results have shown the capabilities to maintain physiological parameter for each subject under test and present also the capability of monitoring the pulmonary parameters, compliance (C), where this is the unique proposed prototype to present

Gabriel Monteiro da Silva, Kyle Lam, David C. Dalgarno, Brenda M. Rubenstein

The intrinsic dynamics of most proteins are central to their function. Protein tyrosine kinases such as Abl1 undergo significant conformational changes that modulate their activity in response to different stimuli. These conformational changes constitute a conserved mechanism for self-regulation that dramatically impacts kinases' affinities for inhibitors. F

Ihsan Caha, Loukya Boddapatti, Aqrab ul Ahmad, Manuel Banobre

CrI3 is a layered ferromagnetic insulator that has recently attracted enormous interest as it was the first example of a stand-alone monolayer ferromagnet, paving the way towards the study of two-dimensional magnetic materials and their use as building blocks of hybrid van der Waals layered heterostructures. Here we go one step down in the dimensionality lad

Joonas Lahikainen, Nadia M. Ady, Christian Guckelsberger

Creativity is already regularly attributed to AI systems outside specialised computational creativity (CC) communities. However, the evaluation of creativity in AI at large typically lacks grounding in creativity theory, which can promote inappropriate attributions and limit the analysis of creative behaviour. While CC researchers have translated psychologic

Marston Copeland, Thomas Mehen

We develop a theoretical framework to match transverse momentum dependent parton distribution functions (TMD PDFs) onto chiral effective theory operators. In this framework the TMD PDF is expressed as a convolution of TMD hadronic distribution functions, which describe fluctuations of initial states into intermediate hadrons in chiral perturbation theory, an

Zexian Zeng, Prajwal Bhagwat, Maryam Saeedifard, Dominic Groß

This paper focuses on the challenging task of bottom-up restoration in a complete blackout system using Grid-forming (GFM) converters. Challenges arise due to the limited current capability of power converters, resulting in distinct dynamic responses and fault current characteristics compared to synchronous generators. Additionally, GFM control needs to addr

Ogan Özsoy, Alexandros Papageorgiou, Matteo Fasiello

Axion-inflation models are a compelling candidate as a mechanism behind the accelerated expansion in the early universe in light of the possibility to embed them in higher dimensional UV complete theories and the exciting prospect of testing them with next-generation cosmological probes. Adding an Abelian gauge sector to axion-inflation models makes for a ri

Kotaro Nagayama, Shota Kato, Manabu Kano

The extraction of variable definitions from scientific and technical papers is essential for understanding these documents. However, the characteristics of variable definitions, such as the length and the words that make up the definition, differ among fields, which leads to differences in the performance of existing extraction methods across fields. Althoug

Chi Hong, Jiyue Huang, Robert Birke, Dick Epema

While diffusion models effectively generate remarkable synthetic images, a key limitation is the inference inefficiency, requiring numerous sampling steps. To accelerate inference and maintain high-quality synthesis, teacher-student distillation is applied to compress the diffusion models in a progressive and binary manner by retraining, e.g., reducing the 1

Siamak Attarian, Dane Morgan, Izabela Szlufarska

In nuclear and/or solar applications that involve molten salts, impurities frequently enter the salt as either fission products or via corrosion. Impurities can interact and make complexes, but the impact of such complexation on the properties of the salts and corrosion rates has not been understood. Common impurities in molten salts, such as FLiBe, include

Jiaxu Wang, Junhao He, Ziyi Zhang, Mingyuan Sun

Event cameras offer promising advantages such as high dynamic range and low latency, making them well-suited for challenging lighting conditions and fast-moving scenarios. However, reconstructing 3D scenes from raw event streams is difficult because event data is sparse and does not carry absolute color information. To release its potential in 3D reconstruct

Soumyadeep Chaudhuri, Frank Ferrari

We compute the scalar determinants $\det(\Delta+M^{2})$ on the two-dimensional round disks of constant curvature $R=0$, $\mp 2$, for any finite boundary length $\ell$ and mass $M$, with Dirichlet boundary conditions, using the $\zeta$-function prescription. When $M^{2}=\pm q(q+1)$, $q\in\mathbb N$, a simple expression involving only elementary functions and

Juan Molina, Mircea Petrache, Francisco Sahli Costabal, Matías Courdurier

Recent works have shown that traditional Neural Network (NN) architectures display a marked frequency bias in the learning process. Namely, the NN first learns the low-frequency features before learning the high-frequency ones. In this study, we rigorously develop a partial differential equation (PDE) that unravels the frequency dynamics of the error for a 2

Hector Kohler, Quentin Delfosse, Riad Akrour, Kristian Kersting

Deep reinforcement learning agents are prone to goal misalignments. The black-box nature of their policies hinders the detection and correction of such misalignments, and the trust necessary for real-world deployment. So far, solutions learning interpretable policies are inefficient or require many human priors. We propose INTERPRETER, a fast distillation me

Rebecca Dorn, Lee Kezar, Fred Morstatter, Kristina Lerman

Content moderation on social media platforms shapes the dynamics of online discourse, influencing whose voices are amplified and whose are suppressed. Recent studies have raised concerns about the fairness of content moderation practices, particularly for aggressively flagging posts from transgender and non-binary individuals as toxic. In this study, we inve

Luca Pasquini, Dinko Milaković

Astrophysics demands higher precision in measurements across photometry, spectroscopy, and astrometry. Several science cases necessitate not only precision but also a high level of accuracy. We highlight the challenges involved, particularly in achieving spectral fidelity, which refers to our ability to accurately replicate the input spectrum of an astrophys

Yoshito Watanabe, Simon Trebst, Ciarán Hickey

Local moments with a spin $S>1/2$ can exhibit a rich variety of elementary quasiparticle excitations, such as quadrupolar excitations, that go beyond the dipolar magnons of conventional spin-$1/2$ systems. However, the experimental observation of such quadrupolar excitations is often challenging due to the dipolar selection rules of many linear response prob

Haoxian Chen, Hanyang Zhao, Henry Lam, David Yao

Direct Preference Optimization (DPO) has recently emerged as a popular approach to improve reinforcement learning with human feedback (RLHF), leading to better techniques to fine-tune large language models (LLM). A weakness of DPO, however, lies in its lack of capability to characterize the diversity of human preferences. Inspired by Mallows' theory of prefe

Wynn C. G. Ho, Nihan Pol, Adam T. Deller, Werner Becker

PSR J0837-2454 is a young 629 ms radio pulsar whose uncertain distance has important implications. A large distance would place the pulsar far out of the Galactic plane and suggest it is the result of a runaway star, while a short distance would mean the pulsar is extraordinarily cold. Here we present further radio observations and the first deep X-ray obser

J. Tucker, P. Strange, P. Mironowicz, J. Quintanilla

We study quantum advantage in one-step rendezvous games on simple graphs analytically, numerically, and using noisy intermediate-scale quantum (NISQ) processors. Our protocols realise the recently discovered [arXiv:2207.14404] optimal bounds for small cycle graphs and cubic graphs. In the case of cycle graphs, we generalise the protocols to arbitrary graph s

Raz Halifa Levi, Yacov Kantor

We consider a $d$-dimensional correlated percolation problem of sites {\em not} visited by a random walk on a hypercubic lattice $L^d$ for $d=3$, 4 and 5. The length of the random walk is ${\cal N}=uL^d$. Close to the critical value $u=u_c$, many geometrical properties of the problem can be described as powers (critical exponents) of $u_c-u$, such as $\beta$

Ruochen Ma

We prove the Lieb-Schultz-Mattis theorem in $d$-dimensional spin systems exhibiting $SO(3)$ spin rotation and lattice translation symmetries in the presence of $k-$local interactions decaying as $\sim 1/r^\alpha$ with distance $r$. Two types of Hamiltonians are considered: Type I comprises long-range spin-spin couplings, while Type II features long-range cou

Donghyeon J. Khim, Dennis Zaritsky, Mika Lambert, Richard Donnerstein

Using the SMUDGes and SDSS catalogs, and our own reprocessing of the Legacy Surveys imaging, we investigate the properties of nuclear star clusters (NSCs) in galaxies having central surface brightnesses as low as 27 mag arcsec$^{-2}$. We identify 273 (123 with known redshift) and 32 NSC-bearing galaxies in the two samples, respectively, where we require cand

Carla Rieger, Michele Grossi, Gian Giacomo Guerreschi, Sofia Vallecorsa

Databases are an essential component of modern computing infrastructures and allow efficient manipulation of inherently structured data. The structure depends on the type and relationships of the individual data elements and on the access pattern. Extending the concept of databases to the quantum domain is expected to increase both the storage efficiency and

Gonzalo Herrera, Shunsaku Horiuchi, Xiaolin Qi

Energetic cosmic rays scatter off the cosmic neutrino background throughout the history of the Universe, yielding a diffuse flux of cosmic relic neutrinos boosted to high energies. We calculate this flux under different assumptions of the cosmic-ray flux spectral slope and redshift evolution. The non-observation of the diffuse flux of boosted relic neutrinos

Giulia Fumagalli, Isobel Romero-Shaw, Davide Gerosa, Viola De Renzis

Resolving the formation channel(s) of merging binary black holes is a key goal in gravitational-wave astronomy. The orbital eccentricity is believed to be a precious tracer of the underlying formation pathway, but is largely dissipated during the usually long inspiral between black hole formation and merger. Most gravitational-wave sources are thus expected

David Benisty, Moshe M. Chaichian, Anca Tureanu

The boundaries of galaxy groups and clusters are defined by the interplay between the Newtonian attractive force and the decoupling from the local expansion of the Universe. This work extends the definition of a zero radial acceleration surface (ZRAS) and the turnaround surface (TS) for a general distribution of the masses in an expanding background, governe

K. Kreckel, O. V. Egorov, E. Egorova, G. A. Blanc

The Orion Molecular Cloud complex, one of the nearest (D = 406 pc) and most extensively studied massive star-forming regions, is ideal for constraining the physics of stellar feedback, but its ~12 deg diameter on the sky requires a dedicated approach to mapping ionized gas structures within and around the nebula. The Sloan Digital Sky Survey (SDSS-V) Local V

Jean Alexandre, Drew Backhouse, Eleni-Alexandra Kontou, Diego Pardo Santos

Both black hole thermodynamics and finite volume effects in quantum field theory violate the null energy condition. Motivated by this, we compare thermodynamic features between two $1+1$-dimensional systems: (i) a scalar field confined to a periodic spatial interval of length $a$ and tunneling between two degenerate vacua; (ii) a dilatonic black hole at temp

Bjorn Larsen, Chiara M. F. Mingarelli, Jeffrey S. Hazboun, Aurelien Chalumeau

Pulsar timing arrays (PTAs) are designed to detect low-frequency gravitational waves (GWs). GWs induce achromatic signals in PTA data, meaning that the timing delays do not depend on radio-frequency. However, pulse arrival times are also affected by radio-frequency dependent "chromatic" noise from sources such as dispersion measure (DM) and scattering delay

A. Mert Bozkurt, Sebastian Miles, Sebastiaan L. D. ten Haaf, Chun-Xiao Liu

We theoretically explore the emergence of strong zero modes in a two-site chain consisting of two quantum dots coupled due to a central dot that mediates electron hopping and singlet superconducting pairing. In the presence of time-reversal symmetry, the on-site Coulomb interaction leads to a three-fold ground-state degeneracy when tuning the system to a swe

Da-Chuan Lu, Zhengdi Sun, Yi-Zhuang You

In this paper, we study the twisted gauging on the (1+1)d lattice and construct various non-local mappings on the lattice operators. To be specific, we define the twisted Gauss law operator and implement the twisted gauging of the finite group on the lattice motivated by the orbifolding procedure in the conformal field theory, which involves the data of non-

William Lake, Claire E. Williams, Smadar Naoz, Federico Marinacci

At high redshifts ($z\gtrsim12$), the relative velocity between baryons and dark matter (the so-called streaming velocity) significantly affects star formation in low-mass objects. Streaming substantially reduces the abundance of low-mass gas objects while simultaneously allowing for the formation of supersonically-induced gas objects (SIGOs) and their assoc

Makoto Sawada, Liyi Gu, Ryo Yamazaki

Over-ionized, recombining plasma is an emerging class of X-ray bright supernova remnants (SNRs). This unique thermal state where the ionization temperature ($T_{\rm z}$) is significantly higher than the electron temperature ($T_{\rm e}$) is not expected from the standard evolution model assuming a point explosion in a uniform interstellar medium, requiring a

Haining Pan, Sriram Ganeshan, Thomas Iadecola, Justin H. Wilson

We theoretically study the topology of the phase diagram of a family of quantum models inspired by the classical Bernoulli map under stochastic control. The quantum models inherit a control-induced phase transition from the classical model and also manifest an entanglement phase transition intrinsic to the quantum setting. This measurement-induced phase tran

Rohit Mukherjee, Haoyu Guo, Keiran Lewellen, Debanjan Chowdhury

Ergodic quantum many-body systems evolving under unitary time dynamics typically lose memory of their initial state via information scrambling. Here we consider a paradigmatic translationally invariant many-body Hamiltonian of interacting bosons -- a Josephson junction array in the transmon regime -- in the presence of a strong Floquet drive. Generically, su

Zakariya Chaouai, Mohamed Tamaazousti

Most of the recent literature on image Super-Resolution (SR) can be classified into two main approaches. The first one involves learning a corruption model tailored to a specific dataset, aiming to mimic the noise and corruption in low-resolution images, such as sensor noise. However, this approach is data-specific, tends to lack adaptability, and its accura

Bjørnulf Brekke, Pavlo Sukhachov, Hans Gløckner Giil, Arne Brataas

New unconventional compensated magnets with a $p$-wave spin polarization protected by a composite time-reversal translation symmetry have been proposed in the wake of altermagnets. To facilitate the experimental discovery and applications of these unconventional magnets, we construct an effective analytical model. The effective model is based on a minimal ti

Adam M. Dillamore, Stephanie Monty, Vasily Belokurov, N. Wyn Evans

We search for globular clusters (GCs) trapped in resonances with the bar of the Milky Way. By integrating their orbits in a potential with a decelerating bar, we select 10 whose orbits are significantly changed by its presence. Most of these are trapped in the corotation resonance (CR), including M22 and 47 Tuc. The decelerating bar is capable of transportin

Dorian W. P. Amaral, Shixiao Liang, Juehang Qin, Christopher Tunnell

Multi-dimensional parameter spaces are commonly encountered in physics theories that go beyond the Standard Model. However, they often possess complicated posterior geometries that are expensive to traverse using techniques traditional to astroparticle physics. Several recent innovations, which are only beginning to make their way into this field, have made

Arthur Platschorre

In this article we study a modification of axion physics in which the dual axion acquires a mass. This mass explicitly breaks the shift symmetry of the dual axion. The potential breaking of this shift symmetry poses a dual axion quality problem. When the dual axion acquires a mass, the axion gets eaten and becomes the longitudinal degree of freedom of a mass

Michael J. Smith, Ryan J. Roberts, Eirini Angeloudi, Marc Huertas-Company

This work presents AstroPT, an autoregressive pretrained transformer developed with astronomical use-cases in mind. The AstroPT models presented here have been pretrained on 8.6 million $512 \times 512$ pixel $grz$-band galaxy postage stamp observations from the DESI Legacy Survey DR8. We train a selection of foundation models of increasing size from 1 milli

Ruizhi Liu, Jinmin Yi, Shiyu Zhou, Liujun Zou

We establish multiple interrelated, fundamental results in quantum many-body systems that can have long-range interactions. For a sufficiently long quantum spin chain, we first show that if the multi-spin interactions in the Hamiltonian decay fast enough as their ranges increase and the Hamiltonian is gapped, then the ground states satisfy the entanglement a

P. Molignini, B. Chakrabarti

We study the non-equilibrium dynamics of a one-dimensional Bose gas with long-range interactions that decay as $(\frac{1}{r^{\alpha}})$ $(0.5 < \alpha <4.0$). We investigate exotic dynamics when the interactions are suddenly switched from strongly repulsive to strongly attractive, a procedure known to generate super-Tonks-Girardeau gases in systems with cont

Benedikt Placke, S. A. Parameswaran

We study the non-unitary dynamics of a class of quantum circuits based on stochastically measuring check operators of subsystem quantum error-correcting codes, such as the Bacon-Shor code and its various generalizations. Our focus is on how properties of the underlying code are imprinted onto the measurement-only dynamics. We find that in a large class of co

Jonas B. Profe, Luke C. Rhodes, Matteo Dürrnagel, Rebecca Bisset

Understanding of unconventional superconductivity is crucial for engineering materials with specific order parameters or elevated superconducting transition temperatures. However, for many materials, the pairing mechanism and symmetry of the order parameter remain unclear: reliable and efficient methods of predicting the order parameter and its response to t

Matteo Poggi, Fabio Tosi

We introduce a novel approach for adapting deep stereo networks in a collaborative manner. By building over principles of federated learning, we develop a distributed framework allowing for demanding the optimization process to a number of clients deployed in different environments. This makes it possible, for a deep stereo network running on resourced-const

Hao Geng

The Karch-Randall braneworld provides a natural set-up to study the Hawking radiation from a black hole using holographic tools. Such a black hole lives on a brane and is highly quantum yet has a holographic dual as a higher dimensional classical theory that lives in the ambient space. Moreover, such a black hole is coupled to a nongravitational bath which i

Dor Verbin, Pratul P. Srinivasan, Peter Hedman, Ben Mildenhall

Neural Radiance Fields (NeRFs) typically struggle to reconstruct and render highly specular objects, whose appearance varies quickly with changes in viewpoint. Recent works have improved NeRF's ability to render detailed specular appearance of distant environment illumination, but are unable to synthesize consistent reflections of closer content. Moreover, t

Jiahao Sun, Chunmei Qing, Xiang Xu, Lingdong Kong

In the rapidly evolving field of autonomous driving, precise segmentation of LiDAR data is crucial for understanding complex 3D environments. Traditional approaches often rely on disparate, standalone codebases, hindering unified advancements and fair benchmarking across models. To address these challenges, we introduce MMDetection3D-lidarseg, a comprehensiv

Yuliang Xiu, Yufei Ye, Zhen Liu, Dimitrios Tzionas

Generating personalized 3D avatars is crucial for AR/VR. However, recent text-to-3D methods that generate avatars for celebrities or fictional characters, struggle with everyday people. Methods for faithful reconstruction typically require full-body images in controlled settings. What if a user could just upload their personal "OOTD" (Outfit Of The Day) phot

Basile Van Hoorick, Rundi Wu, Ege Ozguroglu, Kyle Sargent

Accurate reconstruction of complex dynamic scenes from just a single viewpoint continues to be a challenging task in computer vision. Current dynamic novel view synthesis methods typically require videos from many different camera viewpoints, necessitating careful recording setups, and significantly restricting their utility in the wild as well as in terms o

Tianwei Yin, Michaël Gharbi, Taesung Park, Richard Zhang

Recent approaches have shown promises distilling diffusion models into efficient one-step generators. Among them, Distribution Matching Distillation (DMD) produces one-step generators that match their teacher in distribution, without enforcing a one-to-one correspondence with the sampling trajectories of their teachers. However, to ensure stable training, DM

Hanzhang Tu, Ruizhi Shao, Xue Dong, Shunyuan Zheng

In this paper, we present a low-budget and high-authenticity bidirectional telepresence system, Tele-Aloha, targeting peer-to-peer communication scenarios. Compared to previous systems, Tele-Aloha utilizes only four sparse RGB cameras, one consumer-grade GPU, and one autostereoscopic screen to achieve high-resolution (2048x2048), real-time (30 fps), low-late

Tobias Schnurrenberger, Lucas Happ, Maxim A. Efremov

We show the existence of Borromean bound states in a one-dimensional quantum three-body system composed of two identical bosons and a distinguishable particle. It is assumed that there is no interaction between the two bosons, while the mass-imbalanced two-body subsystems can be tuned to be either bound or unbound. Within the framework of the Faddeev equatio

Zeqi Xiao, Yifan Zhou, Shuai Yang, Xingang Pan

Video generation primarily aims to model authentic and customized motion across frames, making understanding and controlling the motion a crucial topic. Most diffusion-based studies on video motion focus on motion customization with training-based paradigms, which, however, demands substantial training resources and necessitates retraining for diverse models

Asaf Yehudai, Taelin Karidi, Gabriel Stanovsky, Ariel Goldstein

Cross-domain alignment refers to the task of mapping a concept from one domain to another. For example, ``If a \textit{doctor} were a \textit{color}, what color would it be?''. This seemingly peculiar task is designed to investigate how people represent concrete and abstract concepts through their mappings between categories and their reasoning processes ove

Dawid J. Kopiczko, Tijmen Blankevoort, Yuki M. Asano

Decoder-only large language models typically rely solely on masked causal attention, which limits their expressiveness by restricting information flow to one direction. We propose Bitune, a method that enhances pretrained decoder-only LLMs by incorporating bidirectional attention into prompt processing. We evaluate Bitune in instruction-tuning and question-a

Gen Li, Yuling Yan

This paper investigates score-based diffusion models when the underlying target distribution is concentrated on or near low-dimensional manifolds within the higher-dimensional space in which they formally reside, a common characteristic of natural image distributions. Despite previous efforts to understand the data generation process of diffusion models, exi

Joshua Engels, Eric J. Michaud, Isaac Liao, Wes Gurnee

Recent work has proposed that language models perform computation by manipulating one-dimensional representations of concepts ("features") in activation space. In contrast, we explore whether some language model representations may be inherently multi-dimensional. We begin by developing a rigorous definition of irreducible multi-dimensional features based on

S. Alex Breitweiser, Mathieu Ouellet, Tzu-Yung Huang, Tim H. Taminiau

Nuclear quadrupolar resonance (NQR) spectroscopy reveals chemical bonding patterns in materials and molecules through the unique coupling between nuclear spins and local fields. However, traditional NQR techniques require macroscopic ensembles of nuclei to yield a detectable signal, which precludes the study of individual molecules and obscures molecule-to-m

Feng Wang, Jiahao Wang, Sucheng Ren, Guoyizhe Wei

Similar to Vision Transformers, this paper identifies artifacts also present within the feature maps of Vision Mamba. These artifacts, corresponding to high-norm tokens emerging in low-information background areas of images, appear much more severe in Vision Mamba -- they exist prevalently even with the tiny-sized model and activate extensively across backgr

Julian Cremer, Tuan Le, Frank Noé, Djork-Arné Clevert

The generation of ligands that both are tailored to a given protein pocket and exhibit a range of desired chemical properties is a major challenge in structure-based drug design. Here, we propose an in-silico approach for the $\textit{de novo}$ generation of 3D ligand structures using the equivariant diffusion model PILOT, combining pocket conditioning with

Manoj Kumar, Neil Houlsby, Emiel Hoogeboom

Generating image variations, where a model produces variations of an input image while preserving the semantic context has gained increasing attention. Current image variation techniques involve adapting a text-to-image model to reconstruct an input image conditioned on the same image. We first demonstrate that a diffusion model trained to reconstruct an inp

Adam R. Brown

I explore the limitations on the capacity of a relativistic channel to transmit power and information that arise because of the finiteness of the transverse speed of light. As a model system, I consider a rope constructed from a fundamental string, for which relativistic invariance is built in. By wiggling one end of the string, both power and information ma

Yizhou Zhao, Tuanfeng Y. Wang, Bhiksha Raj, Min Xu

Remarkable strides have been made in reconstructing static scenes or human bodies from monocular videos. Yet, the two problems have largely been approached independently, without much synergy. Most visual SLAM methods can only reconstruct camera trajectories and scene structures up to scale, while most HMR methods reconstruct human meshes in metric scale but

Xudong Lu, Aojun Zhou, Ziyi Lin, Qi Liu

Recent developments in large-scale pre-trained text-to-image diffusion models have significantly improved the generation of high-fidelity images, particularly with the emergence of diffusion transformer models (DiTs). Among diffusion models, diffusion transformers have demonstrated superior image-generation capabilities, boosting lower FID scores and higher

Edward S. Hu, James Springer, Oleh Rybkin, Dinesh Jayaraman

We need to look at our shoelaces as we first learn to tie them but having mastered this skill, can do it from touch alone. We call this phenomenon "sensory scaffolding": observation streams that are not needed by a master might yet aid a novice learner. We consider such sensory scaffolding setups for training artificial agents. For example, a robot arm may n

Sayan Das, Duncan Dauvergne, Bálint Virág

Starting from one-point tail bounds, we establish an upper tail large deviation principle for the directed landscape at the metric level. Metrics of finite rate are in one-to-one correspondence with measures supported on a set of countably many paths, and the rate function is given by a certain Kruzhkov entropy of these measures. As an application of our mai

Vladimir Malinovskii, Denis Mazur, Ivan Ilin, Denis Kuznedelev

There has been significant interest in "extreme" compression of large language models (LLMs), i.e., to 1-2 bits per parameter, which allows such models to be executed efficiently on resource-constrained devices. Existing work focused on improved one-shot quantization techniques and weight representations; yet, purely post-training approaches are reaching dim

Can Cui1, Sam Liu, Jaesuk Kwon, Jean Anne C. Incorvia

In spiking neural networks, neuron dynamics are described by the biologically realistic integrate-and-fire model that captures membrane potential accumulation and above-threshold firing behaviors. Among the hardware implementations of integrate-and-fire neuron devices, one important feature, reset, has been largely ignored. Here, we present the design and fa

Szilárd Szabó, Filip Živanović

We classify equivariant $\mathbb{C}^*$-actions on moduli spaces of Higgs bundles corresponding to the Painlev\'e equations. Using this, we compute the Floer-theoretic filtrations on the cohomology of these spaces, introduced by Ritter and the second author in arXiv:2304.13026. We compare it with the ``$P=W$'' and the filtration obtained by multiplicities of

Adam Sky, Michael Neunteufel, Peter Lewintan, Panos Gourgiotis

In this work we present a consistent reduction of the relaxed micromorphic model to its corresponding two-dimensional planar model, such that its capacity to capture discontinuous dilatation fields is preserved. As a direct consequence of our approach, new conforming finite elements for $H^\mathrm{dev}(\mathrm{Curl},A)$ become necessary. We present two novel

Jacqueline Maasch, Kyra Gan, Violet Chen, Agni Orfanoudaki

Identifying the causal pathways of unfairness is a critical objective for improving policy design and algorithmic decision-making. Prior work in causal fairness analysis often requires knowledge of the causal graph, hindering practical applications in complex or low-knowledge domains. Moreover, global discovery methods that learn causal structure from data c

Liwen Wu, Sai Bi, Zexiang Xu, Fujun Luan

Novel-view synthesis of specular objects like shiny metals or glossy paints remains a significant challenge. Not only the glossy appearance but also global illumination effects, including reflections of other objects in the environment, are critical components to faithfully reproduce a scene. In this paper, we present Neural Directional Encoding (NDE), a vie

Mihajlo Cekić, Thibault Lefeuvre

Let $G$ be a compact Lie group. We introduce a semiclassical framework, called Borel-Weil calculus, to investigate $G$-equivariant (pseudo)differential operators acting on $G$-principal bundles over closed manifolds. In this calculus, the semiclassical parameters correspond to the highest roots in the Weyl chamber of the group $G$ that parametrize irreducibl

Stefan Ivanov, Alexander Petkov

We establish in the present paper two sub-gradient estimates for the quaternionic contact (qc) heat equation on a compact qc manifold of dimension $4n+3$, provided some positivity conditions are satisfied. These are qc versions of the prominent Li-Yau gradient estimate in Riemannian geometry. Another goal of this paper is to get two Perelman-type entropy for

Leonardo Medel, Rahul Ghosh, Alberto Martín-Ruiz, Ipsita Mandal

We continue our investigation of the response tensors in planar Hall (or planar thermal Hall) configurations where a three-dimensional Weyl/multi-Weyl semimetal is subjected to the combined influence of an electric field $\mathbf E $ (and/or temperature gradient $\nabla_{\mathbf r } T$) and an effective magnetic field $\mathbf B_\chi $, generalizing the cons

Eduardo M. Gutiérrez, Luciano Combi, Geoffrey Ryan

In this chapter, we give an overview of our current understanding of the physics of accreting massive black hole binaries (MBHBs), with a special focus on the latest developments in numerical simulations and General-Relativistic Magnetohydrodynamics (GRMHD) simulations in particular. We give a self-contained global picture of how to model accretion onto MBHB

Eugene J. Tsao, Alexander J. Lind, Connor Fredrick, Ryan K. Cole

The detection of light of thermal origin is the principal means by which humanity has learned about our world and the cosmos. In optical astronomy, in particular, direct detection of thermal photons and the resolution of their spectra have enabled discoveries of the broadest scope and impact. Such measurements, however, do not capture the phase of the therma

Yihong Sun, Bharath Hariharan

Embodied agents must detect and localize objects of interest, e.g. traffic participants for self-driving cars. Supervision in the form of bounding boxes for this task is extremely expensive. As such, prior work has looked at unsupervised instance detection and segmentation, but in the absence of annotated boxes, it is unclear how pixels must be grouped into

Johannes Zenn, Robert Bamler

Differentiable annealed importance sampling (DAIS), proposed by Geffner & Domke (2021) and Zhang et al. (2021), allows optimizing over the initial distribution of AIS. In this paper, we show that, in the limit of many transitions, DAIS minimizes the symmetrized Kullback-Leibler divergence between the initial and target distribution. Thus, DAIS can be seen as

Yue Yang, Mona Gandhi, Yufei Wang, Yifan Wu

While deep networks have achieved broad success in analyzing natural images, when applied to medical scans, they often fail in unexcepted situations. We investigate this challenge and focus on model sensitivity to domain shifts, such as data sampled from different hospitals or data confounded by demographic variables such as sex, race, etc, in the context of

Yuntian Deng, Yejin Choi, Stuart Shieber

When leveraging language models for reasoning tasks, generating explicit chain-of-thought (CoT) steps often proves essential for achieving high accuracy in final outputs. In this paper, we investigate if models can be taught to internalize these CoT steps. To this end, we propose a simple yet effective method for internalizing CoT steps: starting with a mode

Aashu Katharria, Kanchan Rajwar, Millie Pant, Juan D. Velásquez

Machine learning (ML) is a rapidly evolving technology with expanding applications across various fields. This paper presents a comprehensive survey of recent ML applications in agriculture for sustainability and efficiency. Existing reviews mainly focus on narrow subdomains or lack a fusion-driven perspectives. This study provides a combined analysis of ML

Ruihan Zhang, Jun Sun

Adversarial examples pose a security threat to many critical systems built on neural networks. Given that deterministic robustness often comes with significantly reduced accuracy, probabilistic robustness (i.e., the probability of having the same label with a vicinity is $\ge 1-\kappa$) has been proposed as a promising way of achieving robustness whilst main

Jose Arjona-Medina, Ramil Nugmanov

Despite the rapid and significant advancements in deep learning for Quantitative Structure-Activity Relationship (QSAR) models, the challenge of learning robust molecular representations that effectively generalize in real-world scenarios to novel compounds remains an elusive and unresolved task. This study examines how atom-level pretraining with quantum me

Alberto Larrauri, Guillem Perarnau

We consider limit probabilities of first order properties in random graphs with a given degree sequence. Under mild conditions on the degree sequence, we show that the closure set of limit probabilities is a finite union of closed intervals. Moreover, we characterize the degree sequences for which this closure set is the interval $[0,1]$, a property that is

Sepehr Assadi, Prantar Ghosh, Bruno Loff, Parth Mittal

The following question arises naturally in the study of graph streaming algorithms: "Is there any graph problem which is "not too hard", in that it can be solved efficiently with total communication (nearly) linear in the number $n$ of vertices, and for which, nonetheless, any streaming algorithm with $\tilde{O}(n)$ space (i.e., a semi-streaming algorithm) n

Sun-Kai Leung

Assuming the generalized Lindel\"{o}f hypothesis (GLH), a weak version of the generalized Ramanujan conjecture and a Rankin--Selberg type partial sum estimate, we establish the normality of the sum of coefficients of a general $L$-function in short intervals of appropriate length. The novelty lies in the degree aspect under GLH. In particular, this generaliz

Adam LaClair

A famous theorem of Kalai and Meshulam is that $\mathrm{reg}(I + J) \leq \mathrm{reg}(I) + \mathrm{reg}(J) -1$ for any squarefree monomial ideals $I$ and $J$. This result was subsequently extended by Herzog to the case where $I$ and $J$ are any monomial ideals. In this paper we conjecture that the Castelnuovo--Mumford regularity is subadditive on binomial ed

Shuang Wu, Youtian Lin, Feihu Zhang, Yifei Zeng

Generating high-quality 3D assets from text and images has long been challenging, primarily due to the absence of scalable 3D representations capable of capturing intricate geometry distributions. In this work, we introduce Direct3D, a native 3D generative model scalable to in-the-wild input images, without requiring a multiview diffusion model or SDS optimi

Bernal Jiménez Gutiérrez, Yiheng Shu, Yu Gu, Michihiro Yasunaga

In order to thrive in hostile and ever-changing natural environments, mammalian brains evolved to store large amounts of knowledge about the world and continually integrate new information while avoiding catastrophic forgetting. Despite the impressive accomplishments, large language models (LLMs), even with retrieval-augmented generation (RAG), still struggl

Xinxin Cai, Habitamu Y. Walelign, John M. Nichol

Silicon-based qubits are often made by trapping individual electrons in quantum dots defined by electric gates. Quantum information can then be stored using the spin states of the electrons. However, the nuclei of the surrounding atoms also have spin degrees of freedom that couple to the electron spin qubits and cause decoherence. The emergence of a nuclear-

Anuj Apte, Anthony Ashmore, Clay Cordova, Tzu-Chen Huang

Monte Carlo methods have led to profound insights into the strong-coupling behaviour of lattice gauge theories and produced remarkable results such as first-principles computations of hadron masses. Despite tremendous progress over the last four decades, fundamental challenges such as the sign problem and the inability to simulate real-time dynamics remain.