Search

It is usually assumed that $4D$ instantons can only arise in non-Abelian theories. In this paper we re-examine this conventional wisdom by explicitly constructing instantons in an Abelian gauge theory: ${\rm QED}_4$ with $N_f$ flavors of Dirac fermions, in the background of a Dirac monopole. This is the low-energy effective field theory for fermions interacting with a 't Hooft-Polyakov monopole, in the limit where the monopole is infinitely heavy (hence pointlike) and static. This theory, whose non-topological sectors were studied by Rubakov and Callan, has a far richer structure than previously explored. We show how to calculate the topological instanton number, demonstrate the existence of 't Hooft zero modes localized around such instantons, and show how instantons in the path integral provide the underlying mechanism for the Callan-Rubakov process: monopole-catalyzed baryon decay with a cross section that saturates the unitarity bound. Our computation relies on correctly identifying the relevant $2D$ EFT for monopole catalysis as Axial ${\rm QED}_2$ in an effective $AdS_2$ metric., Comment: 43 pages main text, 12 pages appendix, 4 figures

The Dicke model is a staple of theoretical cavity Quantum Electrodynamics (cavity QED), describing the interaction between an ensemble of atoms and a single radiation mode of an optical cavity. It has been studied both quantum mechanically and semiclassically for two-level atoms, and demonstrates a rich variety of dynamics such as phase transitions, phase multistability, and chaos. In this work we explore an open, spin-1 Dicke model with independent co- and counter-rotating coupling terms as well as a quadratic Zeeman shift enabling control over the atomic energy-level structure. We investigate the stability of operator and moment equations under two approximations and show the system undergoes phase transitions. To compliment these results, we relax the aforementioned approximations and investigate the system semiclassically. We show evidence of phase transitions to steady-state and oscillatory superradiance in this semiclassical model, as well as the emergence of chaotic dynamics. The varied and complex behaviours admitted by the model highlights the need to more rigorously map its dynamics., Comment: 16 pages, 12 figures

Image collection summarization techniques aim to present a compact representation of an image gallery through a carefully selected subset of images that captures its semantic content. When it comes to web content, however, the ideal selection can vary based on the user's specific intentions and preferences. This is particularly relevant at Booking.com, where presenting properties and their visual summaries that align with users' expectations is crucial. To address this challenge, we consider user intentions in the summarization of property visuals by analyzing property reviews and extracting the most significant aspects mentioned by users. By incorporating the insights from reviews in our visual summaries, we enhance the summaries by presenting the relevant content to a user. Moreover, we achieve it without the need for costly annotations. Our experiments, including human perceptual studies, demonstrate the superiority of our cross-modal approach, which we coin as CrossSummarizer over the no-personalization and image-based clustering baselines.

Multi-label text classification is a critical task in the industry. It helps to extract structured information from large amount of textual data. We propose Text to Topic (Text2Topic), which achieves high multi-label classification performance by employing a Bi-Encoder Transformer architecture that utilizes concatenation, subtraction, and multiplication of embeddings on both text and topic. Text2Topic also supports zero-shot predictions, produces domain-specific text embeddings, and enables production-scale batch-inference with high throughput. The final model achieves accurate and comprehensive results compared to state-of-the-art baselines, including large language models (LLMs). In this study, a total of 239 topics are defined, and around 1.6 million text-topic pairs annotations (in which 200K are positive) are collected on approximately 120K texts from 3 main data sources on Booking.com. The data is collected with optimized smart sampling and partial labeling. The final Text2Topic model is deployed on a real-world stream processing platform, and it outperforms other models with 92.9% micro mAP, as well as a 75.8% macro mAP score. We summarize the modeling choices which are extensively tested through ablation studies, and share detailed in-production decision-making steps.

Can classical systems be described analytically at all orders in their interaction strength? For periodic and approximately periodic systems, the answer is yes, as we show in this work. Our analytical approach, which we call the \textit{Quantum Spectral Method}, is based on a novel application of Bohr's correspondence principle, obtaining non-perturbative classical dynamics as the classical limit of \textit{quantum matrix elements}. A major application of our method is the calculation of self-force as the classical limit of atomic radiative transitions. We demonstrate this by calculating an adiabatic electromagnetic inspiral, along with its associated radiation, at all orders in the multipole expansion. Finally, we propose a future application of the QSM to compute scalar and gravitational self-force in Schwarzschild, analytically.

Through digital imaging, microscopy has evolved from primarily being a means for visual observation of life at the micro- and nano-scale, to a quantitative tool with ever-increasing resolution and throughput. Artificial intelligence, deep neural networks, and machine learning are all niche terms describing computational methods that have gained a pivotal role in microscopy-based research over the past decade. This Roadmap is written collectively by prominent researchers and encompasses selected aspects of how machine learning is applied to microscopy image data, with the aim of gaining scientific knowledge by improved image quality, automated detection, segmentation, classification and tracking of objects, and efficient merging of information from multiple imaging modalities. We aim to give the reader an overview of the key developments and an understanding of possibilities and limitations of machine learning for microscopy. It will be of interest to a wide cross-disciplinary audience in the physical sciences and life sciences.

We present Neural Congealing -- a zero-shot self-supervised framework for detecting and jointly aligning semantically-common content across a given set of images. Our approach harnesses the power of pre-trained DINO-ViT features to learn: (i) a joint semantic atlas -- a 2D grid that captures the mode of DINO-ViT features in the input set, and (ii) dense mappings from the unified atlas to each of the input images. We derive a new robust self-supervised framework that optimizes the atlas representation and mappings per image set, requiring only a few real-world images as input without any additional input information (e.g., segmentation masks). Notably, we design our losses and training paradigm to account only for the shared content under severe variations in appearance, pose, background clutter or other distracting objects. We demonstrate results on a plethora of challenging image sets including sets of mixed domains (e.g., aligning images depicting sculpture and artwork of cats), sets depicting related yet different object categories (e.g., dogs and tigers), or domains for which large-scale training data is scarce (e.g., coffee mugs). We thoroughly evaluate our method and show that our test-time optimization approach performs favorably compared to a state-of-the-art method that requires extensive training on large-scale datasets., Comment: Project page: https://neural-congealing.github.io/

We study representations of the Poincar\'e group that have a privileged transformation law along a p-dimensional hyperplane, and uncover their associated spinor helicity variables in D spacetime dimensions. Our novel representations generalize the recently introduced celestial states and transform as conformal primaries of SO(p,1), the symmetry group of the p-hyperplane. We will refer to our generalized states as ``partially celestial." Following Wigner's method, we find the induced representations, including spin degrees of freedom. Defining generalized spinor helicity variables for every D and p, we are able to construct the little group covariant part of partially celestial amplitudes. Finally, we briefly examine the application of the pairwise little group to partially celestial states with mutually non-local charges., Comment: 10 pages, RevTeX

We present a careful study of the chiral symmetry breaking minima and the baryonic directions in supersymmetric QCD ($SU(N_c)$ with $N_f$ flavors) perturbed by Anomaly Mediated Supersymmetry Breaking (AMSB). For the s-confining case of $N_f = N_c + 1$ and most of the free-magnetic phase ($N_f \leq 1.43 N_c$) we find that naive tree level baryonic runaways are stabilized by loop effects. Runaways are present, however, for the upper end of the free magnetic phase ($N_f \gtrsim 1.43 N_c$) and into conformal window, signaling the existence of incalculable minima at large field values of ${\cal O} (\Lambda)$. Nevertheless, the chiral symmetry breaking points are locally stable, and are expected to continuously connect to the vacua of QCD for large SUSY breaking. The case of $N_f = N_c$ requires particular care due to the inherently strongly coupled nature of the quantum modified moduli space. Due to the incalculability of critical K\"ahler potential terms, the stability of the chiral symmetry breaking point along baryonic directions cannot be determined for $N_f=N_c$. With the exception of this case, all theories to which AMSB can be applied ($N_f < 3 N_c$) possess stable chiral symmetry breaking minima, and all theories with $N_f \lesssim 1.43 N_c$ (aside from $N_f = N_f$) are protected from runaways to incalculable minima., Comment: 10 pages

We construct the Faddeev-Kulish dressed multiparticle states of electrically and magnetically charged particles, incorporating the effects of real and virtual soft photons. We calculate the properties of such dressed states under Lorentz transformations, and find that they can be identified with the pairwise multi-particle states that transform under the pairwise little group. The shifts in the dressing factors under Lorentz transformations are finite and have a simple geometric interpretation. Using the transformation properties of the dressed states we also present a novel, fully quantum field theoretic derivation of the geometric (Berry) phase obtained by an adiabatic rotation of the Dirac string, and also of the Dirac quantization condition. For half integer pairwise helcity, we show that these multiparticle states have flipped spin-statistics, reproducing the surprising fact that fermions can be made out of bosons.

We present a method for zero-shot, text-driven appearance manipulation in natural images and videos. Given an input image or video and a target text prompt, our goal is to edit the appearance of existing objects (e.g., object's texture) or augment the scene with visual effects (e.g., smoke, fire) in a semantically meaningful manner. We train a generator using an internal dataset of training examples, extracted from a single input (image or video and target text prompt), while leveraging an external pre-trained CLIP model to establish our losses. Rather than directly generating the edited output, our key idea is to generate an edit layer (color+opacity) that is composited over the original input. This allows us to constrain the generation process and maintain high fidelity to the original input via novel text-driven losses that are applied directly to the edit layer. Our method neither relies on a pre-trained generator nor requires user-provided edit masks. We demonstrate localized, semantic edits on high-resolution natural images and videos across a variety of objects and scenes., Comment: Project page: https://text2live.github.io

Seagrass meadows are twice as efficient as forests at capturing and storing carbon, but over the last two decades they have been disappearing due to human activities. We take a nature-centered design approach using contextual inquiry and iterative participatory designs methods to consolidate knowledge from the marine and material sciences to industrial design. The sketches and renders documented evolved into the design and fabrication guidelines. This pictorial documents a dialogue between designers and scientists to design an ecological intervention using digital fabrication to manufacture morphing ceramics for seagrass meadow restoration., Comment: 12 pages with 32 figures, ACM C&C Pictorial

We extract the relativistic classical radial action from scattering amplitudes, to all orders in perturbation theory, in the probe limit. Our sources include point charges and monopoles, as well as the Schwarzschild and pure-NUT gravitational backgrounds. A characteristic relativistic effect, that scattering trajectories may wind around these sources any number of times, can be recovered when all-order amplitudes are available. We show that the amplitude for scattering a probe off a pure NUT is given by the solution of a transcendental equation involving continued fractions, and explain how to solve this equation to any desired loop order., Comment: Fixed typo in Schwarzschild scattering angle

We present a method that decomposes, or "unwraps", an input video into a set of layered 2D atlases, each providing a unified representation of the appearance of an object (or background) over the video. For each pixel in the video, our method estimates its corresponding 2D coordinate in each of the atlases, giving us a consistent parameterization of the video, along with an associated alpha (opacity) value. Importantly, we design our atlases to be interpretable and semantic, which facilitates easy and intuitive editing in the atlas domain, with minimal manual work required. Edits applied to a single 2D atlas (or input video frame) are automatically and consistently mapped back to the original video frames, while preserving occlusions, deformation, and other complex scene effects such as shadows and reflections. Our method employs a coordinate-based Multilayer Perceptron (MLP) representation for mappings, atlases, and alphas, which are jointly optimized on a per-video basis, using a combination of video reconstruction and regularization losses. By operating purely in 2D, our method does not require any prior 3D knowledge about scene geometry or camera poses, and can handle complex dynamic real world videos. We demonstrate various video editing applications, including texture mapping, video style transfer, image-to-video texture transfer, and segmentation/labeling propagation, all automatically produced by editing a single 2D atlas image.

We resolve the decades old mystery of what happens when a positron scatters off a minimal GUT monopole in an s-wave, first discussed by Callan in 1983. Using the language of on-shell amplitudes and pairwise helicity we show that the final state contains two up quarks and a down quark entangled with angular momentum stored in the gauge fields, which is the only particle final state that satisfies angular momentum and gauge charge conservation. The cross section for this process is as large as in the original Rubakov-Callan effect, only suppressed by the QCD scale. The final state we find cannot be seen in Callan's truncated 2D theory, since our entanglement requires more than 2 dimensions.

We investigate the IR phases of non-supersymmetric (non-SUSY) $SO(N_c)$ gauge theories with $N_F$ fermions in the vector representation obtained by perturbing the SUSY theory with anomaly mediated SUSY breaking (AMSB). We find that of the wide variety of phases appearing in the SUSY theory only two survive: for $N_F<\frac{3}{2} (N_c-2)$ the theory confines, breaking the $SU(N_F)$ global symmetry to $SO(N_F)$, while for $\frac{3}{2} (N_c-2)

We demonstrate that $SO(N_{c})$ gauge theories with matter fields in the vector representation confine due to monopole condensation and break the $SU(N_{F})$ chiral symmetry to $SO(N_{F})$ via the quark bilinear. Our results are obtained by perturbing the ${\cal N}=1$ supersymmetric theory with anomaly-mediated supersymmetry breaking., Comment: arXiv admin note: substantial text overlap with arXiv:2107.02813

We study dynanics of $SU(N-4)$ gauge theories with fermions in rank-2 symmetric tensor and $N$ anti-fundamental representations, by perturbing supersymmetric theories with anomaly-mediated supersymmetry breaking. We find the $SU(N)\times U(1)$ global symmetry is dynamically broken to $SO(N)$ for $N\geq 17$, a different result from conjectures in the literature. For $N<17$, theories flow to infrared fixed points.

We analyze dynamics of chiral gauge theories based on the $SU(N)$ gauge group with one anti-symmetric tensor $A$ and $(N-4)$ anti-fundamentals $F_{i}$ when $N$ is odd. Based on the continuity to the supersymmetric gauge theories with anomaly-mediated supersymmetry breaking, we claim that the global $SU(N-4)$ symmetry is spontaneously broken to $Sp(N-5)$. There are $N-5$ massless fermions as a fundamental representation of $Sp(N-5)$, and another massless fermion, together saturating the anomaly matching conditions. When $N$ is even, the unbroken flavor symmetry is $Sp(N-4)$ while there are no massless fermions. Our result is different from the dynamics suggested by tumbling where the full $SU(N-4)$ symmetry is unbroken, but the tumbling picture can be modified via the addition of a second condensate to produce the symmetry breaking pattern predicted from our method.