Search

Diffusion-based text-to-image models have demonstrated impressive achievements in diversity and aesthetics but struggle to generate images with legible visual texts. Existing backbone models have limitations such as misspelling, failing to generate texts, and lack of support for Chinese text, but their development shows promising potential. In this paper, we propose a series of methods, aiming to empower backbone models to generate visual texts in English and Chinese. We first conduct a preliminary study revealing that Byte Pair Encoding (BPE) tokenization and the insufficient learning of cross-attention modules restrict the performance of the backbone models. Based on these observations, we make the following improvements: (1) We design a mixed granularity input strategy to provide more suitable text representations; (2) We propose to augment the conventional training objective with three glyph-aware training losses, which enhance the learning of cross-attention modules and encourage the model to focus on visual texts. Through experiments, we demonstrate that our methods can effectively empower backbone models to generate semantic relevant, aesthetically appealing, and accurate visual text images, while maintaining their fundamental image generation quality.

This study delves into Microtearing Modes (MTMs) in tokamak plasmas, employing advanced simulations within the BOUT++ framework. The research, centering on collisional MTMs influenced by the time-dependent thermal force, enhances our understanding of plasma dynamics. It achieves this through the simplification and linearization of control equations in detailed linear simulations. The study meticulously evaluates various conductivity models, including those proposed by Larakers, Drake, and Hassam, under diverse plasma conditions and collision regimes. A notable achievement of this research is the derivation of a unified dispersion relation that encompasses both MTM and Drift-Alfven Wave (DAW) instabilities. It interestingly reveals that DAW and MTM exhibit instability at different proximities to the rational surface. Specifically, MTMs become unstable near the rational surface but stabilize farther away, whereas the drift-Alfven instability manifests away from the rational surface. Further, the study re-derives MTM dispersion relations based on Ohm's law and the vorticity equation, providing a thorough analysis of electromagnetic and electrostatic interactions in tokamaks. Global simulations demonstrate an inverse correlation between MTM growth rates and collisionality, and a direct correlation with temperature gradients. The nonalignment of the rational surface with the peak of electron local diamagnetic frequency stabilizes the MTMs. Nonlinear simulations highlight electron temperature relaxation as the primary saturation mechanism for MTMs, with magnetic flutter identified as the dominant mode of electron thermal transport., Comment: 31 pages, 13 figures

Utilizing a tritter with variable parameter $T$ and induced by vacuum--one-photon superpositions $\left\vert 0\right\rangle +\alpha\left\vert 1\right\rangle $ with $\alpha =\left\vert \alpha \right\vert e^{i\phi }$, we propose a scheme to prepare a class of three-qubit pure states. These states take the form of $\left\vert \psi \right\rangle _{123}=c_{0}\left\vert 000\right\rangle +c_{1}\left\vert 100\right\rangle +c_{2}\left\vert 010\right\rangle +c_{3}\left\vert 001\right\rangle $. The coefficients ($c_{0}$, $c_{1}$, $c_{2}$, and $c_{3}$) can be manipulated through interaction parameters ($\left\vert \alpha \right\vert $, $\phi $, and $T$). In line with Xie and Eberly's work[Phys. Rev. Lett. 127, 040403 (2021)], we investigate the genuine tripartite entanglement for $\left\vert \psi \right\rangle _{123}$ by using the measure of concurrence fill. Drawing on Hao \textit{et al.}'s research [Phys. Rev. Lett. 128, 120402 (2021)], we examine tripartite steering for $\left\vert \psi \right\rangle _{123}$ under certain measurements based on the uncertainty relations criterion. We identify nine potential configurations exhibiting varying steerability across different parameter spaces. It is important to highlight that, while the state $\left\vert \psi \right\rangle _{123}$ exhibits entanglement, steering remains unattainable in a substantial portion of the parameter space., Comment: 12 pages, 8 figures, close to published version

Existing text-to-image diffusion models primarily generate images from text prompts. However, the inherent conciseness of textual descriptions poses challenges in faithfully synthesizing images with intricate details, such as specific entities or scenes. This paper presents UNIMO-G, a simple multimodal conditional diffusion framework that operates on multimodal prompts with interleaved textual and visual inputs, which demonstrates a unified ability for both text-driven and subject-driven image generation. UNIMO-G comprises two core components: a Multimodal Large Language Model (MLLM) for encoding multimodal prompts, and a conditional denoising diffusion network for generating images based on the encoded multimodal input. We leverage a two-stage training strategy to effectively train the framework: firstly pre-training on large-scale text-image pairs to develop conditional image generation capabilities, and then instruction tuning with multimodal prompts to achieve unified image generation proficiency. A well-designed data processing pipeline involving language grounding and image segmentation is employed to construct multi-modal prompts. UNIMO-G excels in both text-to-image generation and zero-shot subject-driven synthesis, and is notably effective in generating high-fidelity images from complex multimodal prompts involving multiple image entities., Comment: Accepted by ACL 2024, Main Conference, Long Paper

In this paper, we introduce a noiselessly amplified thermal state (ATS), by operating the noiseless amplification operator ($g^{\hat{n}}$) on the thermal state (TS) with corresponding mean photon number (MPN) $\bar{n}$. Actually, the ATS is an new TS with MPN $\bar{N}=g^{2}\bar{n}/[1-\bar{n}\left(g^{2}-1\right)]$. Furthermore, we introduce photon-added-ATS (PAATS) and photon-subtracted-ATS (PSATS) by operating $m$-photon addition ($\hat{a}^{\dag m}$) and $m$-photon subtraction ($\hat{a}^{m}$) on the ATS, respectively. We study photon number distributions (PNDs), purities, and Wigner functions (WFs) for all these states., Comment: 8 pages, 8 figures

The fine-grained action analysis of the existing action datasets is challenged by insufficient action categories, low fine granularities, limited modalities, and tasks. In this paper, we propose a Multi-modality and Multi-task dataset of Figure Skating (MMFS) which was collected from the World Figure Skating Championships. MMFS, which possesses action recognition and action quality assessment, captures RGB, skeleton, and is collected the score of actions from 11671 clips with 256 categories including spatial and temporal labels. The key contributions of our dataset fall into three aspects as follows. (1) Independently spatial and temporal categories are first proposed to further explore fine-grained action recognition and quality assessment. (2) MMFS first introduces the skeleton modality for complex fine-grained action quality assessment. (3) Our multi-modality and multi-task dataset encourage more action analysis models. To benchmark our dataset, we adopt RGB-based and skeleton-based baseline methods for action recognition and action quality assessment.

The edge plasma turbulence and transport dynamics, as well as the divertor power loads during the thermal quench phase of tokamak disruptions are numerically investigated with BOUT++'s flux-driven, six-field electromagnetic turbulence model. Here a transient yet intense particle and energy sources are applied at the pedestal top to mimic the plasma power drive at the edge induced by a core thermal collapse, which flattens core temperature profile. Interesting features such as surging of divertor heat load (up to 50 times), and broadening of heat flux width (up to 4 times) on the outer divertor target plate, are observed in the simulation, in qualitative agreement with experimental observations. The dramatic changes of divertor heat load and width are due to the enhanced plasma turbulence activities inside the separatrix. Two cross-field transport mechanisms, namely the $E\times B$ turbulent convection and the stochastic parallel advection/conduction, are identified to play important roles in this process. Firstly, elevated edge pressure gradient drives instabilities and subsequent turbulence in the entire pedestal region. The enhanced turbulence not only transports particles and energy radially across the separatrix via $E\times B$ convection which causes the initial divertor heat load burst, but also induces an amplified magnetic fluctuation $\tilde{B}$. Once the magnetic fluctuation is large enough to break the magnetic flux surface, magnetic flutter effect provides an additional radial transport channel. In the late stage of our simulation, $|\tilde{B}_r/B_0|$ reaches to $10^{-4}$ level that completely breaks magnetic flux surfaces such that stochastic field-lines are directly connecting pedestal top plasma to the divertor target plates or first wall, further contributing to the divertor heat flux width broadening., Comment: 20 pages, 12 figures

Multipartite Einstein-Podolsky-Rosen (EPR) steering and multimode quantum squeezing are essential resources for various quantum applications. The paper focuses on studying a coupled three-mode squeezed vacuum (C3MSV), which is a typical multimode squeezed Gaussian state and will exhibit peculiar steering property. Using the technique of integration within ordered products, we give the normal-ordering form for the coupled three-mode squeezing operator and derive the general analytical expressions of the statistical quantities for the C3MSV. Under Gaussian measurements, we analyze all bipartite Gaussian steerings (including no steering, one-way steering and two-way steering) in details and study the monogamy relations for the C3MSV. Then, we study the decoherence of all these steerings in noisy channels and find that sudden death will happen in a certain threshold time. Through the steerings shared in the C3MSV, we propose conceptual (and ideal) schemes of remotely generating Wigner negativity (WN) by performing appropriate photon subtraction(s) in the local position. Our obtained results may lay a solid theoretical foundation for a future practical study. We also believe that the C3MSV will be one of good candidate resources in future quantum protocols., Comment: 16 pages, 11 figures, an improvement of the first pre-print version. The published Ref.[104] has been corrected as "J. Phys.A: Math. Gen. 23, 1833 (1990)"

Diffusion generative models have recently greatly improved the power of text-conditioned image generation. Existing image generation models mainly include text conditional diffusion model and cross-modal guided diffusion model, which are good at small scene image generation and complex scene image generation respectively. In this work, we propose a simple yet effective approach, namely UPainting, to unify simple and complex scene image generation, as shown in Figure 1. Based on architecture improvements and diverse guidance schedules, UPainting effectively integrates cross-modal guidance from a pretrained image-text matching model into a text conditional diffusion model that utilizes a pretrained Transformer language model as the text encoder. Our key findings is that combining the power of large-scale Transformer language model in understanding language and image-text matching model in capturing cross-modal semantics and style, is effective to improve sample fidelity and image-text alignment of image generation. In this way, UPainting has a more general image generation capability, which can generate images of both simple and complex scenes more effectively. To comprehensively compare text-to-image models, we further create a more general benchmark, UniBench, with well-written Chinese and English prompts in both simple and complex scenes. We compare UPainting with recent models and find that UPainting greatly outperforms other models in terms of caption similarity and image fidelity in both simple and complex scenes. UPainting project page \url{https://upainting.github.io/}., Comment: First Version, 16 pages

Based on heralded interference on a six-port Mach-Zehnder interferometer, we propose protocols to generate a series of multiphoton states in primary output port, by injecting a coherent state in primary input port and two Fock states in two ancillary input ports, and measuring two Fock states in two ancillary output ports. Only manipulating at the single-photon level (i.e, |0> or |1>) in all ancillary ports, we generate sixteen types (six categories) of multiphoton nonclassical states, whose state vectors are unified as superposition of a new coherent state, a single-photon added coherent state, and a two-photon added coherent state. Indeed, a wide range of nonclassical phenomena can be created by modulating the interaction parameters (including coherent field strength and shift phase). We mainly analyze quadrature-squeezing effects for all our considered states. Of particular interest is maximum squeezing of up to 2.57dB, with success probability 6.7%, at least in our present cases., Comment: 7 pages, 6 figures, comments are welcome

State $g^{\hat{n}}\hat{a}^{\dag m}\left\vert \alpha \right\rangle $ and state $\hat{a}^{\dag m}g^{\hat{n}}\left\vert \alpha \right\rangle $ are same to state $\hat{a}^{\dag m}\left\vert g\alpha \right\rangle $, which is called as multi-photon-addition amplified coherent state (MPAACS) by us. Here, $\hat{n}$, $\hat{a}^{\dag }$, $\left\vert \alpha \right\rangle $, $g$ ( $\geq 1$), and $m$ are photon number operator, creation operator, coherent state, gain facor, and an interger, respectively. We study mathematical and physical properties for these MPAACSs, including normalization, photon component analysis, Wigner function, effective gain, quadrature squeezing, and equivalent input noise. Actually, the MPAACS, which contains more nonclassicality, is an amplified version of photon-added coherent state (PACS) introduced by Agrwal and Tara [Phys. Rev. A 43, 492 (1991)]. Our work provides theoretical references for implementing amplifiers for light fields., Comment: 10 pages, 5 figures

We present a fast and accurate data-driven surrogate model for divertor plasma detachment prediction leveraging the latent feature space concept in machine learning research. Our approach involves constructing and training two neural networks. An autoencoder that finds a proper latent space representation (LSR) of plasma state by compressing the multi-modal diagnostic measurements, and a forward model using multi-layer perception (MLP) that projects a set of plasma control parameters to its corresponding LSR. By combining the forward model and the decoder network from autoencoder, this new data-driven surrogate model is able to predict a consistent set of diagnostic measurements based on a few plasma control parameters. In order to ensure that the crucial detachment physics is correctly captured, highly efficient 1D UEDGE model is used to generate training and validation data in this study. Benchmark between the data-driven surrogate model and UEDGE simulations shows that our surrogate model is capable to provide accurate detachment prediction (usually within a few percent relative error margin) but with at least four orders of magnitude speed-up, indicating that performance-wise, it has the potential to facilitate integrated tokamak design and plasma control. Comparing to the widely used two-point model and/or two-point model formatting, the new data-driven model features additional detachment front prediction and can be easily extended to incorporate richer physics. This study demonstrates that the complicated divertor and scrape-off-layer plasma state has a low-dimensional representation in latent space. Understanding plasma dynamics in latent space and utilizing this knowledge could open a new path for plasma control in magnetic fusion energy research., Comment: 24 pages, 15 figures

It is an extremely challenging task to detect arbitrary shape text in natural scenes on high accuracy and efficiency. In this paper, we propose a scene text detection framework, namely GWNet, which mainly includes two modules: Global module and RCNN module. Specifically, Global module improves the adaptive performance of the DB (Differentiable Binarization) module by adding k submodule and shift submodule. Two submodules enhance the adaptability of amplifying factor k, accelerate the convergence of models and help to produce more accurate detection results. RCNN module fuses global-level and word-level features. The word-level label is generated by obtaining the minimum axis-aligned rectangle boxes of the shrunk polygon. In the inference period, GWNet only uses global-level features to output simple polygon detections. Experiments on four benchmark datasets, including the MSRA-TD500, Total-Text, ICDAR2015 and CTW-1500, demonstrate that our GWNet outperforms the state-of-the-art detectors. Specifically, with a backbone of ResNet-50, we achieve an F-measure of 88.6% on MSRA- TD500, 87.9% on Total-Text, 89.2% on ICDAR2015 and 87.5% on CTW-1500.

We studied the effect of delocalized single-photon addition (DPA) on two input modes containing four cases: two independent coherent states (CSs), two independent thermal states (TSs), two independent single-mode squeezed vacuums (SVs), and an entangled two-mode squeezed vacuum (TMSV). In essence, four types of new non-Gaussian entangled light states are generated. We studied three different resources (including entanglement, discorrelation and Wigner negativity) for each two-mode light state. The output states after DPA are entangled, with more parameters and complex structures, characterizing more Wigner negativity or even discorrelation. In contrast, the CSs case is the most tunable protocol, because its negativity under partial transposition, discorrelation, and Wigner logarithmic negativity are more sensitive to superposition phase than those in TSs, SVs and TMSV cases.