Search

Font design poses unique challenges for logographic languages like Chinese, Japanese, and Korean (CJK), where thousands of unique characters must be individually crafted. This paper introduces a novel Vision Transformer (ViT)-based model for multi-language font generation, effectively addressing the complexities of both logographic and alphabetic scripts. By leveraging ViT and pretraining with a strong visual pretext task (Masked Autoencoding, MAE), our model eliminates the need for complex design components in prior frameworks while achieving comprehensive results with enhanced generalizability. Remarkably, it can generate high-quality fonts across multiple languages for unseen, unknown, and even user-crafted characters. Additionally, we integrate a Retrieval-Augmented Guidance (RAG) module to dynamically retrieve and adapt style references, improving scalability and real-world applicability. We evaluated our approach in various font generation tasks, demonstrating its effectiveness, adaptability, and scalability.

The CompactObject package is an open-source software framework developed to constrain the neutron star equation of state (EOS) through Bayesian statistical inference. It integrates astrophysical observational constraints from X-ray timing, gravitational wave events, and radio measurements, as well as nuclear experimental constraints derived from perturbative Quantum Chromodynamics (pQCD) and Chiral Effective Field Theory ($\chi$EFT). The package supports a diverse range of EOS models, including meta-model like and several physics-motivated EOS models. It comprises three independent components: an EOS generator module that currently provides seven EOS choices, a Tolman-Oppenheimer-Volkoff (TOV) equation solver, that allows the determination of the Mass Radius and Tidal deformability as observables, and a comprehensive Bayesian inference workflow module, including a complete pipeline for implementing EOS Bayesian inference. Each component can be used independently in different scientific research contexts, such as nuclear physics and astrophysics. In addition, CompactObject is designed to work in synergy with existing software such as CompOSE, allowing the use of the CompOSE EOS database to extend the EOS options available., Comment: submitted to JOSS, Github page: https://github.com/ChunHuangPhy/CompactObject , Documentation: https://chunhuangphy.github.io/CompactObject/

Isotope dependences of charge radii, i.e., isotope shifts, calculated by the Skyrme Hartree-Fock, the relativistic mean-field, and the relativistic Hartree-Fock calculations are compared against the experimental data of magic and semimagic nuclei. It is found that the tensor interaction plays a role in reproducing the ``kink'' behavior, irregularity of isotope shifts at the neutron magic number, in the relativistic Hartree-Fock approach. With several Skyrme models, it is found that the kink behavior can be reproduced with the spin-orbit interaction having nonzero isovector channel. The single-particle orbitals near the Fermi energy are crucial to determine the kink size. The effects of the symmetry energy and the pairing interaction are also discussed in relation to the kink behavior., Comment: 22 pages, 34 figures, 10 tables

In this paper, the content popularity prediction problem in fog radio access networks (F-RANs) is investigated. Based on clustered federated learning, we propose a novel mobility-aware popularity prediction policy, which integrates content popularities in terms of local users and mobile users. For local users, the content popularity is predicted by learning the hidden representations of local users and contents. Initial features of local users and contents are generated by incorporating neighbor information with self information. Then, dual-channel neural network (DCNN) model is introduced to learn the hidden representations by producing deep latent features from initial features. For mobile users, the content popularity is predicted via user preference learning. In order to distinguish regional variations of content popularity, clustered federated learning (CFL) is employed, which enables fog access points (F-APs) with similar regional types to benefit from one another and provides a more specialized DCNN model for each F-AP. Simulation results show that our proposed policy achieves significant performance improvement over the traditional policies., Comment: 6 pages, 3 figures. This paper has been accepted by IEEE ICC 2022

The rapid progress of photorealistic synthesis techniques has reached a critical point where the boundary between real and manipulated images starts to blur. Recently, a mega-scale deep face forgery dataset, ForgeryNet which comprised of 2.9 million images and 221,247 videos has been released. It is by far the largest publicly available in terms of data-scale, manipulations (7 image-level approaches, 8 video-level approaches), perturbations (36 independent and more mixed perturbations), and annotations (6.3 million classification labels, 2.9 million manipulated area annotations, and 221,247 temporal forgery segment labels). This paper reports methods and results in the ForgeryNet - Face Forgery Analysis Challenge 2021, which employs the ForgeryNet benchmark. The model evaluation is conducted offline on the private test set. A total of 186 participants registered for the competition, and 11 teams made valid submissions. We will analyze the top-ranked solutions and present some discussion on future work directions., Comment: Technical report. Challenge website: https://competitions.codalab.org/competitions/33386

The tensor-force effects on the evolution of the spin-orbit splittings in the neutron drops are investigated within the framework of the relativistic Hartree-Fock theory. For fair comparisons on the pure mean-field level, the results of the relativistic Brueckner-Hartree-Fock calculation with the Bonn A interaction are adopted as meta-data. Through a quantitative analysis, we certify that the $\pi$-pseudovector ($\pi$-PV) coupling affects the evolutionary trend through the tensor force embedded. The strength of the tensor force is explored by enlarging the strength $f_{\pi}$ of the $\pi$-PV coupling. It is found that weakening the density dependence of $f_{\pi}$ is slightly better than enlarging it with a factor. We thus provide a semiquantitative support for the \textit{renormalization persistency} of the tensor force within the framework of density functional theory. This will serve as an important guidance for the further development of the relativistic effective interactions with particular focus on the tensor force., Comment: 9 pages, 4 figures, 2 tables

WebFG 2020 is an international challenge hosted by Nanjing University of Science and Technology, University of Edinburgh, Nanjing University, The University of Adelaide, Waseda University, etc. This challenge mainly pays attention to the webly-supervised fine-grained recognition problem. In the literature, existing deep learning methods highly rely on large-scale and high-quality labeled training data, which poses a limitation to their practicability and scalability in real world applications. In particular, for fine-grained recognition, a visual task that requires professional knowledge for labeling, the cost of acquiring labeled training data is quite high. It causes extreme difficulties to obtain a large amount of high-quality training data. Therefore, utilizing free web data to train fine-grained recognition models has attracted increasing attentions from researchers in the fine-grained community. This challenge expects participants to develop webly-supervised fine-grained recognition methods, which leverages web images in training fine-grained recognition models to ease the extreme dependence of deep learning methods on large-scale manually labeled datasets and to enhance their practicability and scalability. In this technical report, we have pulled together the top WebFG 2020 solutions of total 54 competing teams, and discuss what methods worked best across the set of winning teams, and what surprisingly did not help., Comment: This is a technical report of the WebFG 2020 challenge (https://sites.google.com/view/webfg2020) associated with ACCV 2020

The evolution of the energy difference between the neutron states $1i_{13/2}$ and $1h_{9/2}$ in the $N=82$ isotones and that between the proton states $1h_{11/2}$ and $1g_{7/2}$ in the $Z = 50$ isotopes are investigated within the framework of the relativistic Hartree-Fock theory, using the density-dependent effective interactions PKA1 and PKO$i$ ($i = 1 $, $ 2 $, $ 3 $). By identifying the contributions of the tensor force, which is naturally induced via the Fock terms, we find that the tensor force plays crucial roles in the evolution of the shell structure. The strength of the tensor force is also explored. It is found that moderately increasing the coupling strength of pion-nucleon coupling, i.e., $f_{\pi}$, will significantly improve the description of the shell-structure evolution. In particular, reducing the density dependence of $f_{\pi}$ is shown to be more preferable, in comparison to enlarging $f_{\pi}$ with a factor. This is in consistence with the idea of "tensor renormalization persistency" and provides valuable guidance for the development of nuclear energy density functional in the relativistic framework., Comment: 15 pages, 7 figures, 3 tables

Visual tracking is typically solved as a discriminative learning problem that usually requires high-quality samples for online model adaptation. It is a critical and challenging problem to evaluate the training samples collected from previous predictions and employ sample selection by their quality to train the model. To tackle the above problem, we propose a joint discriminative learning scheme with the progressive multi-stage optimization policy of sample selection for robust visual tracking. The proposed scheme presents a novel time-weighted and detection-guided self-paced learning strategy for easy-to-hard sample selection, which is capable of tolerating relatively large intra-class variations while maintaining inter-class separability. Such a self-paced learning strategy is jointly optimized in conjunction with the discriminative tracking process, resulting in robust tracking results. Experiments on the benchmark datasets demonstrate the effectiveness of the proposed learning framework., Comment: accepted to IEEE Transactions on Cybernetics