Search

Deep learning-based methods have shown remarkable performance in single JPEG artifacts removal task. However, existing methods tend to degrade on double JPEG images, which are prevalent in real-world scenarios. To address this issue, we propose Offset-Aware Partition Transformer for double JPEG artifacts removal, termed as OAPT. We conduct an analysis of double JPEG compression that results in up to four patterns within each 8x8 block and design our model to cluster the similar patterns to remedy the difficulty of restoration. Our OAPT consists of two components: compression offset predictor and image reconstructor. Specifically, the predictor estimates pixel offsets between the first and second compression, which are then utilized to divide different patterns. The reconstructor is mainly based on several Hybrid Partition Attention Blocks (HPAB), combining vanilla window-based self-attention and sparse attention for clustered pattern features. Extensive experiments demonstrate that OAPT outperforms the state-of-the-art method by more than 0.16dB in double JPEG image restoration task. Moreover, without increasing any computation cost, the pattern clustering module in HPAB can serve as a plugin to enhance other transformer-based image restoration methods. The code will be available at https://github.com/QMoQ/OAPT.git ., Comment: 14 pages, 9 figures. Codes and models are available at https://github.com/QMoQ/OAPT.git

Offline meta-reinforcement learning (meta-RL) methods, which adapt to unseen target tasks with prior experience, are essential in robot control tasks. Current methods typically utilize task contexts and skills as prior experience, where task contexts are related to the information within each task and skills represent a set of temporally extended actions for solving subtasks. However, these methods still suffer from limited performance when adapting to unseen target tasks, mainly because the learned prior experience lacks generalization, i.e., they are unable to extract effective prior experience from meta-training tasks by exploration and learning of continuous latent spaces. We propose a framework called decoupled meta-reinforcement learning (DCMRL), which (1) contrastively restricts the learning of task contexts through pulling in similar task contexts within the same task and pushing away different task contexts of different tasks, and (2) utilizes a Gaussian quantization variational autoencoder (GQ-VAE) for clustering the Gaussian distributions of the task contexts and skills respectively, and decoupling the exploration and learning processes of their spaces. These cluster centers which serve as representative and discrete distributions of task context and skill are stored in task context codebook and skill codebook, respectively. DCMRL can acquire generalizable prior experience and achieve effective adaptation to unseen target tasks during the meta-testing phase. Experiments in the navigation and robot manipulation continuous control tasks show that DCMRL is more effective than previous meta-RL methods with more generalizable prior experience., Comment: Accepted by AAAI 2024 (this version includes appendix)

We present an exploratory $N_f=2$ lattice QCD study of $\psi_2(3823)\to \gamma \chi_{c1}$ at a pion mass $m_{\pi}\approx 350$~MeV. The related two-point and three-piont functions are calculated using the distillation method. The electromagnetic multipole form factor $\hat{V}(0)=2.083(11)$ for $J/\psi\to\gamma \eta_c$ is consistent with previous lattice results, the form factors $\hat{E}_1(0)$, $\hat{M}_2(0)$ and $\hat{E}_3(0)$ for $\Gamma(\chi_{c2}\to\gamma J/\psi)$ have the same hierarchy as that derived from experiments and the predicted decay width $\Gamma(\chi_{c2}\to\gamma J/\psi)=368(5)~\text{keV}$ is in excellent agreement with the PDG value $374(10)~\text{keV}$ and previous lattice QCD results in the quenched approximation. The same strategy is applied to the study of the process $\psi_2(3823)\to \gamma \chi_{c1}$ and the partial decay width is predicted to be $337(27)~\text{keV}$. According to the BESIII constraints on the $\psi_2(3823)$ decay channels and some phenomenological results, we estimate the total width $\Gamma(\psi_2(3823))=520(100)~\text{keV}$.

Accurate segmentation of colonoscopic polyps is considered a fundamental step in medical image analysis and surgical interventions. Many recent studies have made improvements based on the encoder-decoder framework, which can effectively segment diverse polyps. Such improvements mainly aim to enhance local features by using global features and applying attention methods. However, relying only on the global information of the final encoder block can result in losing local regional features in the intermediate layer. In addition, determining the edges between benign regions and polyps could be a challenging task. To address the aforementioned issues, we propose a novel separated edge-guidance transformer (SegT) network that aims to build an effective polyp segmentation model. A transformer encoder that learns a more robust representation than existing CNN-based approaches was specifically applied. To determine the precise segmentation of polyps, we utilize a separated edge-guidance module consisting of separator and edge-guidance blocks. The separator block is a two-stream operator to highlight edges between the background and foreground, whereas the edge-guidance block lies behind both streams to strengthen the understanding of the edge. Lastly, an innovative cascade fusion module was used and fused the refined multi-level features. To evaluate the effectiveness of SegT, we conducted experiments with five challenging public datasets, and the proposed model achieved state-of-the-art performance.

Accurate understanding and prediction of human behaviors are critical prerequisites for autonomous vehicles, especially in highly dynamic and interactive scenarios such as intersections in dense urban areas. In this work, we aim at identifying crossing pedestrians and predicting their future trajectories. To achieve these goals, we not only need the context information of road geometry and other traffic participants but also need fine-grained information of the human pose, motion and activity, which can be inferred from human keypoints. In this paper, we propose a novel multi-task learning framework for pedestrian crossing action recognition and trajectory prediction, which utilizes 3D human keypoints extracted from raw sensor data to capture rich information on human pose and activity. Moreover, we propose to apply two auxiliary tasks and contrastive learning to enable auxiliary supervisions to improve the learned keypoints representation, which further enhances the performance of major tasks. We validate our approach on a large-scale in-house dataset, as well as a public benchmark dataset, and show that our approach achieves state-of-the-art performance on a wide range of evaluation metrics. The effectiveness of each model component is validated in a detailed ablation study., Comment: ICRA 2023

The celebrated anti-de Sitter/conformal field theory (AdS/CFT) correspondence1-3, also known as the gravity/gauge duality, posits a dual relationship between the quantum gravity in an AdS spacetime and the CFT defined on its lower-dimensional boundary. This correspondence not only offers profound insights into the enigmatic nature of quantum gravity but also shows mighty power in addressing strongly-correlated systems. However, despite its importance in contemporary physics, the AdS/CFT correspondence remains a conjecture, and further experimental investigation is highly sought after. Here, we present the first experimental exploration of this conjecture by testing its core corollary: the leading-order effects of a strongly-coupled CFT can be exactly described by a weakly-coupled classical field in a higher-dimensional AdS spacetime. Through measuring the bulk entanglement entropy (BEE) and boundary-boundary correlation function (BBCF) of scalar fields in both conventional type-I and previously overlooked type-II hyperbolic lattices, serving as the discretized regularizations of spatial geometries of pure AdS2+1 spacetime and AdS2+1 black hole, respectively, we experimentally confirm that BEE exhibits a logarithmic scaling with the subsystem size, following the Ryu-Takayanagi (RT) formula, while BBCF showcases an exponential law dependence on the boundary separation, the scaling dimension of which conforms to the Klebanov-Witten (KW) relation, both of which align remarkably well with the established CFT outcomes. This heuristic experimental effort opens a new avenue for in-depth investigations on the gravity/gauge duality and extensive exploration of quantum-gravity-inspired phenomena in classical systems.

In this work we study a more restricted class of quasi-topological gravity theories where the higher curvature terms have no contribution to the equation of motion on general static spherically symmetric metric where $g_{tt} g_{rr} \ne \mathrm{constant}$. We construct such theories up to quintic order in Riemann tensor and observe an important property of these theories: the higher order term in the Lagrangian vanishes identically when evaluated on the most general non-stationary spherically symmetric metric ansatz. This not only signals the higher terms could only have non-trivial effects when considering perturbations, but also makes the theories quasi-topological on a much wider range of metrics. As an example of the holographic effects of such theories, we consider a general Einstein-scalar theory and calculate it's holographic shear viscosity., Comment: 10 + 11 pages and appendices, one supplementary Wolfram Language file; v2: Revised, accepted by JHEP

We present the first theoretical prediction of the production rate of $1^{-+}$ light hybrid meson $\eta_1$ in $J/\psi$ radiative decays. In the $N_f=2$ lattice QCD formalism with the pion mass $m_\pi\approx 350$ MeV, the related electromagnetic multipole form factors are extracted from the three-point functions that involve necessarily quark annihilation diagrams, which are calculated through the distillation method. The partial width of $J/\psi\to \gamma \eta_1$ is determined to be $2.29(77)~\mathrm{eV}$ at the $\eta_1$ mass $m_{\eta_1}=2.23(4)$ GeV. If $\eta_1$ corresponds to the recently observed $\eta_1(1855)$ in the process $J/\psi\to \gamma\eta_1(1855)\to \gamma \eta\eta'$ by BESIII, then the branching fraction $\mathrm{Br}(J/\psi\to \gamma\eta_1(1855))$ is estimated to be $6.2(2.2)\times 10^{-5}$, which implies $\mathrm{Br}(\eta_1(1855)\to \eta\eta')\sim 4.3\%$., Comment: v2: Published in prd

The large radiative production rate for pseudoscalar mesons in the $J/\psi$ radiative decay remains elusive. We present the first lattice QCD calculation of partial decay width of $J/\psi$ radiatively decaying into $\eta_{(2)}$, the $\mathrm{SU(2)}$ flavor singlet pseudoscalar meson, which confirms QCD $\mathrm{U_A(1)}$ anomaly enhancement to the coupling of gluons with flavor singlet pseudoscalar mesons. The lattice simulation is carried out using $N_f=2$ lattice QCD gauge configurations at the pion mass $m_{\pi} \approx 350$ MeV. In particular, the distillation method has been utilized to calculate light quark loops. The results are reported here with the mass $m_{\eta_{(2)}}= 718(8)$ MeV and the decay width $\Gamma(J/\psi\to\gamma \eta_{(2)})=0.385(45)$ keV. By assuming the dominance of $\mathrm{U_A(1)}$ anomaly and flavor singlet-octet mixing angle $\theta=-24.5^\circ$, the production rates for the physical $\eta$ and $\eta'$ in $J/\psi$ radiative decay are predicted to be $1.15(14)\times 10^{-3}$ and $4.49(53)\times 10^{-3}$, respectively, which agree well with the experimental measurement data. Our study manifests the potential of lattice QCD studies on the light hadron production in $J/\psi$ radiative decays., Comment: 6 pages, 4 figures, 2 tables

We perform the first lattice study on the mixing of the isoscalar pseudoscalar meson $\eta$ and the pseudoscalar glueball $G$ in the $N_f=2$ QCD at the pion mass $m_\pi\approx 350$ MeV. The $\eta$ mass is determined to be $m_\eta=714(6)(16)$ MeV. Through the Witten-Veneziano relation, this value can be matched to a mass value of $\sim 981$ MeV for the $\mathrm{SU(3)}$ counterpart of $\eta$. Based on a large gauge ensemble, the $\eta-G$ mixing energy and the mixing angle are determined to be $|x|=107(15)(2)$ MeV and $|\theta|=3.46(46)^\circ$ from the $\eta-G$ correlators that are calculated using the distillation method. We conclude that the $\eta-G$ mixing is tiny and the topology induced interaction contributes most of $\eta$ mass owing to the QCD $\mathrm{U_A(1)}$ anomaly., Comment: 13 pages, 12 figures

We study glueballs on two $N_f=2+1$ RBC/UKQCD gauge ensembles with physical quark masses at two lattice spacings. The statistical uncertainties of the glueball correlation functions are considerably reduced through the cluster decomposition error reduction (CDER) method. The Bethe-Salpeter wave functions are obtained for the scalar, tensor and pseudoscalar glueballs by using spatially extended glueball operators defined through the gauge potential $A_\mu(x)$ in the Coulomb gauge. These wave functions show similar features of non-relativistic two-gluon systems, and they are used to optimize the signals of the related correlation functions at the early time regions. Consequently, the ground state masses can be extracted precisely. To the extent that the excited state contamination is not important, our calculation gives glueball masses at the physical pion mass for the first time.

In this work, we generate gauge configurations with $N_f=2$ dynamical charm quarks on anisotropic lattices. The mass shift of $1S$ and $1P$ charmonia owing to the charm quark annihilation effect can be investigated directly in a manner of unitary theory. The distillation method is adopted to treat the charm quark annihilation diagrams at a very precise level. For $1S$ charmonia, the charm quark annihilation effect almost does not change the $J/\psi$ mass, but lifts the $\eta_c$ mass by approximately 3-4 MeV. For $1P$ charmonia, this effect results in positive mass shifts of approximately 1 MeV for $\chi_{c1}$ and $h_c$, but decreases the $\chi_{c2}$ mass by approximately 3 MeV. We have not obtain a reliable result for the mass shift of $\chi_{c0}$. In addition, it is observed that the spin averaged mass of the spin-triplet $1P$ charmonia is in a good agreement with the $h_c$, as expected by the non-relativistic quark model and measured by experiments.

The K-means algorithm is arguably the most popular data clustering method, commonly applied to processed datasets in some "feature spaces", as is in spectral clustering. Highly sensitive to initializations, however, K-means encounters a scalability bottleneck with respect to the number of clusters K as this number grows in big data applications. In this work, we promote a closely related model called K-indicators model and construct an efficient, semi-convex-relaxation algorithm that requires no randomized initializations. We present extensive empirical results to show advantages of the new algorithm when K is large. In particular, using the new algorithm to start the K-means algorithm, without any replication, can significantly outperform the standard K-means with a large number of currently state-of-the-art random replications.

Compressed sensing in MRI enables high subsampling factors while maintaining diagnostic image quality. This technique enables shortened scan durations and/or improved image resolution. Further, compressed sensing can increase the diagnostic information and value from each scan performed. Overall, compressed sensing has significant clinical impact in improving the diagnostic quality and patient experience for imaging exams. However, a number of challenges exist when moving compressed sensing from research to the clinic. These challenges include hand-crafted image priors, sensitive tuning parameters, and long reconstruction times. Data-driven learning provides a solution to address these challenges. As a result, compressed sensing can have greater clinical impact. In this tutorial, we will review the compressed sensing formulation and outline steps needed to transform this formulation to a deep learning framework. Supplementary open source code in python will be used to demonstrate this approach with open databases. Further, we will discuss considerations in applying data-driven compressed sensing in the clinical setting., Comment: Submitted to the Special Issue on Computational MRI: Compressed Sensing and Beyond in the IEEE Signal Processing Magazine