Showing 201–250 of 767 results for author: Peng, C

Few-Shot Class-Incremental Learning from an Open-Set Perspective

Authors: Can Peng, Kun Zhao, Tianren Wang, Meng Li, Brian C. Lovell

Abstract: The continual appearance of new objects in the visual world poses considerable challenges for current deep learning methods in real-world deployments. The challenge of new task learning is often exacerbated by the scarcity of data for the new categories due to rarity or cost. Here we explore the important task of Few-Shot Class-Incremental Learning (FSCIL) and its extreme data scarcity condition o… ▽ More The continual appearance of new objects in the visual world poses considerable challenges for current deep learning methods in real-world deployments. The challenge of new task learning is often exacerbated by the scarcity of data for the new categories due to rarity or cost. Here we explore the important task of Few-Shot Class-Incremental Learning (FSCIL) and its extreme data scarcity condition of one-shot. An ideal FSCIL model needs to perform well on all classes, regardless of their presentation order or paucity of data. It also needs to be robust to open-set real-world conditions and be easily adapted to the new tasks that always arise in the field. In this paper, we first reevaluate the current task setting and propose a more comprehensive and practical setting for the FSCIL task. Then, inspired by the similarity of the goals for FSCIL and modern face recognition systems, we propose our method -- Augmented Angular Loss Incremental Classification or ALICE. In ALICE, instead of the commonly used cross-entropy loss, we propose to use the angular penalty loss to obtain well-clustered features. As the obtained features not only need to be compactly clustered but also diverse enough to maintain generalization for future incremental classes, we further discuss how class augmentation, data augmentation, and data balancing affect classification performance. Experiments on benchmark datasets, including CIFAR100, miniImageNet, and CUB200, demonstrate the improved performance of ALICE over the state-of-the-art FSCIL methods. △ Less

Submitted 30 July, 2022; originally announced August 2022.

Comments: Accepted to ECCV 2022

Experimental Simulation of Larger Quantum Circuits with Fewer Superconducting Qubits

Authors: Chong Ying, Bin Cheng, Youwei Zhao, He-Liang Huang, Yu-Ning Zhang, Ming Gong, Yulin Wu, Shiyu Wang, Futian Liang, ** Lin, Yu Xu, Hui Deng, Hao Rong, Cheng-Zhi Peng, Man-Hong Yung, Xiaobo Zhu, Jian-Wei Pan

Abstract: Although near-term quantum computing devices are still limited by the quantity and quality of qubits in the so-called NISQ era, quantum computational advantage has been experimentally demonstrated. Moreover, hybrid architectures of quantum and classical computing have become the main paradigm for exhibiting NISQ applications, where low-depth quantum circuits are repeatedly applied. In order to fur… ▽ More Although near-term quantum computing devices are still limited by the quantity and quality of qubits in the so-called NISQ era, quantum computational advantage has been experimentally demonstrated. Moreover, hybrid architectures of quantum and classical computing have become the main paradigm for exhibiting NISQ applications, where low-depth quantum circuits are repeatedly applied. In order to further scale up the problem size solvable by the NISQ devices, it is also possible to reduce the number of physical qubits by "cutting" the quantum circuit into different pieces. In this work, we experimentally demonstrated a circuit-cutting method for simulating quantum circuits involving many logical qubits, using only a few physical superconducting qubits. By exploiting the symmetry of linear-cluster states, we can estimate the effectiveness of circuit-cutting for simulating up to 33-qubit linear-cluster states, using at most 4 physical qubits for each subcircuit. Specifically, for the 12-qubit linear-cluster state, we found that the experimental fidelity bound can reach as much as 0.734, which is about 19\% higher than a direct simulation {on the same} 12-qubit superconducting processor. Our results indicate that circuit-cutting represents a feasible approach of simulating quantum circuits using much fewer qubits, while achieving a much higher circuit fidelity. △ Less

Submitted 1 March, 2023; v1 submitted 28 July, 2022; originally announced July 2022.

JDRec: Practical Actor-Critic Framework for Online Combinatorial Recommender System

Authors: Xin Zhao, Zhiwei Fang, Yuchen Guo, Jie He, Wenlong Chen, Chang** Peng

Abstract: A combinatorial recommender (CR) system feeds a list of items to a user at a time in the result page, in which the user behavior is affected by both contextual information and items. The CR is formulated as a combinatorial optimization problem with the objective of maximizing the recommendation reward of the whole list. Despite its importance, it is still a challenge to build a practical CR system… ▽ More A combinatorial recommender (CR) system feeds a list of items to a user at a time in the result page, in which the user behavior is affected by both contextual information and items. The CR is formulated as a combinatorial optimization problem with the objective of maximizing the recommendation reward of the whole list. Despite its importance, it is still a challenge to build a practical CR system, due to the efficiency, dynamics, personalization requirement in online environment. In particular, we tear the problem into two sub-problems, list generation and list evaluation. Novel and practical model architectures are designed for these sub-problems aiming at jointly optimizing effectiveness and efficiency. In order to adapt to online case, a bootstrap algorithm forming an actor-critic reinforcement framework is given to explore better recommendation mode in long-term user interaction. Offline and online experiment results demonstrate the efficacy of proposed JDRec framework. JDRec has been applied in online JD recommendation, improving click through rate by 2.6% and synthetical value for the platform by 5.03%. We will publish the large-scale dataset used in this study to contribute to the research community. △ Less

Submitted 27 July, 2022; originally announced July 2022.

Comments: 9 pages (7+2), 5 figures, AAAI Templete

TransFA: Transformer-based Representation for Face Attribute Evaluation

Authors: Decheng Liu, Weijie He, Chunlei Peng, Nannan Wang, Jie Li, Xinbo Gao

Abstract: Face attribute evaluation plays an important role in video surveillance and face analysis. Although methods based on convolution neural networks have made great progress, they inevitably only deal with one local neighborhood with convolutions at a time. Besides, existing methods mostly regard face attribute evaluation as the individual multi-label classification task, ignoring the inherent relatio… ▽ More Face attribute evaluation plays an important role in video surveillance and face analysis. Although methods based on convolution neural networks have made great progress, they inevitably only deal with one local neighborhood with convolutions at a time. Besides, existing methods mostly regard face attribute evaluation as the individual multi-label classification task, ignoring the inherent relationship between semantic attributes and face identity information. In this paper, we propose a novel \textbf{trans}former-based representation for \textbf{f}ace \textbf{a}ttribute evaluation method (\textbf{TransFA}), which could effectively enhance the attribute discriminative representation learning in the context of attention mechanism. The multiple branches transformer is employed to explore the inter-correlation between different attributes in similar semantic regions for attribute feature learning. Specially, the hierarchical identity-constraint attribute loss is designed to train the end-to-end architecture, which could further integrate face identity discriminative information to boost performance. Experimental results on multiple face attribute benchmarks demonstrate that the proposed TransFA achieves superior performances compared with state-of-the-art methods. △ Less

Submitted 12 July, 2022; originally announced July 2022.

Determining the Proton's Gluonic Gravitational Form Factors

Authors: B. Duran, Z. -E. Meziani, S. Joosten, M. K. Jones, S. Prasad, C. Peng, W. Armstrong, H. Atac, E. Chudakov, H. Bhatt, D. Bhetuwal, M. Boer, A. Camsonne, J. -P. Chen, M. M. Dalton, N. Deokar, M. Diefenthaler, J. Dunne, L. El Fassi, E. Fuchey, H. Gao, D. Gaskell, O. Hansen, F. Hauenstein, D. Higinbotham , et al. (30 additional authors not shown)

Abstract: The proton is one of the main building blocks of all visible matter in the universe. Among its intrinsic properties are its electric charge, mass, and spin. These emerge from the complex dynamics of its fundamental constituents, quarks and gluons, described by the theory of quantum chromodynamics (QCD). Using electron scattering, its electric charge and spin, shared among the quark constituents, h… ▽ More The proton is one of the main building blocks of all visible matter in the universe. Among its intrinsic properties are its electric charge, mass, and spin. These emerge from the complex dynamics of its fundamental constituents, quarks and gluons, described by the theory of quantum chromodynamics (QCD). Using electron scattering, its electric charge and spin, shared among the quark constituents, have been the topic of active investigation. An example is the novel precision measurement of the proton's electric charge radius. In contrast, little is known about the proton's inner mass density, dominated by the energy carried by the gluons, which are hard to access through electron scattering since gluons carry no electromagnetic charge. Here, we chose to probe this gluonic gravitational density using a small color dipole, the $J/ψ$ particle, through its threshold photoproduction. From our data, we determined, for the first time, the proton's gluonic gravitational form factors. We used a variety of models and determined, in all cases, a mass radius that is notably smaller than the electric charge radius. In some cases, the determined radius, although model dependent, is in excellent agreement with first-principle predictions from lattice QCD. This work paves the way for a deeper understanding of the salient role of gluons in providing gravitational mass to visible matter. △ Less

Submitted 7 February, 2023; v1 submitted 11 July, 2022; originally announced July 2022.

Comments: Accepted for publication

Journal ref: Nature 615, 813-816 (2023)

Spatial-Temporal Frequency Forgery Clue for Video Forgery Detection in VIS and NIR Scenario

Authors: Yukai Wang, Chunlei Peng, Decheng Liu, Nannan Wang, Xinbo Gao

Abstract: In recent years, with the rapid development of face editing and generation, more and more fake videos are circulating on social media, which has caused extreme public concerns. Existing face forgery detection methods based on frequency domain find that the GAN forged images have obvious grid-like visual artifacts in the frequency spectrum compared to the real images. But for synthesized videos, th… ▽ More In recent years, with the rapid development of face editing and generation, more and more fake videos are circulating on social media, which has caused extreme public concerns. Existing face forgery detection methods based on frequency domain find that the GAN forged images have obvious grid-like visual artifacts in the frequency spectrum compared to the real images. But for synthesized videos, these methods only confine to single frame and pay little attention to the most discriminative part and temporal frequency clue among different frames. To take full advantage of the rich information in video sequences, this paper performs video forgery detection on both spatial and temporal frequency domains and proposes a Discrete Cosine Transform-based Forgery Clue Augmentation Network (FCAN-DCT) to achieve a more comprehensive spatial-temporal feature representation. FCAN-DCT consists of a backbone network and two branches: Compact Feature Extraction (CFE) module and Frequency Temporal Attention (FTA) module. We conduct thorough experimental assessments on two visible light (VIS) based datasets WildDeepfake and Celeb-DF (v2), and our self-built video forgery dataset DeepfakeNIR, which is the first video forgery dataset on near-infrared modality. The experimental results demonstrate the effectiveness of our method on detecting forgery videos in both VIS and NIR scenarios. △ Less

Submitted 5 July, 2022; originally announced July 2022.

Potential energy surface and formation of superheavy nuclei with the Skyrme energy-density functional

Authors: Cheng Peng, Zhao-Qing Feng

Abstract: Within the framework of Skyrme energy-density functional theory, the nucleus-nucleus potential is calculated and potential energy surface is obtained with different effective forces for accurately estimating the formation cross sections of superheavy nuclei in massive fusion reactions. The width and height of the potential pocket are influenced by the Skyrme effective forces SkM, SkM$^{\ast}$, SkP… ▽ More Within the framework of Skyrme energy-density functional theory, the nucleus-nucleus potential is calculated and potential energy surface is obtained with different effective forces for accurately estimating the formation cross sections of superheavy nuclei in massive fusion reactions. The width and height of the potential pocket are influenced by the Skyrme effective forces SkM, SkM$^{\ast}$, SkP, SIII, Ska and SLy4, which correspond to the different equation of state for the isospin symmetry nuclear matter. It is found that the nucleus-nucleus potential is associated with the collision orientation and Skyrme parameters. More repulsive nuclear potential is pronounced with increasing the incompressible modulus of nuclear matter. The available data in the fusion-evaporation reaction of $^{48}$Ca+$^{238}$U are nicely reproduced with the SkM$^{\ast}$ parameter by implementing into the dinuclear system model. △ Less

Submitted 27 June, 2022; originally announced June 2022.

Comments: 9 pages, 6 figures, 1 table

Journal ref: Commun. Theor. Phys. 74 (2022) 055302

Bridging quantum many-body scar and quantum integrability in Ising chains with transverse and longitudinal fields

Authors: Cheng Peng, Xiaoling Cui

Abstract: Quantum many-body scar (QMBS) and quantum integrability(QI) have been recognized as two distinct mechanisms for the breakdown of eigenstate thermalization hypothesis(ETH) in an isolated system. In this work, we reveal a smooth route to connect these two ETH-breaking mechanisms in the Ising chain with transverse and longitudinal fields. Specifically, starting from an initial Ising anti-ferromagneti… ▽ More Quantum many-body scar (QMBS) and quantum integrability(QI) have been recognized as two distinct mechanisms for the breakdown of eigenstate thermalization hypothesis(ETH) in an isolated system. In this work, we reveal a smooth route to connect these two ETH-breaking mechanisms in the Ising chain with transverse and longitudinal fields. Specifically, starting from an initial Ising anti-ferromagnetic state, we find that the dynamical system undergoes a smooth non-thermal crossover from QMBS to QI by changing the Ising coupling($J$) and longitudinal field($h$) simultaneously while kee** their ratio fixed, which corresponds to the Rydberg Hamiltonian with an arbitrary nearest-neighbor repulsion. Deviating from this ratio, we further identify a continuous thermalization trajectory in ($h,J$) plane that is exactly given by the Ising transition line, signifying an intimate relation between thermalization and quantum critical point. Finally, we map out a completely different dynamical phase diagram starting from an initial ferromagnetic state, where the thermalization is shown to be equally facilitated by the resonant spin-flip at special ratios of $J$ and $h$. By bridging QMBS and QI in Ising chains, our results demonstrate the breakdown of ETH in much broader physical settings, which also suggest an alternative way to characterize quantum phase transition via thermalization in non-equilibrium dynamics. △ Less

Submitted 13 December, 2022; v1 submitted 22 June, 2022; originally announced June 2022.

Comments: 11 pages, 13 figures; accepted version by PRB

Journal ref: Phys. Rev. B 106, 214311 (2022)

A Novel Long-term Iterative Mining Scheme for Video Salient Object Detection

Authors: Chenglizhao Chen, Hengsen Wang, Yuming Fang, Chong Peng

Abstract: The existing state-of-the-art (SOTA) video salient object detection (VSOD) models have widely followed short-term methodology, which dynamically determines the balance between spatial and temporal saliency fusion by solely considering the current consecutive limited frames. However, the short-term methodology has one critical limitation, which conflicts with the real mechanism of our visual system… ▽ More The existing state-of-the-art (SOTA) video salient object detection (VSOD) models have widely followed short-term methodology, which dynamically determines the balance between spatial and temporal saliency fusion by solely considering the current consecutive limited frames. However, the short-term methodology has one critical limitation, which conflicts with the real mechanism of our visual system -- a typical long-term methodology. As a result, failure cases keep showing up in the results of the current SOTA models, and the short-term methodology becomes the major technical bottleneck. To solve this problem, this paper proposes a novel VSOD approach, which performs VSOD in a complete long-term way. Our approach converts the sequential VSOD, a sequential task, to a data mining problem, i.e., decomposing the input video sequence to object proposals in advance and then mining salient object proposals as much as possible in an easy-to-hard way. Since all object proposals are simultaneously available, the proposed approach is a complete long-term approach, which can alleviate some difficulties rooted in conventional short-term approaches. In addition, we devised an online updating scheme that can grasp the most representative and trustworthy pattern profile of the salient objects, outputting framewise saliency maps with rich details and smoothing both spatially and temporally. The proposed approach outperforms almost all SOTA models on five widely used benchmark datasets. △ Less

Submitted 20 June, 2022; originally announced June 2022.

qrpca: A Package for Fast Principal Component Analysis with GPU Acceleration

Authors: Rafael S. de Souza, Xu Quanfeng, Shiyin Shen, Chen Peng, Zihao Mu

Abstract: We present qrpca, a fast and scalable QR-decomposition principal component analysis package. The software, written in both R and python languages, makes use of torch for internal matrix computations, and enables GPU acceleration, when available. qrpca provides similar functionalities to prcomp (R) and sklearn (python) packages respectively. A benchmark test shows that qrpca can achieve computation… ▽ More We present qrpca, a fast and scalable QR-decomposition principal component analysis package. The software, written in both R and python languages, makes use of torch for internal matrix computations, and enables GPU acceleration, when available. qrpca provides similar functionalities to prcomp (R) and sklearn (python) packages respectively. A benchmark test shows that qrpca can achieve computational speeds 10-20 $\times$ faster for large dimensional matrices than default implementations, and is at least twice as fast for a standard decomposition of spectral data cubes. The qrpca source code is made freely available to the community. △ Less

Submitted 6 September, 2022; v1 submitted 14 June, 2022; originally announced June 2022.

Journal ref: Astronomy and Computing, 41, 100633 (2022)

Enhanced superconductivity by near-neighbor attraction in the doped Hubbard model

Authors: Cheng Peng, Yao Wang, Jiajia Wen, Young Lee, Thomas Devereaux, Hong-Chen Jiang

Abstract: Recent experiment has unveiled an anomalously strong electron-electron attraction in one-dimensional copper-oxide chain Ba$_{2-x}$Sr$_x$CuO$_{3+δ}$. While the near-neighbor electron attraction $V$ in the one-dimensional extended Hubbard chain has been examined recently, its effect in the Hubbard model beyond the one-dimensional chain remains unclear. We report a density-matrix renormalization grou… ▽ More Recent experiment has unveiled an anomalously strong electron-electron attraction in one-dimensional copper-oxide chain Ba$_{2-x}$Sr$_x$CuO$_{3+δ}$. While the near-neighbor electron attraction $V$ in the one-dimensional extended Hubbard chain has been examined recently, its effect in the Hubbard model beyond the one-dimensional chain remains unclear. We report a density-matrix renormalization group study of the extended Hubbard model on long four-leg cylinders on the square lattice. We find that the near-neighbor electron attraction $V$ can notably enhance the long-distance superconducting correlations while simultaneously suppressing the charge-density-wave correlations. Specifically, for a modestly strong electron attraction, the superconducting correlations become dominant over the CDW correlations with a Luttinger exponent $K_{sc}\sim 1$ and strong divergent superconducting susceptibility. Our results provide a promising way to realize long-range superconductivity in the doped Hubbard model in two dimensions. The relevance of our numerical results to cuprate materials is also discussed. △ Less

Submitted 7 June, 2022; originally announced June 2022.

Comments: 6 pages, 5 figures

Black holes Entangled by Radiation

Authors: Yuxuan Liu, Zhuo-Yu Xian, Cheng Peng, Yi Ling

Abstract: We construct three models to describe the scenario where two eternal black holes are separated by a flat space, and can eventually be entangled by exchanging radiations. In the doubly holographic setup, we compute the entanglement entropy and the mutual information among the subsystems and obtain the dynamic phase structure of the entanglement. The formation of entanglement between the two black h… ▽ More We construct three models to describe the scenario where two eternal black holes are separated by a flat space, and can eventually be entangled by exchanging radiations. In the doubly holographic setup, we compute the entanglement entropy and the mutual information among the subsystems and obtain the dynamic phase structure of the entanglement. The formation of entanglement between the two black holes is delayed by the space where the radiations must travel through. Finally, if the two black holes exchange sufficient Hawking modes, the final state is characterized by a connected entanglement wedge; otherwise, the final entanglement wedge contains two separated islands. In the former case, the entanglement wedge of the two black holes forms at the time scale of the size of the flat space between them. While in both cases, unitarity of the evolution is preserved. When the sizes of two black holes are not equal, we observe a loss of entanglement between the smaller black hole and the radiation at late times. In the field theory side, we consider two Sachdev-Ye-Kitaev (SYK) clusters coupled to a Majorana chain, which resemble two black holes connected by a radiation region. We numerically compute the same entanglement measures, and obtain similar phase structures as the bulk results. In general, a time delay of the entanglement between the two SYK clusters is found in cases with a long Majorana chain. In particular, when the two SYK clusters are different in size, similar entanglement loss between the smaller SYK cluster and the Majorana chain is observed. Finally, we investigate a chain model composed of EPR clusters with particle exchanges between neighboring clusters, and reproduce the features of entanglement observed in the other models. △ Less

Submitted 8 November, 2022; v1 submitted 29 May, 2022; originally announced May 2022.

Comments: 38 pages, 16 figures; V2: references added, minor revision; V3: references added, minor revision; typo revision;

Portable ground stations for space-to-ground quantum key distribution

Authors: Ji-Gang Ren, Maimaiti Abulizi, Hai-Lin Yong, Juan Yin, Xue-Jiao Li, Yuan Jiang, Wei-Yang Wang, Hua-Jian Xue, Yu-He Chen, Biao **, Ya-Yun Yin, Zhou-Yu Tu, Xiao-Juan Zhu, Shuang-Qiang Zhao, Feng-Zhi Li, Sheng-Kai Liao, Wen-Qi Cai, Wei-Yue Liu, Yuan Cao, Fei Zhou, Li Li, Nai-Le Liu, Qiang Zhang, Yu-Ao Chen, Cheng-Zhi Peng , et al. (1 additional authors not shown)

Abstract: Quantum key distribution (QKD) uses the fundamental principles of quantum mechanics to share unconditionally secure keys between distant users. Previous works based on the quantum science satellite "Micius" have initially demonstrated the feasibility of a global QKD network. However, the practical applications of space-based QKD still face many technical problems, such as the huge size and weight… ▽ More Quantum key distribution (QKD) uses the fundamental principles of quantum mechanics to share unconditionally secure keys between distant users. Previous works based on the quantum science satellite "Micius" have initially demonstrated the feasibility of a global QKD network. However, the practical applications of space-based QKD still face many technical problems, such as the huge size and weight of ground stations required to receive quantum signals. Here, we report space-to-ground QKD demonstrations based on portable receiving ground stations. The weight of the portable ground station is less than 100 kg, the space required is less than 1 m$^{3}$ and the installation time requires no more than 12 hours, all of the weight, required space and deployment time are about two orders of magnitude lower than those for the previous systems. Moreover, the equipment is easy to handle and can be placed on the roof of buildings in a metropolis. Secure keys have been successfully generated from the "Micius" satellite to these portable ground stations at six different places in China, and an average final secure key length is around 50 kb can be obtained during one passage. Our results pave the way for, and greatly accelerate the practical application of, space-based QKD. △ Less

Submitted 27 May, 2022; originally announced May 2022.

Giant enhancement of third-harmonic generation in graphene-metal heterostructures

Authors: Irati Alonso Calafell, Lee A. Rozema, David Alcaraz Iranzo, Alessandro Trenti, Joel D. Cox, Avinash Kumar, Hlib Bieliaiev, Sebastian Nanot, Cheng Peng, Dmitri K. Efetov, ** Yong Hong, **g Kong, Dirk R. Englund, F. Javier García de Abajo, Frank H. L. Koppens, Philp Walther

Abstract: Nonlinear nanophotonics leverages engineered nanostructures to funnel light into small volumes and intensify nonlinear optical processes with spectral and spatial control. Due to its intrinsically large and electrically tunable nonlinear optical response, graphene is an especially promising nanomaterial for nonlinear optoelectronic applications. Here we report on exceptionally strong optical nonli… ▽ More Nonlinear nanophotonics leverages engineered nanostructures to funnel light into small volumes and intensify nonlinear optical processes with spectral and spatial control. Due to its intrinsically large and electrically tunable nonlinear optical response, graphene is an especially promising nanomaterial for nonlinear optoelectronic applications. Here we report on exceptionally strong optical nonlinearities in graphene-insulator-metal heterostructures, demonstrating an enhancement by three orders of magnitude in the third-harmonic signal compared to bare graphene. Furthermore, by increasing the graphene Fermi energy through an external gate voltage, we find that graphene plasmons mediate the optical nonlinearity and modify the third-harmonic signal. Our findings show that graphene-insulator-metal is a promising heterostructure for optically-controlled and electrically-tunable nano-optoelectronic components. △ Less

Submitted 25 May, 2022; originally announced May 2022.

Journal ref: Nature Nanotechnology 16, 318-324, (2021)

6G Network AI Architecture for Everyone-Centric Customized Services

Authors: Yang Yang, Mulei Ma, Hequan Wu, Quan Yu, ** Zhang, Xiaohu You, Jianjun Wu, Chenghui Peng, Tak-Shing Peter Yum, Sherman Shen, Hamid Aghvami, Geoffrey Y Li, Jiangzhou Wang, Guangyi Liu, Peng Gao, Xiongyan Tang, Chang Cao, John Thompson, Kat-Kit Wong, Shanzhi Chen, Merouane Debbah, Schahram Dustdar, Frank Eliassen, Tao Chen, Xiangyang Duan , et al. (29 additional authors not shown)

Abstract: Mobile communication standards were developed for enhancing transmission and network performance by using more radio resources and improving spectrum and energy efficiency. How to effectively address diverse user requirements and guarantee everyone's Quality of Experience (QoE) remains an open problem. The Sixth Generation (6G) mobile systems will solve this problem by utilizing heterogenous netwo… ▽ More Mobile communication standards were developed for enhancing transmission and network performance by using more radio resources and improving spectrum and energy efficiency. How to effectively address diverse user requirements and guarantee everyone's Quality of Experience (QoE) remains an open problem. The Sixth Generation (6G) mobile systems will solve this problem by utilizing heterogenous network resources and pervasive intelligence to support everyone-centric customized services anywhere and anytime. In this article, we first coin the concept of Service Requirement Zone (SRZ) on the user side to characterize and visualize the integrated service requirements and preferences of specific tasks of individual users. On the system side, we further introduce the concept of User Satisfaction Ratio (USR) to evaluate the system's overall service ability of satisfying a variety of tasks with different SRZs. Then, we propose a network Artificial Intelligence (AI) architecture with integrated network resources and pervasive AI capabilities for supporting customized services with guaranteed QoEs. Finally, extensive simulations show that the proposed network AI architecture can consistently offer a higher USR performance than the cloud AI and edge AI architectures with respect to different task scheduling algorithms, random service requirements, and dynamic network conditions. △ Less

Submitted 6 December, 2023; v1 submitted 19 May, 2022; originally announced May 2022.

Comments: The current version has partial Insufficient completion, so we would like to withdraw it. We hope you agree, thank you

NDGGNET-A Node Independent Gate based Graph Neural Networks

Authors: Ye Tang, Xuesong Yang, Xinrui Liu, Xiwei Zhao, Zhangang Lin, Chang** Peng

Abstract: Graph Neural Networks (GNNs) is an architecture for structural data, and has been adopted in a mass of tasks and achieved fabulous results, such as link prediction, node classification, graph classification and so on. Generally, for a certain node in a given graph, a traditional GNN layer can be regarded as an aggregation from one-hop neighbors, thus a set of stacked layers are able to fetch and u… ▽ More Graph Neural Networks (GNNs) is an architecture for structural data, and has been adopted in a mass of tasks and achieved fabulous results, such as link prediction, node classification, graph classification and so on. Generally, for a certain node in a given graph, a traditional GNN layer can be regarded as an aggregation from one-hop neighbors, thus a set of stacked layers are able to fetch and update node status within multi-hops. For nodes with sparse connectivity, it is difficult to obtain enough information through a single GNN layer as not only there are only few nodes directly connected to them but also can not propagate the high-order neighbor information. However, as the number of layer increases, the GNN model is prone to over-smooth for nodes with the dense connectivity, which resulting in the decrease of accuracy. To tackle this issue, in this thesis, we define a novel framework that allows the normal GNN model to accommodate more layers. Specifically, a node-degree based gate is employed to adjust weight of layers dynamically, that try to enhance the information aggregation ability and reduce the probability of over-smoothing. Experimental results show that our proposed model can effectively increase the model depth and perform well on several datasets. △ Less

Submitted 11 May, 2022; originally announced May 2022.

ACM Class: F.4.1; I.2.4

Half-Wormholes and Ensemble Averages

Authors: Cheng Peng, Jia Tian, Yingyu Yang

Abstract: We study "half-wormhole-like" saddle point contributions to spectral correlators in a variety of ensemble average models, including various statistical models, generalized 0d SYK models, 1d Brownian SYK models and an extension of it. In statistical ensemble models, where more general distributions of the random variables could be studied in great details, we find the accuracy of the previously pro… ▽ More We study "half-wormhole-like" saddle point contributions to spectral correlators in a variety of ensemble average models, including various statistical models, generalized 0d SYK models, 1d Brownian SYK models and an extension of it. In statistical ensemble models, where more general distributions of the random variables could be studied in great details, we find the accuracy of the previously proposed approximation for the half-wormholes could be improved when the distribution of the random variables deviate significantly from Gaussian distributions. We propose a modified approximation scheme of the half-wormhole contributions that also work well in these more general theories. In various generalized 0d SYK models we identify new half-wormhole-like saddle point contributions. In the 0d SYK model and 1d Brownian SYK model, apart from the wormhole and half-wormhole saddles, we find new non-trivial saddles in the spectral correlators that would potentially give contributions of the same order as the trivial self-averaging saddles. However after a careful Lefschetz-thimble analysis we show that these non-trivial saddles should not be included. We also clarify the difference between "linked half-wormholes" and "unlinked half-wormholes" in some models. △ Less

Submitted 6 May, 2022; v1 submitted 2 May, 2022; originally announced May 2022.

Comments: 87 pages, version 2, refs added and minor changes

Gating-adapted Wavelet Multiresolution Analysis for Exposure Sequence Modeling in CTR prediction

Authors: Xiaoxiao Xu, Zhiwei Fang, Qian Yu, Ruoran Huang, \\Chaosheng Fan, Yong Li, Yang He, Chang** Peng, Zhangang Lin, **g** Shao

Abstract: The exposure sequence is being actively studied for user interest modeling in Click-Through Rate (CTR) prediction. However, the existing methods for exposure sequence modeling bring extensive computational burden and neglect noise problems, resulting in an excessively latency and the limited performance in online recommenders. In this paper, we propose to address the high latency and noise problem… ▽ More The exposure sequence is being actively studied for user interest modeling in Click-Through Rate (CTR) prediction. However, the existing methods for exposure sequence modeling bring extensive computational burden and neglect noise problems, resulting in an excessively latency and the limited performance in online recommenders. In this paper, we propose to address the high latency and noise problems via Gating-adapted wavelet multiresolution analysis (Gama), which can effectively denoise the extremely long exposure sequence and adaptively capture the implied multi-dimension user interest with linear computational complexity. This is the first attempt to integrate non-parametric multiresolution analysis technique into deep neural networks to model user exposure sequence. Extensive experiments on large scale benchmark dataset and real production dataset confirm the effectiveness of Gama for exposure sequence modeling, especially in cold-start scenarios. Benefited from its low latency and high effecitveness, Gama has been deployed in our real large-scale industrial recommender, successfully serving over hundreds of millions users. △ Less

Submitted 29 April, 2022; originally announced April 2022.

Comments: In proceedings of the 45th International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR '22), July 11--15, 2022, Madrid, Spain. 5 pages

Log-based Sparse Nonnegative Matrix Factorization for Data Representation

Authors: Chong Peng, Yiqun Zhang, Yongyong Chen, Zhao Kang, Chenglizhao Chen, Qiang Cheng

Abstract: Nonnegative matrix factorization (NMF) has been widely studied in recent years due to its effectiveness in representing nonnegative data with parts-based representations. For NMF, a sparser solution implies better parts-based representation.However, current NMF methods do not always generate sparse solutions.In this paper, we propose a new NMF method with log-norm imposed on the factor matrices to… ▽ More Nonnegative matrix factorization (NMF) has been widely studied in recent years due to its effectiveness in representing nonnegative data with parts-based representations. For NMF, a sparser solution implies better parts-based representation.However, current NMF methods do not always generate sparse solutions.In this paper, we propose a new NMF method with log-norm imposed on the factor matrices to enhance the sparseness.Moreover, we propose a novel column-wisely sparse norm, named $\ell_{2,\log}$-(pseudo) norm to enhance the robustness of the proposed method.The $\ell_{2,\log}$-(pseudo) norm is invariant, continuous, and differentiable.For the $\ell_{2,\log}$ regularized shrinkage problem, we derive a closed-form solution, which can be used for other general problems.Efficient multiplicative updating rules are developed for the optimization, which theoretically guarantees the convergence of the objective value sequence.Extensive experimental results confirm the effectiveness of the proposed method, as well as the enhanced sparseness and robustness. △ Less

Submitted 22 April, 2022; originally announced April 2022.

Matrix Entanglement

Authors: Vaibhav Gautam, Masanori Hanada, Antal Jevicki, Cheng Peng

Abstract: In gauge/gravity duality, matrix degrees of freedom on the gauge theory side play important roles for the emergent geometry. In this paper, we discuss how the entanglement on the gravity side can be described as the entanglement between matrix degrees of freedom. Our approach, which we call 'matrix entanglement', is different from 'target-space entanglement' proposed and discussed recently by seve… ▽ More In gauge/gravity duality, matrix degrees of freedom on the gauge theory side play important roles for the emergent geometry. In this paper, we discuss how the entanglement on the gravity side can be described as the entanglement between matrix degrees of freedom. Our approach, which we call 'matrix entanglement', is different from 'target-space entanglement' proposed and discussed recently by several groups. We consider several classes of quantum states to which our approach can play important roles. When applied to fuzzy sphere, matrix entanglement can be used to define the usual spatial entanglement in two-brane or five-brane world-volume theory nonperturbatively in a regularized setup. Another application is to a small black hole in AdS5*S5 that can evaporate without being attached to a heat bath, for which our approach suggests a gauge theory origin of the Page curve. The confined degrees of freedom in the partially-deconfined states play the important roles. △ Less

Submitted 13 May, 2022; v1 submitted 13 April, 2022; originally announced April 2022.

Comments: 37 pages, 11 figures, references added

Report number: DMUS-MP-22/03

On the Adaptation to Concept Drift for CTR Prediction

Authors: Congcong Liu, Yuejiang Li, Fei Teng, Xiwei Zhao, Chang** Peng, Zhangang Lin, **ghe Hu, **g** Shao

Abstract: Click-through rate (CTR) prediction is a crucial task in web search, recommender systems, and online advertisement displaying. In practical application, CTR models often serve with high-speed user-generated data streams, whose underlying distribution rapidly changing over time. The concept drift problem inevitably exists in those streaming data, which can lead to performance degradation due to the… ▽ More Click-through rate (CTR) prediction is a crucial task in web search, recommender systems, and online advertisement displaying. In practical application, CTR models often serve with high-speed user-generated data streams, whose underlying distribution rapidly changing over time. The concept drift problem inevitably exists in those streaming data, which can lead to performance degradation due to the timeliness issue. To ensure model freshness, incremental learning has been widely adopted in real-world production systems. However, it is hard for the incremental update to achieve the balance of the CTR models between the adaptability to capture the fast-changing trends and generalization ability to retain common knowledge. In this paper, we propose adaptive mixture of experts (AdaMoE), a new framework to alleviate the concept drift problem by statistical weighting policy in the data stream of CTR prediction. The extensive offline experiments on both benchmark and a real-world industrial dataset, as well as an online A/B testing show that our AdaMoE significantly outperforms all incremental learning frameworks considered. △ Less

Submitted 22 February, 2023; v1 submitted 1 April, 2022; originally announced April 2022.

IA-GCN: Interactive Graph Convolutional Network for Recommendation

Authors: Yinan Zhang, Pei Wang, Congcong Liu, Xiwei Zhao, Hao Qi, Jie He, Junsheng **, Chang** Peng, Zhangang Lin, **g** Shao

Abstract: Recently, Graph Convolutional Network (GCN) has become a novel state-of-art for Collaborative Filtering (CF) based Recommender Systems (RS). It is a common practice to learn informative user and item representations by performing embedding propagation on a user-item bipartite graph, and then provide the users with personalized item suggestions based on the representations. Despite effectiveness, e… ▽ More Recently, Graph Convolutional Network (GCN) has become a novel state-of-art for Collaborative Filtering (CF) based Recommender Systems (RS). It is a common practice to learn informative user and item representations by performing embedding propagation on a user-item bipartite graph, and then provide the users with personalized item suggestions based on the representations. Despite effectiveness, existing algorithms neglect precious interactive features between user-item pairs in the embedding process. When predicting a user's preference for different items, they still aggregate the user tree in the same way, without emphasizing target-related information in the user neighborhood. Such a uniform aggregation scheme easily leads to suboptimal user and item representations, limiting the model expressiveness to some extent. In this work, we address this problem by building bilateral interactive guidance between each user-item pair and proposing a new model named IA-GCN (short for InterActive GCN). Specifically, when learning the user representation from its neighborhood, we assign higher attention weights to those neighbors similar to the target item. Correspondingly, when learning the item representation, we pay more attention to those neighbors resembling the target user. This leads to interactive and interpretable features, effectively distilling target-specific information through each graph convolutional operation. Our model is built on top of LightGCN, a state-of-the-art GCN model for CF, and can be combined with various GCN-based CF architectures in an end-to-end fashion. Extensive experiments on three benchmark datasets demonstrate the effectiveness and robustness of IA-GCN. △ Less

Submitted 6 May, 2024; v1 submitted 7 April, 2022; originally announced April 2022.

Rethinking Position Bias Modeling with Knowledge Distillation for CTR Prediction

Authors: Congcong Liu, Yuejiang Li, Jian Zhu, Xiwei Zhao, Chang** Peng, Zhangang Lin, **g** Shao

Abstract: Click-through rate (CTR) Prediction is of great importance in real-world online ads systems. One challenge for the CTR prediction task is to capture the real interest of users from their clicked items, which is inherently biased by presented positions of items, i.e., more front positions tend to obtain higher CTR values. A popular line of existing works focuses on explicitly estimating position bi… ▽ More Click-through rate (CTR) Prediction is of great importance in real-world online ads systems. One challenge for the CTR prediction task is to capture the real interest of users from their clicked items, which is inherently biased by presented positions of items, i.e., more front positions tend to obtain higher CTR values. A popular line of existing works focuses on explicitly estimating position bias by result randomization which is expensive and inefficient, or by inverse propensity weighting (IPW) which relies heavily on the quality of the propensity estimation. Another common solution is modeling position as features during offline training and simply adopting fixed value or dropout tricks when serving. However, training-inference inconsistency can lead to sub-optimal performance. Furthermore, post-click information such as position values is informative while less exploited in CTR prediction. This work proposes a simple yet efficient knowledge distillation framework to alleviate the impact of position bias and leverage position information to improve CTR prediction. We demonstrate the performance of our proposed method on a real-world production dataset and online A/B tests, achieving significant improvements over competing baseline models. The proposed method has been deployed in the real world online ads systems, serving main traffic on one of the world's largest e-commercial platforms. △ Less

Submitted 1 April, 2022; originally announced April 2022.

Interactive Multi-scale Fusion of 2D and 3D Features for Multi-object Tracking

Authors: Guangming Wang, Chensheng Peng, **peng Zhang, Hesheng Wang

Abstract: Multiple object tracking (MOT) is a significant task in achieving autonomous driving. Traditional works attempt to complete this task, either based on point clouds (PC) collected by LiDAR, or based on images captured from cameras. However, relying on one single sensor is not robust enough, because it might fail during the tracking process. On the other hand, feature fusion from multiple modalities… ▽ More Multiple object tracking (MOT) is a significant task in achieving autonomous driving. Traditional works attempt to complete this task, either based on point clouds (PC) collected by LiDAR, or based on images captured from cameras. However, relying on one single sensor is not robust enough, because it might fail during the tracking process. On the other hand, feature fusion from multiple modalities contributes to the improvement of accuracy. As a result, new techniques based on different sensors integrating features from multiple modalities are being developed. Texture information from RGB cameras and 3D structure information from Lidar have respective advantages under different circumstances. However, it's not easy to achieve effective feature fusion because of completely distinct information modalities. Previous fusion methods usually fuse the top-level features after the backbones extract the features from different modalities. In this paper, we first introduce PointNet++ to obtain multi-scale deep representations of point cloud to make it adaptive to our proposed Interactive Feature Fusion between multi-scale features of images and point clouds. Specifically, through multi-scale interactive query and fusion between pixel-level and point-level features, our method, can obtain more distinguishing features to improve the performance of multiple object tracking. Besides, we explore the effectiveness of pre-training on each single modality and fine-tuning on the fusion-based model. The experimental results demonstrate that our method can achieve good performance on the KITTI benchmark and outperform other approaches without using multi-scale feature fusion. Moreover, the ablation studies indicates the effectiveness of multi-scale feature fusion and pre-training on single modality. △ Less

Submitted 30 March, 2022; originally announced March 2022.

Comments: 9 pages, 5 figures, under review

A quantum-inspired tensor network method for constrained combinatorial optimization problems

Authors: Tianyi Hao, Xuxin Huang, Chun**g Jia, Cheng Peng

Abstract: Combinatorial optimization is of general interest for both theoretical study and real-world applications. Fast-develo** quantum algorithms provide a different perspective on solving combinatorial optimization problems. In this paper, we propose a quantum-inspired tensor-network-based algorithm for general locally constrained combinatorial optimization problems. Our algorithm constructs a Hamilto… ▽ More Combinatorial optimization is of general interest for both theoretical study and real-world applications. Fast-develo** quantum algorithms provide a different perspective on solving combinatorial optimization problems. In this paper, we propose a quantum-inspired tensor-network-based algorithm for general locally constrained combinatorial optimization problems. Our algorithm constructs a Hamiltonian for the problem of interest, effectively map** it to a quantum problem, then encodes the constraints directly into a tensor network state and solves the optimal solution by evolving the system to the ground state of the Hamiltonian. We demonstrate our algorithm with the open-pit mining problem, which results in a quadratic asymptotic time complexity. Our numerical results show the effectiveness of this construction and potential applications in further studies for general combinatorial optimization problems. △ Less

Submitted 5 September, 2022; v1 submitted 29 March, 2022; originally announced March 2022.

Journal ref: Frontiers in Physics, Volume 10, Article 906590 (2022)

113 km Free-Space Time-Frequency Dissemination at the 19th Decimal Instability

Authors: Qi Shen, Jian-Yu Guan, Ji-Gang Ren, Ting Zeng, Lei Hou, Min Li, Yuan Cao, **-Jian Han, Meng-Zhe Lian, Yan-Wei Chen, Xin-Xin Peng, Shao-Mao Wang, Dan-Yang Zhu, Xi-** Shi, Zheng-Guo Wang, Ye Li, Wei-Yue Liu, Ge-Sheng Pan, Yong Wang, Zhao-Hui Li, **-Cai Wu, Yan-Yan Zhang, Fa-Xi Chen, Chao-Yang Lu, Sheng-Kai Liao , et al. (6 additional authors not shown)

Abstract: Optical clock networks play important roles in various fields, such as precise navigation, redefinition of "second" unit, and gravitational tests. To establish a global-scale optical clock network, it is essential to disseminate time and frequency with a stability of $10^{-19}$ over a long-distance free-space link. However, such attempts were limited to dozens of kilometers in mirror-folded config… ▽ More Optical clock networks play important roles in various fields, such as precise navigation, redefinition of "second" unit, and gravitational tests. To establish a global-scale optical clock network, it is essential to disseminate time and frequency with a stability of $10^{-19}$ over a long-distance free-space link. However, such attempts were limited to dozens of kilometers in mirror-folded configuration. Here, we take a crucial step toward future satellite-based time-frequency disseminations. By develo** the key technologies, including high-power frequency combs, high-stability and high-efficiency optical transceiver systems, and efficient linear optical sampling, we demonstrate free-space time-frequency dissemination over two independent links with femtosecond time deviation, $3\times10^{-19}$ at 10,000 s residual instability and $1.6\times10^{-20}\pm 4.3\times10^{-19}$ offset. This level of the stability retains for an increased channel loss up to 89 dB. Our work can not only be directly used in ground-based application, but also firmly laid the groundwork for future satellite time-frequency dissemination. △ Less

Submitted 22 March, 2022; originally announced March 2022.

Comments: 27 pages, 13 figures, 2 tables

Journal ref: Nature 610, 661 (2022)

Tailoring Dirac fermions by in-situ tunable high-order moire pattern in graphene-monolayer xenon heterostructure

Authors: Chunlong Wu, Qiang Wan, Cao Peng, Shangkun Mo, Renzhe Li, Keming Zhao, Yan** Guo, Shengjun Yuan, Fengcheng Wu, Chendong Zhang, Nan Xu

Abstract: A variety of novel quantum phases have been achieved in twist bilayer graphene (tBLG) and other moire superlattices recently, including correlated insulators, superconductivity, magnetism, and topological states. These phenomena are very sensitive to the moire superlattices, which can hardly be changed rapidly or intensely. Here, we report the experimental realization of a high-order moire pattern… ▽ More A variety of novel quantum phases have been achieved in twist bilayer graphene (tBLG) and other moire superlattices recently, including correlated insulators, superconductivity, magnetism, and topological states. These phenomena are very sensitive to the moire superlattices, which can hardly be changed rapidly or intensely. Here, we report the experimental realization of a high-order moire pattern (a high-order interference pattern) in graphene-monolayer xenon heterostructure (G/mXe), with moire period in-situ tuned from few nanometers to infinity by changing the lattice constant of Xe through different annealing temperatures and pressures. We use angle-resolved photoemission spectroscopy to directly observe that replicas of graphene Dirac cone emerge and move close to each other in momentum-space as moire pattern continuously expands in real-space. When the moire period approaches infinity, the replicas finally overlap with each other and an energy gap is observed at the Dirac point induced by intervalley coupling, which is a manifestation of Kekule distortion. We construct a continuum moire Hamiltonian, which can explain the experimental results well. The form of moire Hamiltonian in G/mXe is similar to that in tBLG, and moire band with narrow bandwidth is predicted in G/mXe. However, the moire Hamiltonian couples Dirac fermions from different valleys in G/mXe, instead of ones from different layers in tBLG. Our work demonstrates a novel platform to study the continuous evolution of moire pattern and its modulation effect on electronic structure, and provides an unprecedented approach for tailoring Dirac fermions with tunable intervalley coupling. △ Less

Submitted 17 March, 2022; originally announced March 2022.

Comments: 17 pages, 4 figures, supplementary materials available from the authors, submitted Feb. 2022

Journal ref: Phys. Rev. Lett. 129, 176402 (2022)

Monolithic Active Pixel Sensors on CMOS technologies

Authors: Nicole Apadula, Whitney Armstrong, James Brau, Martin Breidenbach, R. Caputo, Gabriella Carinii, Alberto Collu, Marcel Demarteau, Grzegorz Deptuch, Angelo Dragone, Gabriele Giacomini, Carl Grace, Norman Graf, Leo Greiner, Ryan Herbst, Gunther Haller, Manoj Jadhav, Sylvester Joosten, Christopher J. Kenney, C. Kierans, Jihee Kim, Thomas Markiewicz, Yuan Mei, Jessica Metcalfe, Zein-Eddine Meziani , et al. (15 additional authors not shown)

Abstract: Collider detectors have taken advantage of the resolution and accuracy of silicon detectors for at least four decades. Future colliders will need large areas of silicon sensors for low mass trackers and sampling calorimetry. Monolithic Active Pixel Sensors (MAPS), in which Si diodes and readout circuitry are combined in the same pixels, and can be fabricated in some of standard CMOS processes, are… ▽ More Collider detectors have taken advantage of the resolution and accuracy of silicon detectors for at least four decades. Future colliders will need large areas of silicon sensors for low mass trackers and sampling calorimetry. Monolithic Active Pixel Sensors (MAPS), in which Si diodes and readout circuitry are combined in the same pixels, and can be fabricated in some of standard CMOS processes, are a promising technology for high-granularity and light detectors. In this paper we review 1) the requirements on MAPS for trackers and electromagnetic calorimeters (ECal) at future colliders experiments, 2) the ongoing efforts towards dedicated MAPS for the Electron-Ion Collider (EIC) at BNL, for which the EIC Silicon Consortium was already instantiated, and 3) space-born applications for MeV $γ$-ray experiments with MAPS based trackers (AstroPix). △ Less

Submitted 28 March, 2022; v1 submitted 14 March, 2022; originally announced March 2022.

Comments: 25 pages, 18 figures, contribution to Snowmass 2021

First measurement of high-energy reactor antineutrinos at Daya Bay

Authors: Daya Bay collaboration, F. P. An, A. B. Balantekin, H. R. Band, M. Bishai, S. Blyth, G. F. Cao, J. Cao, J. F. Chang, Y. Chang, H. S. Chen, S. M. Chen, Y. Chen, Y. X. Chen, J. Cheng, Z. K. Cheng, J. J. Cherwinka, M. C. Chu, J. P. Cummings, O. Dalager, F. S. Deng, Y. Y. Ding, M. V. Diwan, T. Dohnal, J. Dove , et al. (162 additional authors not shown)

Abstract: This Letter reports the first measurement of high-energy reactor antineutrinos at Daya Bay, with nearly 9000 inverse beta decay candidates in the prompt energy region of 8-12~MeV observed over 1958 days of data collection. A multivariate analysis is used to separate 2500 signal events from background statistically. The hypothesis of no reactor antineutrinos with neutrino energy above 10~MeV is rej… ▽ More This Letter reports the first measurement of high-energy reactor antineutrinos at Daya Bay, with nearly 9000 inverse beta decay candidates in the prompt energy region of 8-12~MeV observed over 1958 days of data collection. A multivariate analysis is used to separate 2500 signal events from background statistically. The hypothesis of no reactor antineutrinos with neutrino energy above 10~MeV is rejected with a significance of 6.2 standard deviations. A 29\% antineutrino flux deficit in the prompt energy region of 8-11~MeV is observed compared to a recent model prediction. We provide the unfolded antineutrino spectrum above 7 MeV as a data-based reference for other experiments. This result provides the first direct observation of the production of antineutrinos from several high-$Q_β$ isotopes in commercial reactors. △ Less

Submitted 8 July, 2022; v1 submitted 13 March, 2022; originally announced March 2022.

Comments: 7 pages, 4 figures, accepted by Physical Review Letters

Journal ref: Phys. Rev. Lett. 129, 041801 (2022)

Towards performant and reliable undersampled MR reconstruction via diffusion model sampling

Authors: Cheng Peng, Pengfei Guo, S. Kevin Zhou, Vishal Patel, Rama Chellappa

Abstract: Magnetic Resonance (MR) image reconstruction from under-sampled acquisition promises faster scanning time. To this end, current State-of-The-Art (SoTA) approaches leverage deep neural networks and supervised training to learn a recovery model. While these approaches achieve impressive performances, the learned model can be fragile on unseen degradation, e.g. when given a different acceleration fac… ▽ More Magnetic Resonance (MR) image reconstruction from under-sampled acquisition promises faster scanning time. To this end, current State-of-The-Art (SoTA) approaches leverage deep neural networks and supervised training to learn a recovery model. While these approaches achieve impressive performances, the learned model can be fragile on unseen degradation, e.g. when given a different acceleration factor. These methods are also generally deterministic and provide a single solution to an ill-posed problem; as such, it can be difficult for practitioners to understand the reliability of the reconstruction. We introduce DiffuseRecon, a novel diffusion model-based MR reconstruction method. DiffuseRecon guides the generation process based on the observed signals and a pre-trained diffusion model, and does not require additional training on specific acceleration factors. DiffuseRecon is stochastic in nature and generates results from a distribution of fully-sampled MR images; as such, it allows us to explicitly visualize different potential reconstruction solutions. Lastly, DiffuseRecon proposes an accelerated, coarse-to-fine Monte-Carlo sampling scheme to approximate the most likely reconstruction candidate. The proposed DiffuseRecon achieves SoTA performances reconstructing from raw acquisition signals in fastMRI and SKM-TEA. Code will be open-sourced at www.github.com/cpeng93/DiffuseRecon. △ Less

Submitted 10 March, 2022; v1 submitted 7 March, 2022; originally announced March 2022.

Undersampled MRI Reconstruction with Side Information-Guided Normalisation

Authors: Xinwen Liu, **g Wang, Cheng Peng, Shekhar S. Chandra, Feng Liu, S. Kevin Zhou

Abstract: Magnetic resonance (MR) images exhibit various contrasts and appearances based on factors such as different acquisition protocols, views, manufacturers, scanning parameters, etc. This generally accessible appearance-related side information affects deep learning-based undersampled magnetic resonance imaging (MRI) reconstruction frameworks, but has been overlooked in the majority of current works.… ▽ More Magnetic resonance (MR) images exhibit various contrasts and appearances based on factors such as different acquisition protocols, views, manufacturers, scanning parameters, etc. This generally accessible appearance-related side information affects deep learning-based undersampled magnetic resonance imaging (MRI) reconstruction frameworks, but has been overlooked in the majority of current works. In this paper, we investigate the use of such side information as normalisation parameters in a convolutional neural network (CNN) to improve undersampled MRI reconstruction. Specifically, a Side Information-Guided Normalisation (SIGN) module, containing only few layers, is proposed to efficiently encode the side information and output the normalisation parameters. We examine the effectiveness of such a module on two popular reconstruction architectures, D5C5 and OUCR. The experimental results on both brain and knee images under various acceleration rates demonstrate that the proposed method improves on its corresponding baseline architectures with a significant margin. △ Less

Submitted 7 March, 2022; originally announced March 2022.

Topological unidirectional guided resonances emerged from interband coupling

Authors: Xuefan Yin, Takuya Inoue, Chao Peng, Susumu Noda

Abstract: Unidirectional guided resonances (UGRs) are optical modes in photonic crystal (PhC) slabs that radiate towards one side without the need for mirrors on the other, represented from a topological perspective by the merged points of paired, single-sided, half-integer topological charges. In this work, we report a mechanism to realize UGRs by tuning the interband coupling effect originating from up-do… ▽ More Unidirectional guided resonances (UGRs) are optical modes in photonic crystal (PhC) slabs that radiate towards one side without the need for mirrors on the other, represented from a topological perspective by the merged points of paired, single-sided, half-integer topological charges. In this work, we report a mechanism to realize UGRs by tuning the interband coupling effect originating from up-down symmetry breaking. We theoretically demonstrate that a type of polarization singularity, the circular-polarized states (CPs), emerge from trivial polarization fields owing to the hybridization of two unperturbed states. By tuning structural parameters, two half-charges carried by CPs evolve in momentum space and merge to create UGRs. Our findings show that UGRs are ubiquitous in PhC slabs, and can systematically be found from our method, thus paving the way to new possibilities of light manipulation. △ Less

Submitted 4 March, 2022; originally announced March 2022.

Physics-Informed Graph Learning

Authors: Ciyuan Peng, Feng Xia, Vidya Saikrishna, Huan Liu

Abstract: An expeditious development of graph learning in recent years has found innumerable applications in several diversified fields. Of the main associated challenges are the volume and complexity of graph data. The graph learning models suffer from the inability to efficiently learn graph information. In order to indemnify this inefficacy, physics-informed graph learning (PIGL) is emerging. PIGL incorp… ▽ More An expeditious development of graph learning in recent years has found innumerable applications in several diversified fields. Of the main associated challenges are the volume and complexity of graph data. The graph learning models suffer from the inability to efficiently learn graph information. In order to indemnify this inefficacy, physics-informed graph learning (PIGL) is emerging. PIGL incorporates physics rules while performing graph learning, which has enormous benefits. This paper presents a systematic review of PIGL methods. We begin with introducing a unified framework of graph learning models followed by examining existing PIGL methods in relation to the unified framework. We also discuss several future challenges for PIGL. This survey paper is expected to stimulate innovative research and development activities pertaining to PIGL. △ Less

Submitted 20 October, 2022; v1 submitted 22 February, 2022; originally announced February 2022.

Comments: 8 pages, 3 figures

MSC Class: 68T07; 68T30 ACM Class: I.2.6

Journal ref: 2022 IEEE International Conference on Data Mining Workshops (ICDMW)

Realization of fast all-microwave CZ gates with a tunable coupler

Authors: Shaowei Li, Dao** Fan, Ming Gong, Yangsen Ye, Xiawei Chen, Yulin Wu, Huijie Guan, Hui Deng, Hao Rong, He-Liang Huang, Chen Zha, Kai Yan, Shaojun Guo, Haoran Qian, Haibin Zhang, Fusheng Chen, Qingling Zhu, Youwei Zhao, Shiyu Wang, Chong Ying, Sirui Cao, Jiale Yu, Futian Liang, Yu Xu, ** Lin , et al. (7 additional authors not shown)

Abstract: The development of high-fidelity two-qubit quantum gates is essential for digital quantum computing. Here, we propose and realize an all-microwave parametric Controlled-Z (CZ) gates by coupling strength modulation in a superconducting Transmon qubit system with tunable couplers. After optimizing the design of the tunable coupler together with the control pulse numerically, we experimentally realiz… ▽ More The development of high-fidelity two-qubit quantum gates is essential for digital quantum computing. Here, we propose and realize an all-microwave parametric Controlled-Z (CZ) gates by coupling strength modulation in a superconducting Transmon qubit system with tunable couplers. After optimizing the design of the tunable coupler together with the control pulse numerically, we experimentally realized a 100 ns CZ gate with high fidelity of 99.38%$ \pm$0.34% and the control error being 0.1%. We note that our CZ gates are not affected by pulse distortion and do not need pulse correction, {providing a solution for the real-time pulse generation in a dynamic quantum feedback circuit}. With the expectation of utilizing our all-microwave control scheme to reduce the number of control lines through frequency multiplexing in the future, our scheme draws a blueprint for the high-integrable quantum hardware design. △ Less

Submitted 14 February, 2022; originally announced February 2022.

Journal ref: Chin. Phys. Lett.,39 (3): 030302 (2022)

Emergence of Crystalline Few-body Correlations in Mass-imbalanced Fermi Polarons

Authors: Rui** Liu, Cheng Peng, Xiaoling Cui

Abstract: Polarons can serve as an ideal platform to identify few-body correlations in tackling complex many-body problems. In this work, we reveal various crystalline few-body correlations smoothly emergent from the mass-imbalanced Fermi polarons in two dimensions. A unified variational approach up to three particle-hole excitations allows us to extract the dominant dimer, trimer or tetramer correlation in… ▽ More Polarons can serve as an ideal platform to identify few-body correlations in tackling complex many-body problems. In this work, we reveal various crystalline few-body correlations smoothly emergent from the mass-imbalanced Fermi polarons in two dimensions. A unified variational approach up to three particle-hole excitations allows us to extract the dominant dimer, trimer or tetramer correlation in a single framework. When the fermion-impurity mass ratio is beyond certain critical value, the Fermi polaron is found to undergo a smooth crossover, instead of a sharp transition, from the polaronic to trimer and tetramer regimes as increasing the fermion-impurity attraction. The emergent trimer and tetramer correlations result in the momentum-space crystallization of particle-hole excitations featuring a stable diagonal or triangular structure, as can be directly probed through the density-density correlation of majority fermions. Our results shed light on the intriguing quantum phases in the mass-imbalanced Fermi-Fermi mixtures beyond the pairing superfluid paradigm. △ Less

Submitted 18 July, 2022; v1 submitted 7 February, 2022; originally announced February 2022.

Comments: 11+5 pages, 5+4 figures; to appear in Cell Reports Physical Science

Journal ref: Cell Reports Physical Science 3, 100993 (2022)

Universal tetramer and pentamer in two-dimensional fermionic mixtures

Authors: Rui** Liu, Cheng Peng, Xiaoling Cui

Abstract: We study the emergence of universal tetramer and pentamer bound states in the two-dimensional $(N+1)$ system, which consists of $N$ identical heavy fermions interacting with a light atom. We show that the critical heavy-light mass ratio to support a ($3+1$) tetramer below the trimer threshold is $3.38$, and to support a ($4+1$) pentamer below the tetramer threshold is $5.14$. While these ground st… ▽ More We study the emergence of universal tetramer and pentamer bound states in the two-dimensional $(N+1)$ system, which consists of $N$ identical heavy fermions interacting with a light atom. We show that the critical heavy-light mass ratio to support a ($3+1$) tetramer below the trimer threshold is $3.38$, and to support a ($4+1$) pentamer below the tetramer threshold is $5.14$. While these ground state tetramer and pentamer are both with zero total angular momentum, they exhibit very different density distributions and correlations in momentum space, due to their distinct angular momentum decompositions in the dimer-fermion frame. These universal bound states can be accessible by a number of Fermi-Fermi mixtures now realized in cold atoms laboratories, which also suggest novel few-body correlations dominant in their corresponding many-body systems. △ Less

Submitted 18 July, 2022; v1 submitted 3 February, 2022; originally announced February 2022.

Comments: 6 pages, 3 figures; version to appear in PRL

Journal ref: Phys. Rev. Lett. 129, 073401 (2022)

Coupled power generators require stability buffers in addition to inertia

Authors: Gurupraanesh Raman, Gururaghav Raman, Jimmy Chih-Hsien Peng

Abstract: Increasing the inertia is widely considered to be the solution to resolving unstable interactions between coupled oscillators. In power grids, Virtual Synchronous Generators (VSGs) are proposed to compensate the reducing inertia as rotating synchronous generators are being phased out. Yet, modeling how VSGs and rotating generators simultaneously contribute energy and inertia, we surprisingly find… ▽ More Increasing the inertia is widely considered to be the solution to resolving unstable interactions between coupled oscillators. In power grids, Virtual Synchronous Generators (VSGs) are proposed to compensate the reducing inertia as rotating synchronous generators are being phased out. Yet, modeling how VSGs and rotating generators simultaneously contribute energy and inertia, we surprisingly find that instabilities of a small-signal nature could arise despite fairly high system inertia. Importantly, we show there exist both an optimal and a maximum number of such VSGs that can be safely supported, a previously unknown result directly useful for power utilities in long-term planning and prosumer contracting. Meanwhile, to resolve instabilities in the short term, we argue that the new market should include another commodity that we call stability storage, whereby -- analogous to energy storage buffering energy imbalances -- VSGs act as decentralized stability buffers. While demonstrating the effectiveness of this concept for a wide range of energy futures, we provide policymakers and utilities with a roadmap towards achieving a 100% renewable grid. △ Less

Submitted 27 January, 2022; originally announced January 2022.

Comments: 18 pages, 6 figures

Alleviating Cold-start Problem in CTR Prediction with A Variational Embedding Learning Framework

Authors: Xiaoxiao Xu, Chen Yang, Qian Yu, Zhiwei Fang, Jiaxing Wang, Chaosheng Fan, Yang He, Chang** Peng, Zhangang Lin, **g** Shao

Abstract: We propose a general Variational Embedding Learning Framework (VELF) for alleviating the severe cold-start problem in CTR prediction. VELF addresses the cold start problem via alleviating over-fits caused by data-sparsity in two ways: learning probabilistic embedding, and incorporating trainable and regularized priors which utilize the rich side information of cold start users and advertisements (… ▽ More We propose a general Variational Embedding Learning Framework (VELF) for alleviating the severe cold-start problem in CTR prediction. VELF addresses the cold start problem via alleviating over-fits caused by data-sparsity in two ways: learning probabilistic embedding, and incorporating trainable and regularized priors which utilize the rich side information of cold start users and advertisements (Ads). The two techniques are naturally integrated into a variational inference framework, forming an end-to-end training process. Abundant empirical tests on benchmark datasets well demonstrate the advantages of our proposed VELF. Besides, extended experiments confirmed that our parameterized and regularized priors provide more generalization capability than traditional fixed priors. △ Less

Submitted 17 January, 2022; originally announced January 2022.

Comments: In Proceedings of the Web Conference 2022 (WWW 2022), April 25-29, 2022, Lyon, France. 9 pages

Roadmap on Topological Photonics

Authors: Hannah Price, Yidong Chong, Alexander Khanikaev, Henning Schomerus, Lukas J. Maczewsky, Mark Kremer, Matthias Heinrich, Alexander Szameit, Oded Zilberberg, Yihao Yang, Baile Zhang, Andrea Alù, Ronny Thomale, Iacopo Carusotto, Philippe St-Jean, Alberto Amo, Avik Dutt, Luqi Yuan, Shanhui Fan, Xuefan Yin, Chao Peng, Tomoki Ozawa, Andrea Blanco-Redondo

Abstract: Topological photonics seeks to control the behaviour of the light through the design of protected topological modes in photonic structures. While this approach originated from studying the behaviour of electrons in solid-state materials, it has since blossomed into a field that is at the very forefront of the search for new topological types of matter. This can have real implications for future te… ▽ More Topological photonics seeks to control the behaviour of the light through the design of protected topological modes in photonic structures. While this approach originated from studying the behaviour of electrons in solid-state materials, it has since blossomed into a field that is at the very forefront of the search for new topological types of matter. This can have real implications for future technologies by harnessing the robustness of topological photonics for applications in photonics devices. This Roadmap surveys some of the main emerging areas of research within topological photonics, with a special attention to questions in fundamental science, which photonics is in an ideal position to address. Each section provides an overview of the current and future challenges within a part of the field, highlighting the most exciting opportunities for future research and developments. △ Less

Submitted 17 January, 2022; originally announced January 2022.

Comments: Invited Roadmap submission to Journal of Physics: Photonics

Quantum Neuronal Sensing of Quantum Many-Body States on a 61-Qubit Programmable Superconducting Processor

Authors: Ming Gong, He-Liang Huang, Shiyu Wang, Chu Guo, Shaowei Li, Yulin Wu, Qingling Zhu, Youwei Zhao, Shaojun Guo, Haoran Qian, Yangsen Ye, Chen Zha, Fusheng Chen, Chong Ying, Jiale Yu, Dao** Fan, Dachao Wu, Hong Su, Hui Deng, Hao Rong, Kaili Zhang, Sirui Cao, ** Lin, Yu Xu, Lihua Sun , et al. (11 additional authors not shown)

Abstract: Classifying many-body quantum states with distinct properties and phases of matter is one of the most fundamental tasks in quantum many-body physics. However, due to the exponential complexity that emerges from the enormous numbers of interacting particles, classifying large-scale quantum states has been extremely challenging for classical approaches. Here, we propose a new approach called quantum… ▽ More Classifying many-body quantum states with distinct properties and phases of matter is one of the most fundamental tasks in quantum many-body physics. However, due to the exponential complexity that emerges from the enormous numbers of interacting particles, classifying large-scale quantum states has been extremely challenging for classical approaches. Here, we propose a new approach called quantum neuronal sensing. Utilizing a 61 qubit superconducting quantum processor, we show that our scheme can efficiently classify two different types of many-body phenomena: namely the ergodic and localized phases of matter. Our quantum neuronal sensing process allows us to extract the necessary information coming from the statistical characteristics of the eigenspectrum to distinguish these phases of matter by measuring only one qubit. Our work demonstrates the feasibility and scalability of quantum neuronal sensing for near-term quantum processors and opens new avenues for exploring quantum many-body phenomena in larger-scale systems. △ Less

Submitted 20 November, 2022; v1 submitted 15 January, 2022; originally announced January 2022.

Comments: 7 pages, 3 figures in the main text, and 13 pages, 13 figures, and 1 table in supplementary materials

Journal ref: Science Bulletin, 68(9):906-912 (2023)

Deeply virtual Compton scattering cross section at high Bjorken $x_B$

Authors: F. Georges, M. N. H. Rashad, A. Stefanko, M. Dlamini, B. Karki, S. F. Ali, P-J. Lin, H-S Ko, N. Israel, D. Adikaram, Z. Ahmed, H. Albataineh, B. Aljawrneh, K. Allada, S. Allison, S. Alsalmi, D. Androic, K. Aniol, J. Annand, H. Atac, T. Averett, C. Ayerbe Gayoso, X. Bai, J. Bane, S. Barcus , et al. (137 additional authors not shown)

Abstract: We report high-precision measurements of the Deeply Virtual Compton Scattering (DVCS) cross section at high values of the Bjorken variable $x_B$. DVCS is sensitive to the Generalized Parton Distributions of the nucleon, which provide a three-dimensional description of its internal constituents. Using the exact analytic expression of the DVCS cross section for all possible polarization states of th… ▽ More We report high-precision measurements of the Deeply Virtual Compton Scattering (DVCS) cross section at high values of the Bjorken variable $x_B$. DVCS is sensitive to the Generalized Parton Distributions of the nucleon, which provide a three-dimensional description of its internal constituents. Using the exact analytic expression of the DVCS cross section for all possible polarization states of the initial and final electron and nucleon, and final state photon, we present the first experimental extraction of all four helicity-conserving Compton Form Factors (CFFs) of the nucleon as a function of $x_B$, while systematically including helicity flip amplitudes. In particular, the high accuracy of the present data demonstrates sensitivity to some very poorly known CFFs. △ Less

Submitted 10 January, 2022; originally announced January 2022.

Hyperspectral Image Denoising Using Non-convex Local Low-rank and Sparse Separation with Spatial-Spectral Total Variation Regularization

Authors: Chong Peng, Yang Liu, Yongyong Chen, Xinxin Wu, Andrew Cheng, Zhao Kang, Chenglizhao Chen, Qiang Cheng

Abstract: In this paper, we propose a novel nonconvex approach to robust principal component analysis for HSI denoising, which focuses on simultaneously develo** more accurate approximations to both rank and column-wise sparsity for the low-rank and sparse components, respectively. In particular, the new method adopts the log-determinant rank approximation and a novel $\ell_{2,\log}$ norm, to restrict the… ▽ More In this paper, we propose a novel nonconvex approach to robust principal component analysis for HSI denoising, which focuses on simultaneously develo** more accurate approximations to both rank and column-wise sparsity for the low-rank and sparse components, respectively. In particular, the new method adopts the log-determinant rank approximation and a novel $\ell_{2,\log}$ norm, to restrict the local low-rank or column-wisely sparse properties for the component matrices, respectively. For the $\ell_{2,\log}$-regularized shrinkage problem, we develop an efficient, closed-form solution, which is named $\ell_{2,\log}$-shrinkage operator. The new regularization and the corresponding operator can be generally used in other problems that require column-wise sparsity. Moreover, we impose the spatial-spectral total variation regularization in the log-based nonconvex RPCA model, which enhances the global piece-wise smoothness and spectral consistency from the spatial and spectral views in the recovered HSI. Extensive experiments on both simulated and real HSIs demonstrate the effectiveness of the proposed method in denoising HSIs. △ Less

Submitted 8 January, 2022; originally announced January 2022.

Effects of the momentum dependence of nuclear symmetry potential on pion observables in Sn + Sn collisions at 270 MeV/nucleon

Authors: Gao-Feng Wei, Xin Huang, Qi-Jun Zhi, Ai-Jun Dong, Chang-Gen Peng, Zheng-Wen Long

Abstract: Within a transport model, we study effects of the momentum dependence of nuclear symmetry potential on pion observables in central Sn + Sn collisions at 270 MeV/nucleon. To this end, a quantity $U_{sym}^{\infty}(ρ_{0})$, i.e., the value of nuclear symmetry potential at the saturation density $ρ_{0}$ and infinitely large nucleon momentum, is used to characterise the momentum dependence of nuclear s… ▽ More Within a transport model, we study effects of the momentum dependence of nuclear symmetry potential on pion observables in central Sn + Sn collisions at 270 MeV/nucleon. To this end, a quantity $U_{sym}^{\infty}(ρ_{0})$, i.e., the value of nuclear symmetry potential at the saturation density $ρ_{0}$ and infinitely large nucleon momentum, is used to characterise the momentum dependence of nuclear symmetry potential. It is shown that with a certain $L$ (i.e., slope of nuclear symmetry energy at $ρ_{0}$) the characteristic parameter $U_{sym}^{\infty}(ρ_{0})$ of symmetry potential affects significantly the production of $π^{-}$ and $π^{+}$ as well as their pion ratios. Moreover, through comparing the charged pion yields, pion ratios as well the spectral pion ratios of theoretical simulations for the reactions $^{108}$Sn + $^{112}$Sn and $^{132}$Sn + $^{124}$Sn with the corresponding data in S$π$RIT experiments, we find that our results favor a constraint on $U_{sym}^{\infty}(ρ_{0})$, i.e., $-160^{+18}_{-9}$~MeV, and the $L$ is also suggested within a range, i.e., $62.7<L<93.1$~MeV. In addition, it is shown that the pion observable of $^{197}$Au + $^{197}$Au collisions at 400~MeV/nucleon also supports the extracted value for $U_{sym}^{\infty}(ρ_{0})$. △ Less

Submitted 4 October, 2022; v1 submitted 27 December, 2021; originally announced December 2021.

Comments: 11 pages, 14 figures, add several references and several figures to interpret the results; original conclusion is not changed qualitatively, more stringent constraints on $U_{sym}^{\infty}(ρ_{0})$ are concluded quantitatively. Accepted for publication in Nuclear Science and Techniques

Journal ref: NUCL SCI TECH 33, 163 (2022)

Realization of an Error-Correcting Surface Code with Superconducting Qubits

Authors: Youwei Zhao, Yangsen Ye, He-Liang Huang, Yiming Zhang, Dachao Wu, Huijie Guan, Qingling Zhu, Zuolin Wei, Tan He, Sirui Cao, Fusheng Chen, Tung-Hsun Chung, Hui Deng, Dao** Fan, Ming Gong, Cheng Guo, Shaojun Guo, Lianchen Han, Na Li, Shaowei Li, Yuan Li, Futian Liang, ** Lin, Haoran Qian, Hao Rong , et al. (14 additional authors not shown)

Abstract: Quantum error correction is a critical technique for transitioning from noisy intermediate-scale quantum (NISQ) devices to fully fledged quantum computers. The surface code, which has a high threshold error rate, is the leading quantum error correction code for two-dimensional grid architecture. So far, the repeated error correction capability of the surface code has not been realized experimental… ▽ More Quantum error correction is a critical technique for transitioning from noisy intermediate-scale quantum (NISQ) devices to fully fledged quantum computers. The surface code, which has a high threshold error rate, is the leading quantum error correction code for two-dimensional grid architecture. So far, the repeated error correction capability of the surface code has not been realized experimentally. Here, we experimentally implement an error-correcting surface code, the distance-3 surface code which consists of 17 qubits, on the \textit{Zuchongzhi} 2.1 superconducting quantum processor. By executing several consecutive error correction cycles, the logical error can be significantly reduced after applying corrections, achieving the repeated error correction of surface code for the first time. This experiment represents a fully functional instance of an error-correcting surface code, providing a key step on the path towards scalable fault-tolerant quantum computing. △ Less

Submitted 29 January, 2022; v1 submitted 26 December, 2021; originally announced December 2021.

Journal ref: Phys. Rev. Lett. 129, 030501 (2022)

HyperSegNAS: Bridging One-Shot Neural Architecture Search with 3D Medical Image Segmentation using HyperNet

Authors: Cheng Peng, Andriy Myronenko, Ali Hatamizadeh, Vish Nath, Md Mahfuzur Rahman Siddiquee, Yufan He, Daguang Xu, Rama Chellappa, Dong Yang

Abstract: Semantic segmentation of 3D medical images is a challenging task due to the high variability of the shape and pattern of objects (such as organs or tumors). Given the recent success of deep learning in medical image segmentation, Neural Architecture Search (NAS) has been introduced to find high-performance 3D segmentation network architectures. However, because of the massive computational require… ▽ More Semantic segmentation of 3D medical images is a challenging task due to the high variability of the shape and pattern of objects (such as organs or tumors). Given the recent success of deep learning in medical image segmentation, Neural Architecture Search (NAS) has been introduced to find high-performance 3D segmentation network architectures. However, because of the massive computational requirements of 3D data and the discrete optimization nature of architecture search, previous NAS methods require a long search time or necessary continuous relaxation, and commonly lead to sub-optimal network architectures. While one-shot NAS can potentially address these disadvantages, its application in the segmentation domain has not been well studied in the expansive multi-scale multi-path search space. To enable one-shot NAS for medical image segmentation, our method, named HyperSegNAS, introduces a HyperNet to assist super-net training by incorporating architecture topology information. Such a HyperNet can be removed once the super-net is trained and introduces no overhead during architecture search. We show that HyperSegNAS yields better performing and more intuitive architectures compared to the previous state-of-the-art (SOTA) segmentation networks; furthermore, it can quickly and accurately find good architecture candidates under different computing constraints. Our method is evaluated on public datasets from the Medical Segmentation Decathlon (MSD) challenge, and achieves SOTA performances. △ Less

Submitted 24 March, 2022; v1 submitted 20 December, 2021; originally announced December 2021.

Disordered vector models: from higher spins to incipient strings

Authors: Chi-Ming Chang, Sean Colin-Ellerin, Cheng Peng, Mukund Rangamani

Abstract: We present a one-parameter family of large $N$ disordered models, with and without supersymmetry, in three spacetime dimensions. They interpolate from the critical large $N$ vector model dual to a classical higher spin theory, towards a theory with a classical string dual. We analyze the spectrum and OPE data of the theories. While the supersymmetric model is always well-behaved the non-supersymme… ▽ More We present a one-parameter family of large $N$ disordered models, with and without supersymmetry, in three spacetime dimensions. They interpolate from the critical large $N$ vector model dual to a classical higher spin theory, towards a theory with a classical string dual. We analyze the spectrum and OPE data of the theories. While the supersymmetric model is always well-behaved the non-supersymmetric model is unitary only over a small parameter range. We offer some speculations on the origin of strings from the higher spins. △ Less

Submitted 7 July, 2022; v1 submitted 16 December, 2021; originally announced December 2021.

Comments: 7 pages + appendix, several figures. v2: minor changes, published version

Persistent Corner Spin Mode at the Quantum Critical Point of a Plaquette Heisenberg Model

Authors: Yining Xu, Chen Peng, Zijian Xiong, Long Zhang

Abstract: Gapless edge states are the hallmark of a large class of topological states of matter. Recently, intensive research has been devoted to understanding the physical properties of the edge states at the quantum phase transitions of the bulk topological states. A higher-order symmetry-protected topological state is realized in a plaquette Heisenberg model on the square lattice. In its disordered phase… ▽ More Gapless edge states are the hallmark of a large class of topological states of matter. Recently, intensive research has been devoted to understanding the physical properties of the edge states at the quantum phase transitions of the bulk topological states. A higher-order symmetry-protected topological state is realized in a plaquette Heisenberg model on the square lattice. In its disordered phase, the lattice with an open boundary hosts either dangling corner states with spin-$1/2$ degeneracy characterizing the topological phase, or nondangling corner states without degeneracy, which depends on the bond configuration near the corners. In this work, we study the critical behavior of these corner states at the quantum critical point (QCP), and find that the spin-$1/2$ corner state induces a new universality class of the corner critical behavior, which is distinct from the ordinary transition of the nondangling corners. In particular, we find that the dangling spin-$1/2$ corner state persists at the QCP despite its coupling to the critical spin fluctuations in the bulk. This shows the robustness of the corner state of the higher-order topological state. △ Less

Submitted 13 December, 2022; v1 submitted 8 December, 2021; originally announced December 2021.

Comments: 6 pages, 5 figures; v2: revised version

Journal ref: Phys. Rev. B 106, 214409 (2022)

RSBNet: One-Shot Neural Architecture Search for A Backbone Network in Remote Sensing Image Recognition

Authors: Cheng Peng, Yangyang Li, Ronghua Shang, Licheng Jiao

Abstract: Recently, a massive number of deep learning based approaches have been successfully applied to various remote sensing image (RSI) recognition tasks. However, most existing advances of deep learning methods in the RSI field heavily rely on the features extracted by the manually designed backbone network, which severely hinders the potential of deep learning models due the complexity of RSI and the… ▽ More Recently, a massive number of deep learning based approaches have been successfully applied to various remote sensing image (RSI) recognition tasks. However, most existing advances of deep learning methods in the RSI field heavily rely on the features extracted by the manually designed backbone network, which severely hinders the potential of deep learning models due the complexity of RSI and the limitation of prior knowledge. In this paper, we research a new design paradigm for the backbone architecture in RSI recognition tasks, including scene classification, land-cover classification and object detection. A novel one-shot architecture search framework based on weight-sharing strategy and evolutionary algorithm is proposed, called RSBNet, which consists of three stages: Firstly, a supernet constructed in a layer-wise search space is pretrained on a self-assembled large-scale RSI dataset based on an ensemble single-path training strategy. Next, the pre-trained supernet is equipped with different recognition heads through the switchable recognition module and respectively fine-tuned on the target dataset to obtain task-specific supernet. Finally, we search the optimal backbone architecture for different recognition tasks based on the evolutionary algorithm without any network training. Extensive experiments have been conducted on five benchmark datasets for different recognition tasks, the results show the effectiveness of the proposed search paradigm and demonstrate that the searched backbone is able to flexibly adapt different RSI recognition tasks and achieve impressive performance. △ Less

Submitted 6 December, 2021; originally announced December 2021.

On Chern minimal surfaces in Hermitian surfaces

Authors: Chiakuei Peng, Xiaowei Xu

Abstract: In this paper we introduce the Chern minimal surface in Hermitian surfaces by using the Chern connection, and we show that it only has isolated complex and anticomplex points for a generic one (neither holomorphic nor antiholomorphic). For a generic Chern minimal $f$ from compact Riemann surface $Σ$ in a Hermitian surface $M$, we establish two identities which related to the sum of the orders of a… ▽ More In this paper we introduce the Chern minimal surface in Hermitian surfaces by using the Chern connection, and we show that it only has isolated complex and anticomplex points for a generic one (neither holomorphic nor antiholomorphic). For a generic Chern minimal $f$ from compact Riemann surface $Σ$ in a Hermitian surface $M$, we establish two identities which related to the sum of the orders of all complex points, anticomplex points denoted by $P$, $Q$ respectively, the cap product of the pull-back of the first Chern class $f^*c_1(M)$ and $[Σ]$, the Euler characteristic of tangent bundle $χ(TΣ)$ and the Euler characteristic of normal bundle $χ(T^\perpΣ)$. More precisely, we obtain the formulae $P-Q=-f^*c_1(M)[Σ]$ and $P+Q=-\big(χ(TΣ)+χ(T^\perpΣ)\big)$. We also give some applications of these formulae. △ Less

Submitted 4 December, 2021; originally announced December 2021.

Comments: All comments are welcome!

How global observation works in Federated Learning: Integrating vertical training into Horizontal Federated Learning

Authors: Shuo Wan, Jiaxun Lu, **yi Fan, Yunfeng Shao, Chenghui Peng, Khaled B. Letaief

Abstract: Federated learning (FL) has recently emerged as a transformative paradigm that jointly train a model with distributed data sets in IoT while avoiding the need for central data collection. Due to the limited observation range, such data sets can only reflect local information, which limits the quality of trained models. In practice, the global information and local observations would require a join… ▽ More Federated learning (FL) has recently emerged as a transformative paradigm that jointly train a model with distributed data sets in IoT while avoiding the need for central data collection. Due to the limited observation range, such data sets can only reflect local information, which limits the quality of trained models. In practice, the global information and local observations would require a joint consideration for learning to make a reasonable policy. However, in horizontal FL, the central agency only acts as a model aggregator without utilizing its global observation to further improve the model. This could significantly degrade the performance in some missions such as traffic flow prediction in network systems, where the global information may enhance the accuracy. Meanwhile, the global feature may not be directly transmitted to agents for data security. How to utilize the global observation residing in the central agency while protecting its safety thus rises up as an important problem in FL. In this paper, we develop a vertical-horizontal federated learning (VHFL) process, where the global feature is shared with the agents in a procedure similar to that of vertical FL without any extra communication rounds. By considering the delay and packet loss, we will analyze VHFL convergence and validate its performance by experiments. It is shown that the proposed VHFL could enhance the accuracy compared with horizontal FL while still protecting the security of global data. △ Less

Submitted 10 December, 2021; v1 submitted 2 December, 2021; originally announced December 2021.