Showing 51–100 of 360 results for author: Zou, C

Transporting cold atoms towards a GaN-on-sapphire chip via an optical conveyor belt

Authors: Lei Xu, Ling-Xiao Wang, Guang-Jie Chen, Liang Chen, Yuan-Hao Yang, Xin-Biao Xu, Ai** Liu, Chuan-Feng Li, Guang-Can Guo, Chang-Ling Zou, Guo-Yong Xiang

Abstract: Trapped atoms on photonic structures inspire many novel quantum devices for quantum information processing and quantum sensing. Here, we have demonstrated a hybrid photonic-atom chip platform based on a GaN-on-sapphire chip and the transport of an ensemble of atoms from free space towards the chip with an optical conveyor belt. The maximum transport efficiency of atoms is about 50% with a transpor… ▽ More Trapped atoms on photonic structures inspire many novel quantum devices for quantum information processing and quantum sensing. Here, we have demonstrated a hybrid photonic-atom chip platform based on a GaN-on-sapphire chip and the transport of an ensemble of atoms from free space towards the chip with an optical conveyor belt. The maximum transport efficiency of atoms is about 50% with a transport distance of 500 $\mathrm{μm}$. Our results open up a new route toward the efficiently loading of cold atoms into the evanescent-field trap formed by the photonic integrated circuits, which promises strong and controllable interactions between single atoms and single photons. △ Less

Submitted 13 May, 2023; originally announced May 2023.

Comments: 7 pages, 5 figures

ARDIE: AR, Dialogue, and Eye Gaze Policies for Human-Robot Collaboration

Authors: Chelsea Zou, Kishan Chandan, Yan Ding, Shiqi Zhang

Abstract: Human-robot collaboration (HRC) has become increasingly relevant in industrial, household, and commercial settings. However, the effectiveness of such collaborations is highly dependent on the human and robots' situational awareness of the environment. Improving this awareness includes not only aligning perceptions in a shared workspace, but also bidirectionally communicating intent and visualizin… ▽ More Human-robot collaboration (HRC) has become increasingly relevant in industrial, household, and commercial settings. However, the effectiveness of such collaborations is highly dependent on the human and robots' situational awareness of the environment. Improving this awareness includes not only aligning perceptions in a shared workspace, but also bidirectionally communicating intent and visualizing different states of the environment to enhance scene understanding. In this paper, we propose ARDIE (Augmented Reality with Dialogue and Eye Gaze), a novel intelligent agent that leverages multi-modal feedback cues to enhance HRC. Our system utilizes a decision theoretic framework to formulate a joint policy that incorporates interactive augmented reality (AR), natural language, and eye gaze to portray current and future states of the environment. Through object-specific AR renders, the human can visualize future object interactions to make adjustments as needed, ultimately providing an interactive and efficient collaboration between humans and robots. △ Less

Submitted 8 May, 2023; originally announced May 2023.

MINN: Learning the dynamics of differential-algebraic equations and application to battery modeling

Authors: Yicun Huang, Changfu Zou, Yang Li, Torsten Wik

Abstract: The concept of integrating physics-based and data-driven approaches has become popular for modeling sustainable energy systems. However, the existing literature mainly focuses on the data-driven surrogates generated to replace physics-based models. These models often trade accuracy for speed but lack the generalisability, adaptability, and interpretability inherent in physics-based models, which a… ▽ More The concept of integrating physics-based and data-driven approaches has become popular for modeling sustainable energy systems. However, the existing literature mainly focuses on the data-driven surrogates generated to replace physics-based models. These models often trade accuracy for speed but lack the generalisability, adaptability, and interpretability inherent in physics-based models, which are often indispensable in the modeling of real-world dynamic systems for optimization and control purposes. In this work, we propose a novel architecture for generating model-integrated neural networks (MINN) to allow integration on the level of learning physics-based dynamics of the system. The obtained hybrid model solves an unsettled research problem in control-oriented modeling, i.e., how to obtain an optimally simplified model that is physically insightful, numerically accurate, and computationally tractable simultaneously. We apply the proposed neural network architecture to model the electrochemical dynamics of lithium-ion batteries and show that MINN is extremely data-efficient to train while being sufficiently generalizable to previously unseen input data, owing to its underlying physical invariants. The MINN battery model has an accuracy comparable to the first principle-based model in predicting both the system outputs and any locally distributed electrochemical behaviors but achieves two orders of magnitude reduction in the solution time. △ Less

Submitted 27 April, 2023; originally announced April 2023.

Self-do** effect in confined copper selenide semiconducting quantum dots for efficient photoelectrocatalytic oxygen evolution

Authors: Jie Ren, Chenya Zhao, Lanshan He, Congcong Wu, Wenting Jia, Shengwen Xu, Daojian Ye, Weiyang Xu, Fu** Huang, Hang Zhou, Chengwu Zou, Ce Hu, Ting Yu, Xingfang Luo, Cailei Yuan

Abstract: Self-do** can not only suppress the photogenerated charge recombination of semiconducting quantum dots by self-introducing trap** states within the bandgap, but also provide high-density catalytic active sites as the consequence of abundant non-saturated bonds associated with the defects. Here, we successfully prepared semiconducting copper selenide (CuSe) confined quantum dots with abundant v… ▽ More Self-do** can not only suppress the photogenerated charge recombination of semiconducting quantum dots by self-introducing trap** states within the bandgap, but also provide high-density catalytic active sites as the consequence of abundant non-saturated bonds associated with the defects. Here, we successfully prepared semiconducting copper selenide (CuSe) confined quantum dots with abundant vacancies and systematically investigated their photoelectrochemical characteristics. Photoluminescence characterizations reveal that the presence of vacancies reduces the emission intensity dramatically, indicating a low recombination rate of photogenerated charge carriers due to the self-introduced trap** states within the bandgap. In addition, the ultra-low charge transfer resistance measured by electrochemical impedance spectroscopy implies the efficient charge transfer of CuSe semiconducting quantum dots-based photoelectrocatalysts, which is guaranteed by the high conductivity of their confined structure as revealed by room-temperature electrical transport measurements. Such high conductivity and low photogenerated charge carriers recombination rate, combined with high-density active sites and confined structure, guaranteeing the remarkable photoelectrocatalytic performance and stability as manifested by photoelectrocatalysis characterizations. This work promotes the development of semiconducting quantum dots-based photoelectrocatalysis and demonstrates CuSe semiconducting quantum confined catalysts as an advanced photoelectrocatalysts for oxygen evolution reaction. △ Less

Submitted 13 April, 2023; originally announced April 2023.

Weight Try-Once-Discard Protocol-Based L_2 L_infinity State Estimation for Markovian Jum** Neural Networks with Partially Known Transition Probabilities

Authors: Cong Zou, Wei Chen

Abstract: It was the L_2 L_infinity performance index that for the first time is initiated into the discussion on state estimation of delayed MJNNs with with partially known transition probabilities, which provides a more general promotion for the estimation error.The WTOD protocol is adopted to dispatch the sensor nodes so as to effectively alleviate the updating frequency of output signals. The hybrid eff… ▽ More It was the L_2 L_infinity performance index that for the first time is initiated into the discussion on state estimation of delayed MJNNs with with partially known transition probabilities, which provides a more general promotion for the estimation error.The WTOD protocol is adopted to dispatch the sensor nodes so as to effectively alleviate the updating frequency of output signals. The hybrid effects of the time delays, Markov chain, and protocol parameters are apparently reflected in the co-designed estimator which can be solved by a combination of comprehensive matrix inequalities. △ Less

Submitted 10 April, 2023; originally announced April 2023.

Atom-referenced on-chip soliton microcomb

Authors: Rui Niu, Shuai Wan, Tian-Peng Hua, Wei-Qiang Wang, Zheng-Yu Wang, ** Li, Zhu-Bo Wang, Ming Li, Zhen Shen, Y. R. Sun, Shui-Ming Hu, B. E. Little, S. T. Chu, Wei Zhao, Guang-Can Guo, Chang-Ling Zou, Yun-Feng Xiao, Wen-Fu Zhang, Chun-Hua Dong

Abstract: For the applications of the frequency comb in microresonators, it is essential to obtain a fully frequency-stabilized microcomb laser source. Here, we demonstrate an atom-referenced stabilized soliton microcomb generation system based on the integrated microring resonator. The pump light around $1560.48\,\mathrm{nm}$ locked to an ultra-low-expansion (ULE) cavity, is frequency-doubled and reference… ▽ More For the applications of the frequency comb in microresonators, it is essential to obtain a fully frequency-stabilized microcomb laser source. Here, we demonstrate an atom-referenced stabilized soliton microcomb generation system based on the integrated microring resonator. The pump light around $1560.48\,\mathrm{nm}$ locked to an ultra-low-expansion (ULE) cavity, is frequency-doubled and referenced to the atomic transition of $^{87}\mathrm{Rb}$. The repetition rate of the soliton microcomb is injection-locked to an atomic-clock-stabilized radio frequency (RF) source, leading to mHz stabilization at $1$ seconds. As a result, all comb lines have been frequency-stabilized based on the atomic reference and could be determined with very high precision reaching $\sim18\,\mathrm{Hz}$ at 1 second, corresponding to the frequency stability of $9.5\times10^{-14}$. Our approach provides an integrated and fully stabilized microcomb experiment scheme with no requirement of $f-2f$ technique, which could be easily implemented and generalized to various photonic platforms, thus paving the way towards the portable and ultraprecise optical sources for high precision spectroscopy. △ Less

Submitted 4 May, 2023; v1 submitted 3 April, 2023; originally announced April 2023.

Comments: 7 pages, 8 figures

CAP-VSTNet: Content Affinity Preserved Versatile Style Transfer

Authors: Linfeng Wen, Chengying Gao, Changqing Zou

Abstract: Content affinity loss including feature and pixel affinity is a main problem which leads to artifacts in photorealistic and video style transfer. This paper proposes a new framework named CAP-VSTNet, which consists of a new reversible residual network and an unbiased linear transform module, for versatile style transfer. This reversible residual network can not only preserve content affinity but n… ▽ More Content affinity loss including feature and pixel affinity is a main problem which leads to artifacts in photorealistic and video style transfer. This paper proposes a new framework named CAP-VSTNet, which consists of a new reversible residual network and an unbiased linear transform module, for versatile style transfer. This reversible residual network can not only preserve content affinity but not introduce redundant information as traditional reversible networks, and hence facilitate better stylization. Empowered by Matting Laplacian training loss which can address the pixel affinity loss problem led by the linear transform, the proposed framework is applicable and effective on versatile style transfer. Extensive experiments show that CAP-VSTNet can produce better qualitative and quantitative results in comparison with the state-of-the-art methods. △ Less

Submitted 31 March, 2023; originally announced March 2023.

Comments: CVPR 2023

MXM-CLR: A Unified Framework for Contrastive Learning of Multifold Cross-Modal Representations

Authors: Ye Wang, Bowei Jiang, Changqing Zou, Rui Ma

Abstract: Multifold observations are common for different data modalities, e.g., a 3D shape can be represented by multi-view images and an image can be described with different captions. Existing cross-modal contrastive representation learning (XM-CLR) methods such as CLIP are not fully suitable for multifold data as they only consider one positive pair and treat other pairs as negative when computing the c… ▽ More Multifold observations are common for different data modalities, e.g., a 3D shape can be represented by multi-view images and an image can be described with different captions. Existing cross-modal contrastive representation learning (XM-CLR) methods such as CLIP are not fully suitable for multifold data as they only consider one positive pair and treat other pairs as negative when computing the contrastive loss. In this paper, we propose MXM-CLR, a unified framework for contrastive learning of multifold cross-modal representations. MXM-CLR explicitly models and learns the relationships between multifold observations of instances from different modalities for more comprehensive representation learning. The key of MXM-CLR is a novel multifold-aware hybrid loss which considers multiple positive observations when computing the hard and soft relationships for the cross-modal data pairs. We conduct quantitative and qualitative comparisons with SOTA baselines for cross-modal retrieval tasks on the Text2Shape and Flickr30K datasets. We also perform extensive evaluations on the adaptability and generalizability of MXM-CLR, as well as ablation studies on the loss design and effects of batch sizes. The results show the superiority of MXM-CLR in learning better representations for the multifold data. The code is available at https://github.com/JLU-ICL/MXM-CLR. △ Less

Submitted 20 March, 2023; v1 submitted 19 March, 2023; originally announced March 2023.

Comments: 16 pages, 14 figures

Ensemble Nonlinear Model Predictive Control for Residential Solar-Battery Energy Management

Authors: Yang Li, D. Mahinda Vilathgamuwa, Daniel E. Quevedo, Chih Feng Lee, Changfu Zou

Abstract: In a dynamic distribution market environment, residential prosumers with solar power generation and battery energy storage devices can flexibly interact with the power grid via power exchange. Providing a schedule of this bidirectional power dispatch can facilitate the operational planning for the grid operator and bring additional benefits to the prosumers with some economic incentives. However,… ▽ More In a dynamic distribution market environment, residential prosumers with solar power generation and battery energy storage devices can flexibly interact with the power grid via power exchange. Providing a schedule of this bidirectional power dispatch can facilitate the operational planning for the grid operator and bring additional benefits to the prosumers with some economic incentives. However, the major obstacle to achieving this win-win situation is the difficulty in 1) predicting the nonlinear behaviors of battery degradation under unknown operating conditions and 2) addressing the highly uncertain generation/load patterns, in a computationally viable way. This paper thus establishes a robust short-term dispatch framework for residential prosumers equipped with rooftop solar photovoltaic panels and household batteries. The objective is to achieve the minimum-cost operation under the dynamic distribution energy market environment with stipulated dispatch rules. A general nonlinear optimization problem is formulated, taking into consideration the operating costs due to electricity trading, battery degradation, and various operating constraints. The optimization problem is solved in real-time using a proposed ensemble nonlinear model predictive control-based economic dispatch strategy, where the uncertainty in the forecast has been addressed adequately albeit with limited local data. The effectiveness of the proposed algorithm has been validated using real-world prosumer datasets. △ Less

Submitted 18 March, 2023; originally announced March 2023.

Comments: 13 pages, 9 figures

Journal ref: IEEE Transactions on Control Systems Technology, vol. 31, no. 5, pp. 2188-2200, Sep. 2023

Absence of localized $5d^1$ electrons in KTaO$_3$ interface superconductors

Authors: Xinqiang Cai, Jungho Kim, Leonardo Martinelli, Piero Florio, Matteo Corti, Weiliang Qiao, Yanqiu Sun, Jiasen Niu, Quentin Faure, Christoph Sahle, Qingzheng Qiu, Qian Xiao, Xiquan Zheng, Qizhi Li, Changwei Zou, Xinyi Jiang, Giacomo Ghiringhelli, Wei Han, Yanwu Xie, Yi Lu, Marco Moretti Sala, Yingying Peng

Abstract: Recently, an exciting discovery of orientation-dependent superconductivity was made in two-dimensional electron gas (2DEG) at the interfaces of LaAlO$_3$/KTaO$_3$ (LAO/KTO) or EuO/KTaO$_3$ (EuO/KTO). The superconducting transition temperature can reach a $T_c$ of up to $\sim$ 2.2 K, which is significantly higher than its 3$d$ counterpart LaAlO$_3$/SrTiO$_3$ (LAO/STO) with a $T_c$ of $\sim$ 0.2 K.… ▽ More Recently, an exciting discovery of orientation-dependent superconductivity was made in two-dimensional electron gas (2DEG) at the interfaces of LaAlO$_3$/KTaO$_3$ (LAO/KTO) or EuO/KTaO$_3$ (EuO/KTO). The superconducting transition temperature can reach a $T_c$ of up to $\sim$ 2.2 K, which is significantly higher than its 3$d$ counterpart LaAlO$_3$/SrTiO$_3$ (LAO/STO) with a $T_c$ of $\sim$ 0.2 K. However, the underlying origin remains to be understood. To uncover the nature of electrons in KTO-based interfaces, we employ x-ray absorption spectroscopy (XAS) and resonant inelastic x-ray spectroscopy (RIXS) to study LAO/KTO and EuO/KTO with different orientations. We reveal the absence of $dd$ orbital excitations in all the measured samples. Our RIXS results are well reproduced by calculations that considered itinerant $5d$ electrons hybridized with O $2p$ electrons. This suggests that there is a lack of localized Ta $5d^1$ electrons in KTO interface superconductors, which is consistent with the absence of magnetic hysteresis observed in magneto-resistance (MR) measurements. These findings offer new insights into our understanding of superconductivity in Ta $5d$ interface superconductors and their potential applications. △ Less

Submitted 3 March, 2023; originally announced March 2023.

Comments: 8 pages, 6 figures

Journal ref: Phys. Rev. B 108, 235167 (2023)

DC-Former: Diverse and Compact Transformer for Person Re-Identification

Authors: Wen Li, Cheng Zou, Meng Wang, Furong Xu, Jianan Zhao, Ruobing Zheng, Yuan Cheng, Wei Chu

Abstract: In person re-identification (re-ID) task, it is still challenging to learn discriminative representation by deep learning, due to limited data. Generally speaking, the model will get better performance when increasing the amount of data. The addition of similar classes strengthens the ability of the classifier to identify similar identities, thereby improving the discrimination of representation.… ▽ More In person re-identification (re-ID) task, it is still challenging to learn discriminative representation by deep learning, due to limited data. Generally speaking, the model will get better performance when increasing the amount of data. The addition of similar classes strengthens the ability of the classifier to identify similar identities, thereby improving the discrimination of representation. In this paper, we propose a Diverse and Compact Transformer (DC-Former) that can achieve a similar effect by splitting embedding space into multiple diverse and compact subspaces. Compact embedding subspace helps model learn more robust and discriminative embedding to identify similar classes. And the fusion of these diverse embeddings containing more fine-grained information can further improve the effect of re-ID. Specifically, multiple class tokens are used in vision transformer to represent multiple embedding spaces. Then, a self-diverse constraint (SDC) is applied to these spaces to push them away from each other, which makes each embedding space diverse and compact. Further, a dynamic weight controller(DWC) is further designed for balancing the relative importance among them during training. The experimental results of our method are promising, which surpass previous state-of-the-art methods on several commonly used person re-ID benchmarks. △ Less

Submitted 28 February, 2023; originally announced February 2023.

Comments: Accepted by AAAI23

Protecting quantum entanglement between error-corrected logical qubits

Authors: Weizhou Cai, Xianghao Mu, Weiting Wang, Jie Zhou, Yuwei Ma, Xiaoxuan Pan, Ziyue Hua, Xinyu Liu, Guangming Xue, Haifeng Yu, Haiyan Wang, Yipu Song, Chang-Ling Zou, Luyan Sun

Abstract: Entanglement represents one of the most important conceptual advances in physics during the last century and is also one of the most essential resources in quantum information science. However, entanglement is fragile and its potential advantages in applications are hindered by decoherence in practice. Here, we experimentally realize entangled logical qubits (ELQ) with a bosonic quantum module by… ▽ More Entanglement represents one of the most important conceptual advances in physics during the last century and is also one of the most essential resources in quantum information science. However, entanglement is fragile and its potential advantages in applications are hindered by decoherence in practice. Here, we experimentally realize entangled logical qubits (ELQ) with a bosonic quantum module by encoding quantum information into spatially separated microwave modes. The entanglement is protected by repetitive quantum error correction, and the coherence time of the purified ELQ via error detection is improved by 45$\%$ compared with the unprotected ELQ and exceeds that of the entangled physical qubits. In addition, violation of the Bell inequality by logical qubits is demonstrated for the first time with the measured Bell signal B=2.250$\pm$0.019 after purification, surpassing the classical bound by 13 standard deviations. The protected ELQ could be applied in future explorations of quantum foundations and applications of quantum networks. △ Less

Submitted 25 February, 2023; originally announced February 2023.

Comments: 26 pages, 15 figures, 8 tables

Statistical Inference for Ultrahigh Dimensional Location Parameter Based on Spatial Median

Authors: Guanghui Cheng, Liuhua Peng, Changliang Zou

Abstract: Motivated by the widely used geometric median-of-means estimator in machine learning, this paper studies statistical inference for ultrahigh dimensionality location parameter based on the sample spatial median under a general multivariate model, including simultaneous confidence intervals construction, global tests, and multiple testing with false discovery rate control. To achieve these goals, we… ▽ More Motivated by the widely used geometric median-of-means estimator in machine learning, this paper studies statistical inference for ultrahigh dimensionality location parameter based on the sample spatial median under a general multivariate model, including simultaneous confidence intervals construction, global tests, and multiple testing with false discovery rate control. To achieve these goals, we derive a novel Bahadur representation of the sample spatial median with a maximum-norm bound on the remainder term, and establish Gaussian approximation for the sample spatial median over the class of hyperrectangles. In addition, a multiplier bootstrap algorithm is proposed to approximate the distribution of the sample spatial median. The approximations are valid when the dimension diverges at an exponentially rate of the sample size, which facilitates the application of the spatial median in the ultrahigh dimensional region. The proposed approaches are further illustrated by simulations and analysis of a genomic dataset from a microarray study. △ Less

Submitted 8 January, 2023; originally announced January 2023.

Selective conformal inference with false coverage-statement rate control

Authors: Yajie Bao, Yuyang Huo, Haojie Ren, Changliang Zou

Abstract: Conformal inference is a popular tool for constructing prediction intervals (PI). We consider here the scenario of post-selection/selective conformal inference, that is PIs are reported only for individuals selected from an unlabeled test data. To account for multiplicity, we develop a general split conformal framework to construct selective PIs with the false coverage-statement rate (FCR) control… ▽ More Conformal inference is a popular tool for constructing prediction intervals (PI). We consider here the scenario of post-selection/selective conformal inference, that is PIs are reported only for individuals selected from an unlabeled test data. To account for multiplicity, we develop a general split conformal framework to construct selective PIs with the false coverage-statement rate (FCR) control. We first investigate the Benjamini and Yekutieli (2005)'s FCR-adjusted method in the present setting, and show that it is able to achieve FCR control but yields uniformly inflated PIs. We then propose a novel solution to the problem, named as Selective COnditional conformal Predictions (SCOP), which entails performing selection procedures on both calibration set and test set and construct marginal conformal PIs on the selected sets by the aid of conditional empirical distribution obtained by the calibration set. Under a unified framework and exchangeable assumptions, we show that the SCOP can exactly control the FCR. More importantly, we provide non-asymptotic miscoverage bounds for a general class of selection procedures beyond exchangeablity and discuss the conditions under which the SCOP is able to control the FCR. As special cases, the SCOP with quantile-based selection or conformal p-values-based multiple testing procedures enjoys valid coverage guarantee under mild conditions. Numerical results confirm the effectiveness and robustness of SCOP in FCR control and show that it achieves more narrowed PIs over existing methods in many settings. △ Less

Submitted 12 March, 2024; v1 submitted 2 January, 2023; originally announced January 2023.

Hierarchical Deep Reinforcement Learning for VWAP Strategy Optimization

Authors: Xiaodong Li, Pang**g Wu, Chenxin Zou, Qing Li

Abstract: Designing an intelligent volume-weighted average price (VWAP) strategy is a critical concern for brokers, since traditional rule-based strategies are relatively static that cannot achieve a lower transaction cost in a dynamic market. Many studies have tried to minimize the cost via reinforcement learning, but there are bottlenecks in improvement, especially for long-duration strategies such as the… ▽ More Designing an intelligent volume-weighted average price (VWAP) strategy is a critical concern for brokers, since traditional rule-based strategies are relatively static that cannot achieve a lower transaction cost in a dynamic market. Many studies have tried to minimize the cost via reinforcement learning, but there are bottlenecks in improvement, especially for long-duration strategies such as the VWAP strategy. To address this issue, we propose a deep learning and hierarchical reinforcement learning jointed architecture termed Macro-Meta-Micro Trader (M3T) to capture market patterns and execute orders from different temporal scales. The Macro Trader first allocates a parent order into tranches based on volume profiles as the traditional VWAP strategy does, but a long short-term memory neural network is used to improve the forecasting accuracy. Then the Meta Trader selects a short-term subgoal appropriate to instant liquidity within each tranche to form a mini-tranche. The Micro Trader consequently extracts the instant market state and fulfils the subgoal with the lowest transaction cost. Our experiments over stocks listed on the Shanghai stock exchange demonstrate that our approach outperforms baselines in terms of VWAP slippage, with an average cost saving of 1.16 base points compared to the optimal baseline. △ Less

Submitted 11 December, 2022; originally announced December 2022.

Deep quantum neural networks equipped with backpropagation on a superconducting processor

Authors: Xiaoxuan Pan, Zhide Lu, Weiting Wang, Ziyue Hua, Yifang Xu, Weikang Li, Weizhou Cai, Xuegang Li, Haiyan Wang, Yi-Pu Song, Chang-Ling Zou, Dong-Ling Deng, Luyan Sun

Abstract: Deep learning and quantum computing have achieved dramatic progresses in recent years. The interplay between these two fast-growing fields gives rise to a new research frontier of quantum machine learning. In this work, we report the first experimental demonstration of training deep quantum neural networks via the backpropagation algorithm with a six-qubit programmable superconducting processor. I… ▽ More Deep learning and quantum computing have achieved dramatic progresses in recent years. The interplay between these two fast-growing fields gives rise to a new research frontier of quantum machine learning. In this work, we report the first experimental demonstration of training deep quantum neural networks via the backpropagation algorithm with a six-qubit programmable superconducting processor. In particular, we show that three-layer deep quantum neural networks can be trained efficiently to learn two-qubit quantum channels with a mean fidelity up to 96.0% and the ground state energy of molecular hydrogen with an accuracy up to 93.3% compared to the theoretical value. In addition, six-layer deep quantum neural networks can be trained in a similar fashion to achieve a mean fidelity up to 94.8% for learning single-qubit quantum channels. Our experimental results explicitly showcase the advantages of deep quantum neural networks, including quantum analogue of the backpropagation algorithm and less stringent coherence-time requirement for their constituting physical qubits, thus providing a valuable guide for quantum machine learning applications with both near-term and future quantum devices. △ Less

Submitted 5 December, 2022; originally announced December 2022.

Comments: 7 pages (main text) + 11 pages (Supplementary Information), 10 figures

Journal ref: Nature Communications volume 14, Article number: 4006 (2023)

Knowledge Graph Quality Evaluation under Incomplete Information

Authors: Xiaodong Li, Chenxin Zou, Yi Cai, Yuelong Zhu

Abstract: Knowledge graphs (KGs) have attracted more and more attentions because of their fundamental roles in many tasks. Quality evaluation for KGs is thus crucial and indispensable. Existing methods in this field evaluate KGs by either proposing new quality metrics from different dimensions or measuring performances at KG construction stages. However, there are two major issues with those methods. First,… ▽ More Knowledge graphs (KGs) have attracted more and more attentions because of their fundamental roles in many tasks. Quality evaluation for KGs is thus crucial and indispensable. Existing methods in this field evaluate KGs by either proposing new quality metrics from different dimensions or measuring performances at KG construction stages. However, there are two major issues with those methods. First, they highly rely on raw data in KGs, which makes KGs' internal information exposed during quality evaluation. Second, they consider more about the quality at data level instead of ability level, where the latter one is more important for downstream applications. To address these issues, we propose a knowledge graph quality evaluation framework under incomplete information (QEII). The quality evaluation task is transformed into an adversarial Q&A game between two KGs. Winner of the game is thus considered to have better qualities. During the evaluation process, no raw data is exposed, which ensures information protection. Experimental results on four pairs of KGs demonstrate that, compared with baselines, the QEII implements a reasonable quality evaluation at ability level under incomplete information. △ Less

Submitted 12 April, 2023; v1 submitted 2 December, 2022; originally announced December 2022.

QFF: Quantized Fourier Features for Neural Field Representations

Authors: Jae Yong Lee, Yuqun Wu, Chuhang Zou, Shenlong Wang, Derek Hoiem

Abstract: Multilayer perceptrons (MLPs) learn high frequencies slowly. Recent approaches encode features in spatial bins to improve speed of learning details, but at the cost of larger model size and loss of continuity. Instead, we propose to encode features in bins of Fourier features that are commonly used for positional encoding. We call these Quantized Fourier Features (QFF). As a naturally multiresolut… ▽ More Multilayer perceptrons (MLPs) learn high frequencies slowly. Recent approaches encode features in spatial bins to improve speed of learning details, but at the cost of larger model size and loss of continuity. Instead, we propose to encode features in bins of Fourier features that are commonly used for positional encoding. We call these Quantized Fourier Features (QFF). As a naturally multiresolution and periodic representation, our experiments show that using QFF can result in smaller model size, faster training, and better quality outputs for several applications, including Neural Image Representations (NIR), Neural Radiance Field (NeRF) and Signed Distance Function (SDF) modeling. QFF are easy to code, fast to compute, and serve as a simple drop-in addition to many neural field representations. △ Less

Submitted 1 December, 2022; originally announced December 2022.

Beating the break-even point with a discrete-variable-encoded logical qubit

Authors: Zhongchu Ni, Sai Li, Xiaowei Deng, Yanyan Cai, Libo Zhang, Weiting Wang, Zhen-Biao Yang, Haifeng Yu, Fei Yan, Song Liu, Chang-Ling Zou, Luyan Sun, Shi-Biao Zheng, Yuan Xu, Dapeng Yu

Abstract: Quantum error correction (QEC) aims to protect logical qubits from noises by utilizing the redundancy of a large Hilbert space, where an error, once it occurs, can be detected and corrected in real time. In most QEC codes, a logical qubit is encoded in some discrete variables, e.g., photon numbers. Such encoding schemes make the codewords orthogonal, so that the encoded quantum information can be… ▽ More Quantum error correction (QEC) aims to protect logical qubits from noises by utilizing the redundancy of a large Hilbert space, where an error, once it occurs, can be detected and corrected in real time. In most QEC codes, a logical qubit is encoded in some discrete variables, e.g., photon numbers. Such encoding schemes make the codewords orthogonal, so that the encoded quantum information can be unambiguously extracted after processing. Based on such discrete-variable encodings, repetitive QEC demonstrations have been reported on various platforms, but there the lifetime of the encoded logical qubit is still shorter than that of the best available physical qubit in the entire system, which represents a break-even point that needs to be surpassed for any QEC code to be of practical use. Here we demonstrate a QEC procedure with a logical qubit encoded in photon-number states of a microwave cavity, dispersively coupled to an ancilla superconducting qubit. By applying a pulse featuring a tailored frequency comb to the ancilla, we can repetitively extract the error syndrome with high fidelity and perform error correction with feedback control accordingly, thereby exceeding the break-even point by about 16% lifetime enhancement. Our work illustrates the potential of the hardware-efficient discrete-variable QEC codes towards a reliable quantum information processor. △ Less

Submitted 5 April, 2023; v1 submitted 16 November, 2022; originally announced November 2022.

Comments: Main text: 8 pages, 3 figures, 1 table; Supplement: 12 pages, 8 figures, 2 tables

Journal ref: Nature 616, 56-60 (2023)

Synthetic Topological Vacua of Yang-Mills Fields in Bose-Einstein Condensates

Authors: Jia-Zhen Li, Cong-Jun Zou, Yan-Xiong Du, Qing-Xian Lv, Wei Huang, Zhen-Tao Liang, Dan-Wei Zhang, Hui Yan, Shanchao Zhang, Shi-Liang Zhu

Abstract: Topological vacua are a family of degenerate ground states of Yang-Mills fields with zero field strength but nontrivial topological structures. They play a fundamental role in particle physics and quantum field theory, but have not yet been experimentally observed. Here we report the first theoretical proposal and experimental realization of synthetic topological vacua with a cloud of atomic Bose-… ▽ More Topological vacua are a family of degenerate ground states of Yang-Mills fields with zero field strength but nontrivial topological structures. They play a fundamental role in particle physics and quantum field theory, but have not yet been experimentally observed. Here we report the first theoretical proposal and experimental realization of synthetic topological vacua with a cloud of atomic Bose-Einstein condensates. Our setup provides a promising platform to demonstrate the fundamental concept that a vacuum, rather than being empty, has rich spatial structures. The Hamiltonian for the vacuum of topological number n = 1 is synthesized and the related Hopf index is measured. The vacuum of topological number n = 2 is also realized, and we find that vacua with different topological numbers have distinctive spin textures and Hopf links. Our work opens up opportunities for exploring topological vacua and related long-sought-after instantons in tabletop experiments. △ Less

Submitted 13 November, 2022; originally announced November 2022.

Journal ref: Physical Review Letters 129, 220402 (2022)

Model-free controlled variable selection via data splitting

Authors: Yixin Han, Xu Guo, Changliang Zou

Abstract: Addressing the simultaneous identification of contributory variables while controlling the false discovery rate (FDR) in high-dimensional data is a crucial statistical challenge. In this paper, we propose a novel model-free variable selection procedure in sufficient dimension reduction framework via a data splitting technique. The variable selection problem is first converted to a least squares pr… ▽ More Addressing the simultaneous identification of contributory variables while controlling the false discovery rate (FDR) in high-dimensional data is a crucial statistical challenge. In this paper, we propose a novel model-free variable selection procedure in sufficient dimension reduction framework via a data splitting technique. The variable selection problem is first converted to a least squares procedure with several response transformations. We construct a series of statistics with global symmetry property and leverage the symmetry to derive a data-driven threshold aimed at error rate control. Our approach demonstrates the capability for achieving finite-sample and asymptotic FDR control under mild theoretical conditions. Numerical experiments confirm that our procedure has satisfactory FDR control and higher power compared with existing methods. △ Less

Submitted 22 April, 2024; v1 submitted 22 October, 2022; originally announced October 2022.

Comments: 55 pages, 5 figures, 6 tables

Deep PatchMatch MVS with Learned Patch Coplanarity, Geometric Consistency and Adaptive Pixel Sampling

Authors: Jae Yong Lee, Chuhang Zou, Derek Hoiem

Abstract: Recent work in multi-view stereo (MVS) combines learnable photometric scores and regularization with PatchMatch-based optimization to achieve robust pixelwise estimates of depth, normals, and visibility. However, non-learning based methods still outperform for large scenes with sparse views, in part due to use of geometric consistency constraints and ability to optimize over many views at high res… ▽ More Recent work in multi-view stereo (MVS) combines learnable photometric scores and regularization with PatchMatch-based optimization to achieve robust pixelwise estimates of depth, normals, and visibility. However, non-learning based methods still outperform for large scenes with sparse views, in part due to use of geometric consistency constraints and ability to optimize over many views at high resolution. In this paper, we build on learning-based approaches to improve photometric scores by learning patch coplanarity and encourage geometric consistency by learning a scaled photometric cost that can be combined with reprojection error. We also propose an adaptive pixel sampling strategy for candidate propagation that reduces memory to enable training on larger resolution with more views and a larger encoder. These modifications lead to 6-15% gains in accuracy and completeness on the challenging ETH3D benchmark, resulting in higher F1 performance than the widely used state-of-the-art non-learning approaches ACMM and ACMP. △ Less

Submitted 14 October, 2022; originally announced October 2022.

Controllable atomic collision in a tight optical dipole trap

Authors: Zhu-Bo Wang, Chen-yue Gu, Xin-Xin Hu, Ya-Ting Zhang, Ji-Zhe Zhang, Gang Li, Xiao-Dong He, Xu-Bo Zou, Chun-Hua Dong, Guang-Can Guo, Chang-Ling Zou

Abstract: Single atoms are interesting candidates for studying quantum optics and quantum information processing. Recently, trap** and manipulation of single atoms using tight optical dipole traps have generated considerable interest. Here we report an experimental investigation of the dynamics of atoms in a modified optical dipole trap with a backward propagating dipole trap beam, where a change in the t… ▽ More Single atoms are interesting candidates for studying quantum optics and quantum information processing. Recently, trap** and manipulation of single atoms using tight optical dipole traps have generated considerable interest. Here we report an experimental investigation of the dynamics of atoms in a modified optical dipole trap with a backward propagating dipole trap beam, where a change in the two-atom collision rate by six times has been achieved. The theoretical model presented gives a prediction of high probabilities of few-atom loading rates under proper experimental conditions. This work provides an alternative approach to the control of the few-atom dynamics in a dipole trap and the study of the collective quantum optical effects of a few atoms. △ Less

Submitted 24 October, 2022; v1 submitted 13 October, 2022; originally announced October 2022.

Comments: 5 pages, 3 figures

Self-induced optical non-reciprocity

Authors: Zhu-Bo Wang, Yan-Lei Zhang, Xin-Xin Hu, Guang-Jie Chen, Ming Li, Peng-Fei Yang, Xu-Bo Zou, Peng-Fei Zhang, Chun-Hua Dong, Gang Li, Tian-Cai Zhang, Guang-Can Guo, Chang-Ling Zou

Abstract: Non-reciprocal optical components are indispensable in optical applications, and their realization without any magnetic field arose increasing research interests in photonics. Exciting experimental progress has been achieved by either introducing spatial-temporal modulation of the optical medium or combining Kerr-type optical nonlinearity with spatial asymmetry in photonic structures. However, ext… ▽ More Non-reciprocal optical components are indispensable in optical applications, and their realization without any magnetic field arose increasing research interests in photonics. Exciting experimental progress has been achieved by either introducing spatial-temporal modulation of the optical medium or combining Kerr-type optical nonlinearity with spatial asymmetry in photonic structures. However, extra driving fields are required for the first approach, while the isolation of noise and the transmission of the signal cannot be simultaneously achieved for the other approach. Here, we experimentally demonstrate a new concept of nonlinear non-reciprocal susceptibility for optical media and realize the completely passive isolation of optical signals without any external bias field. The self-induced isolation by the input signal is demonstrated with an extremely high isolation ratio of 63.4 dB, a bandwidth of 2.1 GHz for 60 dB isolation, and a low insertion loss of around 1 dB. Furthermore, novel functional optical devices are realized, including polarization purification and non-reciprocal leverage. The demonstrated nonlinear non-reciprocity provides a versatile tool to control light and deepen our understanding of light-matter interactions, and enables applications ranging from topological photonics to unidirectional quantum information transfer in a network. △ Less

Submitted 13 October, 2022; originally announced October 2022.

Comments: 9 pages, 6 figures

Exceptional entanglement phenomena: non-Hermiticity meeting non-classicality

Authors: Pei-Rong Han, Fan Wu, Xin-Jie Huang, Huai-Zhi Wu, Chang-Ling Zou, Wei Yi, Mengzhen Zhang, Hekang Li, Kai Xu, Dongning Zheng, Heng Fan, Jianming Wen, Zhen-Biao Yang, Shi-Biao Zheng

Abstract: Non-Hermitian (NH) extension of quantum-mechanical Hamiltonians represents one of the most significant advancements in physics. During the past two decades, numerous captivating NH phenomena have been revealed and demonstrated, but all of which can appear in both quantum and classical systems. This leads to the fundamental question: what NH signature presents a radical departure from classical phy… ▽ More Non-Hermitian (NH) extension of quantum-mechanical Hamiltonians represents one of the most significant advancements in physics. During the past two decades, numerous captivating NH phenomena have been revealed and demonstrated, but all of which can appear in both quantum and classical systems. This leads to the fundamental question: what NH signature presents a radical departure from classical physics? The solution of this problem is indispensable for exploring genuine NH quantum mechanics, but remains experimentally untouched so far. Here, we resolve this basic issue by unveiling distinct exceptional entanglement phenomena, exemplified by an entanglement transition, occurring at the exceptional point of NH interacting quantum systems. We illustrate and demonstrate such purely quantum-mechanical NH effects with a naturally dissipative light-matter system, engineered in a circuit quantum electrodynamics architecture. Our results lay the foundation for studies of genuinely quantum-mechanical NH physics, signified by exceptional-point-enabled entanglement behaviors. △ Less

Submitted 31 December, 2023; v1 submitted 10 October, 2022; originally announced October 2022.

Journal ref: Phys. Rev. Lett. 131, 260201 (2023)

Quantum Non-Demolition Measurement on the Spin Precession of Laser-Trapped $^{171}$Yb Atoms

Authors: Y. A. Yang, T. A. Zheng, S. -Z. Wang, W. -K. Hu, Chang-Ling Zou, T. Xia, Z. -T. Lu

Abstract: Quantum non-demolition (QND) measurement enhances the detection efficiency and measurement fidelity, and is highly desired for its applications in precision measurements and quantum information processing. We propose and demonstrate a QND measurement scheme for the spin states of laser-trapped atoms. On $^{171}$Yb atoms held in an optical dipole trap, a transition that is simultaneously cycling, s… ▽ More Quantum non-demolition (QND) measurement enhances the detection efficiency and measurement fidelity, and is highly desired for its applications in precision measurements and quantum information processing. We propose and demonstrate a QND measurement scheme for the spin states of laser-trapped atoms. On $^{171}$Yb atoms held in an optical dipole trap, a transition that is simultaneously cycling, spin-state selective, and spin-state preserving is created by introducing a circularly polarized beam of control laser to optically dress the spin states in the excited level, while leaving the spin states in the ground level unperturbed. We measure the phase of spin precession of $5\times10^{4}$ atoms in a bias magnetic field of 20 mG. This QND approach reduces the optical absorption detection noise by $\sim$19 dB, to a level of 2.3 dB below the atomic quantum projection noise. In addition to providing a general approach for efficient spin-state readout, this all-optical technique allows quick switching and real-time programming for quantum sensing and quantum information processing. △ Less

Submitted 16 September, 2022; originally announced September 2022.

Consistent Selection of the Number of Groups in Panel Models via Sample-Splitting

Authors: Zhe Li, Xuening Zhu, Changliang Zou

Abstract: Group number selection is a key question for group panel data modelling. In this work, we develop a cross validation method to tackle this problem. Specifically, we split the panel data into a training dataset and a testing dataset on the time span. We first use the training dataset to estimate the parameters and group memberships. Then we apply the fitted model to the testing dataset and then the… ▽ More Group number selection is a key question for group panel data modelling. In this work, we develop a cross validation method to tackle this problem. Specifically, we split the panel data into a training dataset and a testing dataset on the time span. We first use the training dataset to estimate the parameters and group memberships. Then we apply the fitted model to the testing dataset and then the group number is estimated by minimizing certain loss function values on the testing dataset. We design the loss functions for panel data models either with or without fixed effects. The proposed method has two advantages. First, the method is totally data-driven thus no further tuning parameters are involved. Second, the method can be flexibly applied to a wide range of panel data models. Theoretically, we establish the estimation consistency by taking advantage of the optimization property of the estimation algorithm. Experiments on a variety of synthetic and empirical datasets are carried out to further illustrate the advantages of the proposed method. △ Less

Submitted 21 November, 2023; v1 submitted 12 September, 2022; originally announced September 2022.

Prevailing charge order in overdoped cuprates beyond the superconducting dome

Authors: Qizhi Li, Hsiao-Yu Huang, Tianshuang Ren, Eugen Weschke, Lele Ju, Changwei Zou, Shilong Zhang, Qingzheng Qiu, Jiarui Liu, Shuhan Ding, Amol Singh, Oleksandr Prokhnenko, Di-**g Huang, Ilya Esterlis, Yao Wang, Yanwu Xie, Yingying Peng

Abstract: The extremely overdoped cuprates are generally considered to be Fermi liquid metals without exotic orders, whereas the underdoped cuprates harbor intertwined states. Contrary to this conventional wisdom, using Cu $L_3$ edge and O $K$ edge resonant x-ray scattering, we reveal a charge order (CO) in overdoped La$_{2-x}$Sr$_x$CuO$_4$ (0.35 $\leq$ x $\leq$ 0.6) beyond the superconducting dome. This CO… ▽ More The extremely overdoped cuprates are generally considered to be Fermi liquid metals without exotic orders, whereas the underdoped cuprates harbor intertwined states. Contrary to this conventional wisdom, using Cu $L_3$ edge and O $K$ edge resonant x-ray scattering, we reveal a charge order (CO) in overdoped La$_{2-x}$Sr$_x$CuO$_4$ (0.35 $\leq$ x $\leq$ 0.6) beyond the superconducting dome. This CO has a periodicity of $\sim$ 6 lattice units with correlation lengths of $\sim 3 - 20$ lattice units. It shows similar in-plane momentum and polarization dependence and dispersive excitations as the CO of underdoped cuprates, but its maximum intensity differs along the c-direction and persists up to 300 K. This CO cannot be explained by either the Fermi surface instability or the doped Hubbard model and its origin remains to be understood. Our results suggest that CO is prevailing in the overdoped metallic regime and superconductivity emerges out of the CO phase upon decreasing hole carriers. △ Less

Submitted 17 October, 2022; v1 submitted 18 August, 2022; originally announced August 2022.

Comments: 6 pages, 4 figures

Journal ref: Physical Review Letters 131, 116002 (2023)

Superconducting cavity piezo-electromechanics: the realization of an acoustic frequency comb at microwave frequencies

Authors: Xu Han, Chang-Ling Zou, Wei Fu, Mingrui Xu, Yuntao Xu, Hong X. Tang

Abstract: We present a nonlinear multimode superconducting electroacoustic system, where the interplay between superconducting kinetic inductance and piezoelectric strong coupling establishes an effective Kerr nonlinearity among multiple acoustic modes at 10 GHz that could hardly be achieved via intrinsic mechanical nonlinearity. By exciting this multimode Kerr system with a single microwave tone, we furthe… ▽ More We present a nonlinear multimode superconducting electroacoustic system, where the interplay between superconducting kinetic inductance and piezoelectric strong coupling establishes an effective Kerr nonlinearity among multiple acoustic modes at 10 GHz that could hardly be achieved via intrinsic mechanical nonlinearity. By exciting this multimode Kerr system with a single microwave tone, we further demonstrate a coherent electroacoustic frequency comb and provide theoretical understanding of multimode nonlinear interaction in the superstrong coupling limit. This nonlinear superconducting electroacoustic system sheds light on the active control of multimode resonator systems and offers an enabling platform for the dynamic study of microcombs at microwave frequencies. △ Less

Submitted 12 July, 2022; originally announced August 2022.

Solutions for Fine-grained and Long-tailed Snake Species Recognition in SnakeCLEF 2022

Authors: Cheng Zou, Furong Xu, Meng Wang, Wen Li, Yuan Cheng

Abstract: Automatic snake species recognition is important because it has vast potential to help lower deaths and disabilities caused by snakebites. We introduce our solution in SnakeCLEF 2022 for fine-grained snake species recognition on a heavy long-tailed class distribution. First, a network architecture is designed to extract and fuse features from multiple modalities, i.e. photograph from visual modali… ▽ More Automatic snake species recognition is important because it has vast potential to help lower deaths and disabilities caused by snakebites. We introduce our solution in SnakeCLEF 2022 for fine-grained snake species recognition on a heavy long-tailed class distribution. First, a network architecture is designed to extract and fuse features from multiple modalities, i.e. photograph from visual modality and geographic locality information from language modality. Then, logit adjustment based methods are studied to relieve the impact caused by the severe class imbalance. Next, a combination of supervised and self-supervised learning method is proposed to make full use of the dataset, including both labeled training data and unlabeled testing data. Finally, post processing strategies, such as multi-scale and multi-crop test-time-augmentation, location filtering and model ensemble, are employed for better performance. With an ensemble of several different models, a private score 82.65%, ranking the 3rd, is achieved on the final leaderboard. △ Less

Submitted 4 July, 2022; originally announced July 2022.

Comments: Top solutions for FGVC9, accepted to CLEF2022

Uniqueness and stability of steady vortex rings for 3D incompressible Euler equation

Authors: Daomin Cao, Shanfa Lai, Guolin Qin, Weicheng Zhan, Changjun Zou

Abstract: In this paper, we are concerned with the uniqueness and nonlinear stability of vortex rings for the 3D Euler equation. By utilizing Arnold 's variational principle for steady states of Euler equations and concentrated compactness method introduced by P. L. Lions, we first establish a general stability criteria for vortex rings in rearrangement classes, which allows us to reduce the stability analy… ▽ More In this paper, we are concerned with the uniqueness and nonlinear stability of vortex rings for the 3D Euler equation. By utilizing Arnold 's variational principle for steady states of Euler equations and concentrated compactness method introduced by P. L. Lions, we first establish a general stability criteria for vortex rings in rearrangement classes, which allows us to reduce the stability analysis of certain vortex rings to the problem of their uniqueness. Subsequently, we prove the uniqueness of a special family of vortex rings with a small cross-section and polynomial type distribution function. These vortex rings correspond to global classical solutions to the 3D Euler equation and have been shown to exist by many celebrate works. The proof is achieved by studying carefully asymptotic behaviors of vortex rings as they tend to a circular filament and applying local Pohozaev identities. Consequently, we provide the first family of nonlinear stable classical vortex ring solutions to the 3D Euler equation. △ Less

Submitted 25 December, 2023; v1 submitted 21 June, 2022; originally announced June 2022.

Recurrent Transformer Variational Autoencoders for Multi-Action Motion Synthesis

Authors: Rania Briq, Chuhang Zou, Leonid Pishchulin, Chris Broaddus, Juergen Gall

Abstract: We consider the problem of synthesizing multi-action human motion sequences of arbitrary lengths. Existing approaches have mastered motion sequence generation in single action scenarios, but fail to generalize to multi-action and arbitrary-length sequences. We fill this gap by proposing a novel efficient approach that leverages expressiveness of Recurrent Transformers and generative richness of co… ▽ More We consider the problem of synthesizing multi-action human motion sequences of arbitrary lengths. Existing approaches have mastered motion sequence generation in single action scenarios, but fail to generalize to multi-action and arbitrary-length sequences. We fill this gap by proposing a novel efficient approach that leverages expressiveness of Recurrent Transformers and generative richness of conditional Variational Autoencoders. The proposed iterative approach is able to generate smooth and realistic human motion sequences with an arbitrary number of actions and frames while doing so in linear space and time. We train and evaluate the proposed approach on PROX and Charades datasets, where we augment PROX with ground-truth action labels and Charades with human mesh annotations. Experimental evaluation shows significant improvements in FID score and semantic consistency metrics compared to the state-of-the-art. △ Less

Submitted 27 June, 2022; v1 submitted 14 June, 2022; originally announced June 2022.

Comments: accepted at Transformers for Vision workshop at CVPR 2022

Ultrahigh ion diffusion in oxide crystal by engineering the interfacial transporter channels

Authors: Liang Li, Min Hu, Changlong Hu, Bowen Li, Shanguang Zhao, Guobin Zhang, Liangbin Li, Jun Jiang, Chongwen Zou

Abstract: The mass storage and removal in solid conductors always played vital role on the technological applications such as modern batteries, permeation membranes and neuronal computations, which were seriously lying on the ion diffusion and kinetics in bulk lattice. However, the ions transport was kinetically limited by the low diffusional process, which made it a challenge to fabricate applicable conduc… ▽ More The mass storage and removal in solid conductors always played vital role on the technological applications such as modern batteries, permeation membranes and neuronal computations, which were seriously lying on the ion diffusion and kinetics in bulk lattice. However, the ions transport was kinetically limited by the low diffusional process, which made it a challenge to fabricate applicable conductors with high electronic and ionic conductivities at room temperature. It was known that at essentially all interfaces, the existed space charge layers could modify the charge transport, storage and transfer properties. Thus, in the current study, we proposed an acid solution/WO3/ITO structure and achieved an ultrafast hydrogen transport in WO3 layer by interfacial job-sharing diffusion. In this sandwich structure, the transport pathways of the protons and electrons were spatially separated in acid solution and ITO layer respectively, resulting the pronounced increasing of effective hydrogen diffusion coefficient (Deff) up to 106 times. The experiment and theory simulations also revealed that this accelerated hydrogen transport based on the interfacial job-sharing diffusion was universal and could be extended to other ions and oxide materials as well, which would potentially stimulate systematic studies on ultrafast mixed conductors or faster solid-state electrochemical switching devices in the future. △ Less

Submitted 26 May, 2022; originally announced May 2022.

Comments: 23 pages, 5 figures

Proposal for stable atom trap** on a GaN-on-Sapphire chip

Authors: Ai** Liu, Lei Xu, Xin-Biao Xu, Guang-Jie Chen, Pengfei Zhang, Guo-Yong Xiang, Guang-Can Guo, Qin Wang, Chang-Ling Zou

Abstract: The hybrid photon-atom integrated circuits, which include photonic microcavities and trapped single neutral atom in their evanescent field, are of great potential for quantum information processing. In this platform, the atoms provide the single-photon nonlinearity and long-lived memory, which are complementary to the excellent passive photonics devices in conventional quantum photonic circuits. I… ▽ More The hybrid photon-atom integrated circuits, which include photonic microcavities and trapped single neutral atom in their evanescent field, are of great potential for quantum information processing. In this platform, the atoms provide the single-photon nonlinearity and long-lived memory, which are complementary to the excellent passive photonics devices in conventional quantum photonic circuits. In this work, we propose a stable platform for realizing the hybrid photon-atom circuits based on an unsuspended photonic chip. By introducing high-order modes in the microring, a feasible evanescent-field trap potential well $\sim0.3\,\mathrm{mK}$ could be obtained by only $10\,\mathrm{mW}$-level power in the cavity, compared with $100\,\mathrm{mW}$-level power required in the scheme based on fundamental modes. Based on our scheme, stable single atom trap** with relatively low laser power is feasible for future studies on high-fidelity quantum gates, single-photon sources, as well as many-body quantum physics based on a controllable atom array in a microcavity. △ Less

Submitted 24 May, 2022; originally announced May 2022.

Comments: 8 pages, 6 figures

A Simple Yet Effective SVD-GCN for Directed Graphs

Authors: Chunya Zou, Andi Han, Lequan Lin, Junbin Gao

Abstract: In this paper, we propose a simple yet effective graph neural network for directed graphs (digraph) based on the classic Singular Value Decomposition (SVD), named SVD-GCN. The new graph neural network is built upon the graph SVD-framelet to better decompose graph signals on the SVD ``frequency'' bands. Further the new framelet SVD-GCN is also scaled up for larger scale graphs via using Chebyshev p… ▽ More In this paper, we propose a simple yet effective graph neural network for directed graphs (digraph) based on the classic Singular Value Decomposition (SVD), named SVD-GCN. The new graph neural network is built upon the graph SVD-framelet to better decompose graph signals on the SVD ``frequency'' bands. Further the new framelet SVD-GCN is also scaled up for larger scale graphs via using Chebyshev polynomial approximation. Through empirical experiments conducted on several node classification datasets, we have found that SVD-GCN has remarkable improvements in a variety of graph node learning tasks and it outperforms GCN and many other state-of-the-art graph neural networks for digraphs. Moreover, we empirically demonstate that the SVD-GCN has great denoising capability and robustness to high level graph data attacks. The theoretical and experimental results prove that the SVD-GCN is effective on a variant of graph datasets, meanwhile maintaining stable and even better performance than the state-of-the-arts. △ Less

Submitted 19 May, 2022; originally announced May 2022.

Comments: 14 pages

Model-Free Statistical Inference on High-Dimensional Data

Authors: Xu Guo, Runze Li, Zhe Zhang, Changliang Zou

Abstract: This paper aims to develop an effective model-free inference procedure for high-dimensional data. We first reformulate the hypothesis testing problem via sufficient dimension reduction framework. With the aid of new reformulation, we propose a new test statistic and show that its asymptotic distribution is $χ^2$ distribution whose degree of freedom does not depend on the unknown population distrib… ▽ More This paper aims to develop an effective model-free inference procedure for high-dimensional data. We first reformulate the hypothesis testing problem via sufficient dimension reduction framework. With the aid of new reformulation, we propose a new test statistic and show that its asymptotic distribution is $χ^2$ distribution whose degree of freedom does not depend on the unknown population distribution. We further conduct power analysis under local alternative hypotheses. In addition, we study how to control the false discovery rate of the proposed $χ^2$ tests, which are correlated, to identify important predictors under a model-free framework. To this end, we propose a multiple testing procedure and establish its theoretical guarantees. Monte Carlo simulation studies are conducted to assess the performance of the proposed tests and an empirical analysis of a real-world data set is used to illustrate the proposed methodology. △ Less

Submitted 15 May, 2022; originally announced May 2022.

Massive particle acceleration on a photonic chip via spatial-temporal modulation

Authors: Mai Zhang, Xie-hang Yu, Xin-Biao Xu, Guang-Can Guo, Chang-Ling Zou

Abstract: Recently, the spectral manipulation of single photons has been achieved through spatial-temporal modulation of the optical refractive index. Here, we generalize this mechanism to massive particles, i.e. realizing the acceleration or deceleration of particles through the spatial-temporal modulation of potential induced by lasers. On a photonic integrated chip, we propose a MeV-magnitude acceleratio… ▽ More Recently, the spectral manipulation of single photons has been achieved through spatial-temporal modulation of the optical refractive index. Here, we generalize this mechanism to massive particles, i.e. realizing the acceleration or deceleration of particles through the spatial-temporal modulation of potential induced by lasers. On a photonic integrated chip, we propose a MeV-magnitude acceleration by distributed modulation units driven by lasers. The mechanism could also be applied to atom trap**, which promises a millimeter-scale decelerator to trap atoms. The spatial-temporal modulation approach is universal and could be generalized to other systems, which may play a significant role in hybrid photonic chip and microscale particle manipulation. △ Less

Submitted 25 April, 2022; originally announced April 2022.

Comments: 5 pages, 4 figures

Geometrical bounds on irreversibility in squeezed thermal bath

Authors: Chen-Juan Zou, Yue Li, Jia-Bin You, Qiong Chen, Wan-Li Yang, Mang Feng

Abstract: Irreversible entropy production (IEP) plays an important role in quantum thermodynamic processes. Here we investigate the geometrical bounds of IEP in nonequilibrium thermodynamics by exemplifying a system coupled to a squeezed thermal bath subject to dissipation and dephasing, respectively. We find that the geometrical bounds of the IEP always shift in contrary way under dissipation and dephasing… ▽ More Irreversible entropy production (IEP) plays an important role in quantum thermodynamic processes. Here we investigate the geometrical bounds of IEP in nonequilibrium thermodynamics by exemplifying a system coupled to a squeezed thermal bath subject to dissipation and dephasing, respectively. We find that the geometrical bounds of the IEP always shift in contrary way under dissipation and dephasing, where the lower and upper bounds turning to be tighter occurs in the situation of dephasing and dissipation, respectively. However, either under dissipation or under dephasing, we may reduce both the critical time of the IEP itself and the critical time of the bounds for reaching an equilibrium by harvesting the benefits of squeezing effects, in which the values of the IEP, quantifying the degree of thermodynamic irreversibility, also becomes smaller. Therefore, due to the nonequilibrium nature of the squeezed thermal bath, the system-bath interaction energy brings prominent impact on the IEP, leading to tightness of its bounds. Our results are not contradictory with the second law of thermodynamics by involving squeezing of the bath as an available resource, which can improve the performance of quantum thermodynamic devices. △ Less

Submitted 25 April, 2022; v1 submitted 18 April, 2022; originally announced April 2022.

Nonlinear optical radiation of a lithium niobate microcavity

Authors: Yuan-Hao Yang, Xin-Biao Xu, Jia-Qi Wang, Mai Zhang, Ming Li, Zheng-Xu Zhu, Zhu-Bo Wang, Chun-Hua Dong, Wei Fang, Huakang Yu, Zhiyuan Li, Guang-Can Guo, Chang-Ling Zou

Abstract: The nonlinear optical radiation of an integrated lithium niobate microcavity is demonstrated, which has been neglected in previous studies of nonlinear photonic devices. We find that the nonlinear coupling between confined optical modes on the chip and continuum modes in free space can be greatly enhanced on the platform of integrated microcavity, with feasible relaxation of the phase-matching con… ▽ More The nonlinear optical radiation of an integrated lithium niobate microcavity is demonstrated, which has been neglected in previous studies of nonlinear photonic devices. We find that the nonlinear coupling between confined optical modes on the chip and continuum modes in free space can be greatly enhanced on the platform of integrated microcavity, with feasible relaxation of the phase-matching condition. With an infrared pump laser, we observe the vertical radiation of second-harmonic wave at the visible band, which indicates a robust phase-matching-free chip-to-free-space frequency converter and also unveils an extra energy dissipation channel for integrated devices. Such an unexpected coherent nonlinear interaction between the free-space beam and the confined mode is also validated by the different frequency generation. Furthermore, based on the phase-matching-free nature of the nonlinear radiation, we build an integrated atomic gas sensor to characterize Rb isotopes with a single telecom laser. The unveiled mechanism of nonlinear optical radiation is universal for all dielectric photonic integrated devices, and provides a simple and robust chip-to-free-space as well as visible-to-telecom interface. △ Less

Submitted 17 April, 2022; originally announced April 2022.

Comments: 7 Pages, 4 figures

Efficient and ultra-stable perovskite light-emitting diodes

Authors: Bingbing Guo, Runchen Lai, Sijie Jiang, Yaxiao Lian, Zhixiang Ren, Puyang Li, Xuhui Cao, Shiyu Xing, Yaxin Wang, Weiwei Li, Chen Zou, Mengyu Chen, Cheng Li, Baodan Zhao, Dawei Di

Abstract: Perovskite light-emitting diodes (PeLEDs) have emerged as a strong contender for next-generation display and information technologies. However, similar to perovskite solar cells, the poor operational stability remains the main obstacle toward commercial applications. Here we demonstrate ultra-stable and efficient PeLEDs with extraordinary operational lifetimes (T50) of 1.0x10^4 h, 2.8x10^4 h, 5.4x… ▽ More Perovskite light-emitting diodes (PeLEDs) have emerged as a strong contender for next-generation display and information technologies. However, similar to perovskite solar cells, the poor operational stability remains the main obstacle toward commercial applications. Here we demonstrate ultra-stable and efficient PeLEDs with extraordinary operational lifetimes (T50) of 1.0x10^4 h, 2.8x10^4 h, 5.4x10^5 h, and 1.9x10^6 h at initial radiance (or current densities) of 3.7 W/sr/m2 (~5 mA/cm2), 2.1 W/sr/m2 (~3.2 mA/cm2), 0.42 W/sr/m2 (~1.1 mA/cm2), and 0.21 W/sr/m2 (~0.7 mA/cm2) respectively, and external quantum efficiencies of up to 22.8%. Key to this breakthrough is the introduction of a dipolar molecular stabilizer, which serves two critical roles simultaneously. First, it prevents the detrimental transformation and decomposition of the alpha-phase FAPbI3 perovskite, by inhibiting the formation of lead and iodide intermediates. Secondly, hysteresis-free device operation and microscopic luminescence imaging experiments reveal substantially suppressed ion migration in the emissive perovskite. The record-long PeLED lifespans are encouraging, as they now satisfy the stability requirement for commercial organic LEDs (OLEDs). These results remove the critical concern that halide perovskite devices may be intrinsically unstable, paving the path toward industrial applications. △ Less

Submitted 16 April, 2022; originally announced April 2022.

Comments: This is a preprint of the paper prior to peer review. New and updated results may be available in the final version from the publisher

Journal ref: Nature Photonics 16, 637-643 (2022)

Multi-grating design for integrated single-atom trap**, manipulation, and readout

Authors: Ai** Liu, Jiawei Liu, Wei Peng, Xin-Biao Xu, Guang-Jie Chen, Xifeng Ren, Qin Wang, Chang-Ling Zou

Abstract: An on-chip multi-grating device is proposed to interface single-atoms and integrated photonic circuits, by guiding and focusing lasers to the area with ~10um above the chip for trap**, state manipulation, and readout of single Rubidium atoms. For the optical dipole trap, two 850 nm laser beams are diffracted and overlapped to form a lattice of single-atom dipole trap, with the diameter of optica… ▽ More An on-chip multi-grating device is proposed to interface single-atoms and integrated photonic circuits, by guiding and focusing lasers to the area with ~10um above the chip for trap**, state manipulation, and readout of single Rubidium atoms. For the optical dipole trap, two 850 nm laser beams are diffracted and overlapped to form a lattice of single-atom dipole trap, with the diameter of optical dipole trap around 2.7um. Similar gratings are designed for guiding 780 nm probe laser to excite and also collect the fluorescence of 87Rb atoms. Such a device provides a compact solution for future applications of single atoms, including the single photon source, single-atom quantum register, and sensor. △ Less

Submitted 12 April, 2022; originally announced April 2022.

Comments: 7 pages, 5 figures

Synthetic five-wave mixing in an integrated microcavity for visible-telecom entanglement generation

Authors: Jia-Qi Wang, Yuan-Hao Yang, Ming Li, Hai-Qi Zhou, Xin-Biao Xu, Ji-Zhe Zhang, Chun-Hua Dong, Guang-Can Guo, Chang-Ling Zou

Abstract: Nonlinear optics processes lie at the heart of photonics and quantum optics for their indispensable role in light sources and information processing. During the past decades, the three- and four-wave mixing ($χ^{(2)}$ and $χ^{(3)}$) effects have been extensively studied, especially in the micro-/nano-structures by which the photon-photon interaction strength is greatly enhanced. So far, the high-o… ▽ More Nonlinear optics processes lie at the heart of photonics and quantum optics for their indispensable role in light sources and information processing. During the past decades, the three- and four-wave mixing ($χ^{(2)}$ and $χ^{(3)}$) effects have been extensively studied, especially in the micro-/nano-structures by which the photon-photon interaction strength is greatly enhanced. So far, the high-order nonlinearity beyond the $χ^{(3)}$ has rarely been studied in dielectric materials due to their weak intrinsic nonlinear susceptibility, even in high-quality microcavities. Here, an effective five-wave mixing process ($χ^{(4)}$) is synthesized for the first time, by incorporating $χ^{(2)}$ and $χ^{(3)}$ processes in a single microcavity. The coherence of the synthetic $χ^{(4)}$ is verified by generating time-energy entangled visible-telecom photon-pairs, which requires only one drive laser at the telecom waveband. The photon pair generation rate from the synthetic process shows an enhancement factor over $500$ times upon intrinsic five-wave mixing. Our work demonstrates a universal approach of nonlinear synthesis via photonic structure engineering at the mesoscopic scale rather than material engineering, and thus opens a new avenue for realizing high-order optical nonlinearities and exploring novel functional photonic devices. △ Less

Submitted 3 April, 2022; originally announced April 2022.

Comments: 4 figures

Fano-like resonance due to interference with distant transitions

Authors: Y. -N. Lv, A. -W. Liu, Y. Tan, C. -L. Hu, T. -P. Hua, X. -B. Zou, Y. R. Sun, C. -L. Zou, G. -C. Guo, S. -M. Hu

Abstract: Narrow optical resonances of atoms or molecules have immense significance in various precision measurements, such as testing fundamental physics and the generation of primary frequency standards. In these studies, accurate transition centers derived from fitting the measured spectra are demanded, which critically rely on the knowledge of spectral line profiles. Here, we propose a new mechanism of… ▽ More Narrow optical resonances of atoms or molecules have immense significance in various precision measurements, such as testing fundamental physics and the generation of primary frequency standards. In these studies, accurate transition centers derived from fitting the measured spectra are demanded, which critically rely on the knowledge of spectral line profiles. Here, we propose a new mechanism of Fano-like resonance induced by distant discrete levels %in atoms or molecules and experimentally verify it with Doppler-free spectroscopy of vibration-rotational transitions of CO$_2$. The observed spectrum has an asymmetric profile and its amplitude increases quadratically with the probe laser power. Our results facilitate a broad range of topics based on narrow transitions. %, such as optical clocks, determination of fundamental physical constants, and quantum memory. △ Less

Submitted 12 October, 2022; v1 submitted 23 March, 2022; originally announced March 2022.

Comments: 6 pages, 4 figures

Journal ref: Physical Review Letters, 129: 163201 (2022)

Experimental Quantum End-to-End Learning on a Superconducting Processor

Authors: Xiaoxuan Pan, Xi Cao, Weiting Wang, Ziyue Hua, Weizhou Cai, Xuegang Li, Haiyan Wang, Jiaqi Hu, Yipu Song, Dong-Ling Deng, Chang-Ling Zou, Re-Bing Wu, Luyan Sun

Abstract: Machine learning can be substantially powered by a quantum computer owing to its huge Hilbert space and inherent quantum parallelism. In the pursuit of quantum advantages for machine learning with noisy intermediate-scale quantum devices, it was proposed that the learning model can be designed in an end-to-end fashion, i.e., the quantum ansatz is parameterized by directly manipulable control pulse… ▽ More Machine learning can be substantially powered by a quantum computer owing to its huge Hilbert space and inherent quantum parallelism. In the pursuit of quantum advantages for machine learning with noisy intermediate-scale quantum devices, it was proposed that the learning model can be designed in an end-to-end fashion, i.e., the quantum ansatz is parameterized by directly manipulable control pulses without circuit design and compilation. Such gate-free models are hardware friendly and can fully exploit limited quantum resources. Here, we report the first experimental realization of quantum end-to-end machine learning on a superconducting processor. The trained model can achieve 98% recognition accuracy for two handwritten digits (via two qubits) and 89% for four digits (via three qubits) in the MNIST (Mixed National Institute of Standards and Technology) database. The experimental results exhibit the great potential of quantum end-to-end learning for resolving complex real-world tasks when more qubits are available. △ Less

Submitted 17 March, 2022; originally announced March 2022.

Comments: 6 pages, 4 figures and supplement

MSC Class: 81P05

Break the efficiency limitations of dissipative Kerr soliton using nonlinear couplers

Authors: Ming Li, Xiao-Xiao Xue, Yan-Lei Zhang, Xin-Biao Xu, Chun-Hua Dong, Guang-Can Guo, Chang-Ling Zou

Abstract: Dissipative Kerr soliton (DKS) offers a compact solution of coherent comb sources and holds huge potential for applications, but has long been suffering from poor power conversion efficiency when driving by a continuous-wave laser. Here, a general approach to resolving this challenge is provided. By deriving the critical coupling condition of a multimode nonlinear optics system in a generalized th… ▽ More Dissipative Kerr soliton (DKS) offers a compact solution of coherent comb sources and holds huge potential for applications, but has long been suffering from poor power conversion efficiency when driving by a continuous-wave laser. Here, a general approach to resolving this challenge is provided. By deriving the critical coupling condition of a multimode nonlinear optics system in a generalized theoretical framework, two efficiency limitations of the conventional pump method of DKS are revealed: the effective coupling rate is too small and is also power-dependent. Nonlinear couplers are proposed to sustain the DKS indirectly through nonlinear energy conversion processes, realizing a power-adaptive effective coupling rate to the DKS and matching the total dissipation rate of the system, which promises near-unity power conversion efficiencies. For instance, a conversion efficiency exceeding $90\:\%$ is predicted for aluminum nitride microrings with a nonlinear coupler utilizing second-harmonic generation. The nonlinear coupler approach for high-efficiency generation of DKS is experimentally feasible as its mechanism applies to various nonlinear processes, including Raman and Brillouin scattering, and thus paves the way of micro-solitons towards practical applications. △ Less

Submitted 16 March, 2022; originally announced March 2022.

Comments: 8 pages, 3 figures

Journal ref: Science China Physics, Mechanics & Astronomy 67, 234211 (2024)

High-frequency traveling-wave phononic cavity with sub-micron wavelength

Authors: Xin-Biao Xu, Jia-Qi Wang, Yuan-Hao Yang, Weiting Wang, Yan-Lei Zhang, Bao-Zhen Wang, Chun-Hua Dong, Luyan Sun, Guang-Can Guo, Chang-Ling Zou

Abstract: Thin-film gallium nitride (GaN) as a proven piezoelectric material is a promising platform for the phononic integrated circuits, which hold great potential for scalable information processing processors. Here, an unsuspended traveling phononic resonator based on high-acoustic-index-contrast mechanism is realized in GaN-on-Sapphire with a frequency up to 5 GHz, which matches the typical superconduc… ▽ More Thin-film gallium nitride (GaN) as a proven piezoelectric material is a promising platform for the phononic integrated circuits, which hold great potential for scalable information processing processors. Here, an unsuspended traveling phononic resonator based on high-acoustic-index-contrast mechanism is realized in GaN-on-Sapphire with a frequency up to 5 GHz, which matches the typical superconducting qubit frequency. A sixfold increment in quality factor was found when temperature decreases from room temperature ($Q=5000$) to $7\,\mathrm{K}$ ($Q=30000$) and thus a frequency-quality factor product of $1.5\times10^{14}$ is obtained. Higher quality factors are available when the fabrication process is further optimized. Our system shows great potential in hybrid quantum devices via circuit quantum acoustodynamics. △ Less

Submitted 15 February, 2022; originally announced February 2022.

Adiabatic conversion between gigahertz quasi-Rayleigh and quasi-Love modes for phononic integrated circuits

Authors: Bao-Zhen Wang, Xin-Biao Xu, Yan-Lei Zhang, Weiting Wang, Luyan Sun, Guang-Can Guo, Chang-Ling Zou

Abstract: Unsuspended phononic integrated circuits have been proposed for on-chip acoustic information processing. Limited by the operation mechanism of a conventional interdigital transducer, the excitation of the quasi-Love mode in GaN-on-Sapphire is inefficient and thus a high-efficiency Rayleigh-to-Love mode converter is of great significance for future integrated phononic devices. Here, we propose a hi… ▽ More Unsuspended phononic integrated circuits have been proposed for on-chip acoustic information processing. Limited by the operation mechanism of a conventional interdigital transducer, the excitation of the quasi-Love mode in GaN-on-Sapphire is inefficient and thus a high-efficiency Rayleigh-to-Love mode converter is of great significance for future integrated phononic devices. Here, we propose a high-efficiency and robust phononic mode converter based on an adiabatic conversion mechanism. Utilizing the anisotropic elastic property of the substrate, the adiabatic mode converter is realized by a simple tapered phononic waveguide. A conversion efficiency exceeds $98\%$ with a $3\,\mathrm{dB}$ bandwidth of $1.7\,\mathrm{GHz}$ can be realized for phononic waveguides working at GHz frequency band, and excellent tolerance to the fabrication errors is also numerically validated. The device that we proposed can be useful in both classical and quantum phononic information processing, and the adiabatic mechanism could be generalized to other phononic device designs. △ Less

Submitted 14 February, 2022; originally announced February 2022.

Attention-based Deep Neural Networks for Battery Discharge Capacity Forecasting

Authors: Yadong Zhang, Chenye Zou, Xin Chen

Abstract: Battery discharge capacity forecasting is critically essential for the applications of lithium-ion batteries. The capacity degeneration can be treated as the memory of the initial battery state of charge from the data point of view. The streaming sensor data collected by battery management systems (BMS) reflect the usable battery capacity degradation rates under various operational working conditi… ▽ More Battery discharge capacity forecasting is critically essential for the applications of lithium-ion batteries. The capacity degeneration can be treated as the memory of the initial battery state of charge from the data point of view. The streaming sensor data collected by battery management systems (BMS) reflect the usable battery capacity degradation rates under various operational working conditions. The battery capacity in different cycles can be measured with the temporal patterns extracted from the streaming sensor data based on the attention mechanism. The attention-based similarity regarding the first cycle can describe the battery capacity degradation in the following cycles. The deep degradation network (DDN) is developed with the attention mechanism to measure similarity and predict battery capacity. The DDN model can extract the degeneration-related temporal patterns from the streaming sensor data and perform the battery capacity prediction efficiently online in real-time. Based on the MIT-Stanford open-access battery aging dataset, the root-mean-square error of the capacity estimation is 1.3 mAh. The mean absolute percentage error of the proposed DDN model is 0.06{\%}. The DDN model also performance well in the Oxford Battery Degradation Dataset with dynamic load profiles. Therefore, the high accuracy and strong robustness of the proposed algorithm are verified. △ Less

Submitted 14 February, 2022; originally announced February 2022.

Existence, uniqueness and stability of steady vortex rings of small cross-section

Authors: Daomin Cao, Guolin Qin, Weilin Yu, Weicheng Zhan, Changjun Zou

Abstract: This paper is concerned with steady vortex rings in an ideal fluid of uniform density, which are special global axi-symmetric solutions of the three-dimensional incompressible Euler equation. We systematically establish the existence, uniqueness and nonlinear orbital stability of steady vortex rings of small cross-section for which the potential vorticity is constant throughout the core. The latte… ▽ More This paper is concerned with steady vortex rings in an ideal fluid of uniform density, which are special global axi-symmetric solutions of the three-dimensional incompressible Euler equation. We systematically establish the existence, uniqueness and nonlinear orbital stability of steady vortex rings of small cross-section for which the potential vorticity is constant throughout the core. The latter two answer a long-standing question since the pioneering work of Fraenkel and Berger \cite{BF1} (Acta Math., 1974). To achieve our goal, we rescale the Stokes stream function of vortex ring by its cross-section radius, and expand it at the well-known Rankine vortex using Taylor's formula, where the estimates for coefficients are obtained by a decomposition according to Green's function and local Pohozaev identities. The main observations are: The stream function is even and has a translational invariance in $z$-direction; the zero point for the $r$-coefficient of linear term in its expansion determines the asymptotic location of vortex ring, which appears as the condition to eliminate the degenerate direction in Lyapunov-Schmidt reduction argument for existence; the non-vanishing condition of the second order $r$-coefficient at foresaid zero point is verified as one of the essential factors for uniqueness, while the negativity means that these vortex rings maximizes the functional composed of kinetic energy and impulse. By applying the Arnol'd's dual variational principle together with the uniqueness result, we are finally able to prove the nonlinear orbital stability of thin vortex rings. This result gives a large class of stable vortex rings supported on topological tori, which is different from Hill's spherical vortex discussed by Choi \cite{Choi20} (Comm. Pure Appl. Math., 2023). △ Less

Submitted 5 December, 2023; v1 submitted 20 January, 2022; originally announced January 2022.

Comments: 80 pages

MSC Class: 76B47

Improving Human-Object Interaction Detection via Phrase Learning and Label Composition

Authors: Zhimin Li, Cheng Zou, Yu Zhao, Boxun Li, Sheng Zhong

Abstract: Human-Object Interaction (HOI) detection is a fundamental task in high-level human-centric scene understanding. We propose PhraseHOI, containing a HOI branch and a novel phrase branch, to leverage language prior and improve relation expression. Specifically, the phrase branch is supervised by semantic embeddings, whose ground truths are automatically converted from the original HOI annotations wit… ▽ More Human-Object Interaction (HOI) detection is a fundamental task in high-level human-centric scene understanding. We propose PhraseHOI, containing a HOI branch and a novel phrase branch, to leverage language prior and improve relation expression. Specifically, the phrase branch is supervised by semantic embeddings, whose ground truths are automatically converted from the original HOI annotations without extra human efforts. Meanwhile, a novel label composition method is proposed to deal with the long-tailed problem in HOI, which composites novel phrase labels by semantic neighbors. Further, to optimize the phrase branch, a loss composed of a distilling loss and a balanced triplet loss is proposed. Extensive experiments are conducted to prove the effectiveness of the proposed PhraseHOI, which achieves significant improvement over the baseline and surpasses previous state-of-the-art methods on Full and NonRare on the challenging HICO-DET benchmark. △ Less

Submitted 15 January, 2022; v1 submitted 14 December, 2021; originally announced December 2021.

Comments: Accepted to AAAI2022