arXiv subjects

Chao Li

Publications and source records attributed to Chao Li.

At least 523 records · Page 29Linked to original sources

arXiv2019

Tensor-Ring Nuclear Norm Minimization and Application for Visual Data Completion

Tensor ring (TR) decomposition has been successfully used to obtain the state-of-the-art performance in the visual data completion problem. However, the existing TR-based completion methods are severely non-convex and computationally demanding. In addition, the determination of the optimal TR rank is a tough work in practice. To overcome these drawbacks, we first introduce a class of new tensor nuclear norms by using tensor circular unfolding. Then we theoretically establish connection between the rank of the circularly-unfolded matrices and the TR ranks. We also develop an efficient tensor completion algorithm by minimizing the proposed tensor nuclear norm. Extensive experimental results demonstrate that our proposed tensor completion method outperforms the conventional tensor completion methods in the image/video in-painting problem with striped missing values.

cs.CV↗

arXiv2019

Predicting Research Trends From Arxiv

We perform trend detection on two datasets of Arxiv papers, derived from its machine learning (cs.LG) and natural language processing (cs.CL) categories. Our approach is bottom-up: we first rank papers by their normalized citation counts, then group top-ranked papers into different categories based on the tasks that they pursue and the methods they use. We then analyze these resulting topics. We find that the dominating paradigm in cs.CL revolves around natural language generation problems and those in cs.LG revolve around reinforcement learning and adversarial principles. By extrapolation, we predict that these topics will remain lead problems/approaches in their fields in the short- and mid-term.

cs.CL↗

arXiv2019

Coupled CycleGAN: Unsupervised Hashing Network for Cross-Modal Retrieval

In recent years, hashing has attracted more and more attention owing to its superior capacity of low storage cost and high query efficiency in large-scale cross-modal retrieval. Benefiting from deep leaning, continuously compelling results in cross-modal retrieval community have been achieved. However, existing deep cross-modal hashing methods either rely on amounts of labeled information or have no ability to learn an accuracy correlation between different modalities. In this paper, we proposed Unsupervised coupled Cycle generative adversarial Hashing networks (UCH), for cross-modal retrieval, where outer-cycle network is used to learn powerful common representation, and inner-cycle network is explained to generate reliable hash codes. Specifically, our proposed UCH seamlessly couples these two networks with generative adversarial mechanism, which can be optimized simultaneously to learn representation and hash codes. Extensive experiments on three popular benchmark datasets show that the proposed UCH outperforms the state-of-the-art unsupervised cross-modal hashing methods.

cs.IR↗

arXiv2019

Semantic Adversarial Network with Multi-scale Pyramid Attention for Video Classification

Two-stream architecture have shown strong performance in video classification task. The key idea is to learn spatio-temporal features by fusing convolutional networks spatially and temporally. However, there are some problems within such architecture. First, it relies on optical flow to model temporal information, which are often expensive to compute and store. Second, it has limited ability to capture details and local context information for video data. Third, it lacks explicit semantic guidance that greatly decrease the classification performance. In this paper, we proposed a new two-stream based deep framework for video classification to discover spatial and temporal information only from RGB frames, moreover, the multi-scale pyramid attention (MPA) layer and the semantic adversarial learning (SAL) module is introduced and integrated in our framework. The MPA enables the network capturing global and local feature to generate a comprehensive representation for video, and the SAL can make this representation gradually approximate to the real video semantics in an adversarial manner. Experimental results on two public benchmarks demonstrate our proposed methods achieves state-of-the-art results on standard video datasets.

cs.CV↗

arXiv2019

Collaborative Spatio-temporal Feature Learning for Video Action Recognition

Spatio-temporal feature learning is of central importance for action recognition in videos. Existing deep neural network models either learn spatial and temporal features independently (C2D) or jointly with unconstrained parameters (C3D). In this paper, we propose a novel neural operation which encodes spatio-temporal features collaboratively by imposing a weight-sharing constraint on the learnable parameters. In particular, we perform 2D convolution along three orthogonal views of volumetric video data,which learns spatial appearance and temporal motion cues respectively. By sharing the convolution kernels of different views, spatial and temporal features are collaboratively learned and thus benefit from each other. The complementary features are subsequently fused by a weighted summation whose coefficients are learned end-to-end. Our approach achieves state-of-the-art performance on large-scale benchmarks and won the 1st place in the Moments in Time Challenge 2018. Moreover, based on the learned coefficients of different views, we are able to quantify the contributions of spatial and temporal features. This analysis sheds light on interpretability of the model and may also guide the future design of algorithm for video recognition.

cs.CV↗

arXiv2019

Numerically flat holomorphic bundles over non Kähler manifolds

In this paper, we study numerically flat holomorphic vector bundles over a compact non-Kähler manifold $(X, ω)$ with the Hermitian metric $ω$ satisfying the Gauduchon and Astheno-Kähler conditions. We prove that numerically flatness is equivalent to numerically effectiveness with vanishing first Chern number, semistablity with vanishing first and second Chern numbers, approximate Hermitian flatness and the existence of a filtration whose quotients are Hermitian flat. This gives an affirmative answer to the question proposed by Demailly, Peternell and Schneider.

math.DG↗

arXiv2019

Scalable and Privacy-preserving Design of On/Off-chain Smart Contracts

The rise of smart contract systems such as Ethereum has resulted in a proliferation of blockchain-based decentralized applications including applications that store and manage a wide range of data. Current smart contracts are designed to be executed solely by miners and are revealed entirely on-chain, resulting in reduced scalability and privacy. In this paper, we discuss that scalability and privacy of smart contracts can be enhanced by splitting a given contract into an off-chain contract and an on-chain contract. Specifically, functions of the contract that involve high-cost computation or sensitive information can be split and included as the off-chain contract, that is signed and executed by only the interested participants. The proposed approach allows the participants to reach unanimous agreement off-chain when all of them are honest, allowing computing resources of miners to be saved and content of the off-chain contract to be hidden from the public. In case of a dispute caused by any dishonest participants, a signed copy of the off-chain contract can be revealed so that a verified instance can be created to make miners enforce the true execution result. Thus, honest participants have the ability to redress and penalize any fraudulent or dishonest behavior, which incentivizes all participants to honestly follow the agreed off-chain contract. We discuss techniques for splitting a contract into a pair of on/off-chain contracts and propose a mechanism to address the challenges of handling dishonest participants in the system. Our implementation and evaluation of the proposed approach using an example smart contract demonstrate the effectiveness of the proposed approach in Ethereum.

cs.CR↗

arXiv2019

Decentralized Privacy-preserving Timed Execution in Blockchain-based Smart Contract Platforms

In the age of Big Data, enabling task scheduling while protecting users' privacy is critical for various decentralized applications in blockchain-based smart contract platforms. Such a privacy-preserving task scheduler requires the task input data to be secretly maintained until a prescribed task execution time and be automatically recorded into the blockchain to enabling the execution of the task at the execution time, even if the user goes offline. While straight-forward centralized approaches provide a basic solution to the problem, unfortunately they are limited to a single point of trust and involve a single point of control. This paper presents decentralized techniques for supporting privacy-preserving task scheduling using smart contracts in Ethereum blockchain networks. We design a privacy-preserving task scheduling protocol that is managed by a manager smart contract. The protocol requires a user to schedule a task by deploying a proxy smart contract maintaining the non-sensitive information of the task while creating decentralized secret trust and selecting trustees from the network to maintain the sensitive information of the task. With security techniques including secret sharing and layered encryption as well as security deposit paid by trustees as economic deterrence, the protocol can protect the sensitive information against possible attacks including some trustees destroying the sensitive information (drop attack) or secretly releasing the sensitive information before the execution time (release-ahead attack). We demonstrate the attack-resilience of the proposed protocol through rigorous analysis.Our implementation and experimental evaluation on the Ethereum official test network demonstrate the low monetary cost and the low time overhead associated with the proposed approach.

cs.CR↗

arXiv2019

Decentralized Release of Self-emerging Data using Smart Contracts

In the age of Big Data, releasing protected sensitive data at a future point in time is critical for various applications. Such self-emerging data release requires the data to be protected until a prescribed data release time and be automatically released to the recipient at the release time, even if the data sender goes offline. While straight-forward centralized approaches provide a basic solution to the problem, unfortunately they are limited to a single point of trust and involve a single point of control. This paper presents decentralized techniques for supporting self-emerging data using smart contracts in Ethereum blockchain networks. We design a credible and enforceable smart contract for supporting self-emerging data release. The smart contract employs a set of Ethereum peers to jointly follow the proposed timed-release service protocol allowing the participating peers to earn the remuneration paid by the service users.We model the problem as an extensive-form game with imperfect information to protect against possible adversarial attacks including some peers destroying the private data (drop attack) or secretly releasing the private data before the release time (release-ahead attack). We demonstrate the efficacy and attack-resilience of the proposed techniques through rigorous analysis and experimental evaluation. Our implementation and experimental evaluation on the Ethereum official test network demonstrate the low monetary cost and the low time overhead associated with the proposed approach and validate its guaranteed security properties.

cs.CR↗

arXiv2019

A kilo-Ampere level HTS flux pump

This paper reports a newly developed high current transformer-rectifier High-Tc Superconducting (HTS) flux pump switched by dynamic resistance. A quasi-persistent current of over 1.1 kA has been achieved at 77 K using the device, which is the highest reported operating current by any HTS flux pumps to date. The size of the device is much smaller than traditional current leads and power supplies at the same current level. Parallel YBCO coated conductors are used in the transformer secondary winding as well as in the superconducting load coil to achieve high current. The output current is limited by the critical current of the load rather than the flux pump itself. Moreover, at over 1 kA current level, the device can maintain high flux injection accuracy, and the overall flux ripple is less than 0.2 mili-Weber. The work has shown the potential of using the device to operate high field HTS magnets in ultra-high quasi-persistent current mode, thus substantially reducing the inductance, size, weight, and cost of high field magnets, making them more accessible. It also indicates that the device is promising for powering HTS NMR/MRI magnets, in which the requirement for magnetic field satiability is demanding.

physics.app-ph↗

arXiv2019

New Results about the Boomerang Uniformity of Permutation Polynomials

In EUROCRYPT 2018, Cid et al. \cite{BCT2018} introduced a new concept on the cryptographic property of S-boxes: Boomerang Connectivity Table (BCT for short) for evaluating the subtleties of boomerang-style attacks. Very recently, BCT and the boomerang uniformity, the maximum value in BCT, were further studied by Boura and Canteaut \cite{BC2018}. Aiming at providing new insights, we show some new results about BCT and the boomerang uniformity of permutations in terms of theory and experiment in this paper. Firstly, we present an equivalent technique to compute BCT and the boomerang uniformity, which seems to be much simpler than the original definition from \cite{BCT2018}. Secondly, thanks to Carlet's idea \cite{Carlet2018}, we give a characterization of functions $f$ from $\mathbb{F}_{2}^n$ to itself with boomerang uniformity $δ_{f}$ by means of the Walsh transform. Thirdly, by our method, we consider boomerang uniformities of some specific permutations, mainly the ones with low differential uniformity. Finally, we obtain another class of $4$-uniform BCT permutation polynomials over $\mathbb{F}_{2^n}$, which is the first binomial.

cs.CR↗

arXiv2019

Tunable Thermal Energy Transport across Diamond Membranes and Diamond-Si Interfaces by Nanoscale Graphoepitaxy

The development of electronic devices, especially those that involve heterogeneous integration of materials, has led to increased challenges in addressing their thermal operational-temperature demands. The heat flow in these systems is significantly influenced or even dominated by thermal boundary resistance at interface between dissimilar materials. However, controlling and tuning heat transport across an interface and in the adjacent materials has so far drawn limited attention. In this work, we grow chemical-vapor-deposited (CVD) diamond on silicon substrates by graphoepitaxy and experimentally demonstrate tunable thermal transport across diamond membranes and diamond-silicon interfaces. We observed the highest diamond-silicon thermal boundary conductance (TBC) measured to date and increased diamond thermal conductivity due to strong grain texturing in the diamond near the interface. Additionally, non-equilibrium molecular-dynamics (NEMD) simulations and a Landauer approach are used to understand the diamond-silicon TBC. These findings pave the way for tuning or increasing thermal conductance in heterogeneously integrated electronics that involve polycrystalline materials and will impact applications including electronics thermal management and diamond growth.

cond-mat.mes-hall↗

arXiv2019

Randomized Tensor Ring Decomposition and Its Application to Large-scale Data Reconstruction

Dimensionality reduction is an essential technique for multi-way large-scale data, i.e., tensor. Tensor ring (TR) decomposition has become popular due to its high representation ability and flexibility. However, the traditional TR decomposition algorithms suffer from high computational cost when facing large-scale data. In this paper, taking advantages of the recently proposed tensor random projection method, we propose two TR decomposition algorithms. By employing random projection on every mode of the large-scale tensor, the TR decomposition can be processed at a much smaller scale. The simulation experiment shows that the proposed algorithms are $4-25$ times faster than traditional algorithms without loss of accuracy, and our algorithms show superior performance in deep learning dataset compression and hyperspectral image reconstruction experiments compared to other randomized algorithms.

math.NA↗

arXiv2018

Tensor Ring Decomposition with Rank Minimization on Latent Space: An Efficient Approach for Tensor Completion

In tensor completion tasks, the traditional low-rank tensor decomposition models suffer from the laborious model selection problem due to their high model sensitivity. In particular, for tensor ring (TR) decomposition, the number of model possibilities grows exponentially with the tensor order, which makes it rather challenging to find the optimal TR decomposition. In this paper, by exploiting the low-rank structure of the TR latent space, we propose a novel tensor completion method which is robust to model selection. In contrast to imposing the low-rank constraint on the data space, we introduce nuclear norm regularization on the latent TR factors, resulting in the optimization step using singular value decomposition (SVD) being performed at a much smaller scale. By leveraging the alternating direction method of multipliers (ADMM) scheme, the latent TR factors with optimal rank and the recovered tensor can be obtained simultaneously. Our proposed algorithm is shown to effectively alleviate the burden of TR-rank selection, thereby greatly reducing the computational cost. The extensive experimental results on both synthetic and real-world data demonstrate the superior performance and efficiency of the proposed approach against the state-of-the-art algorithms.

cs.LG↗

arXiv2018

Chat More If You Like: Dynamic Cue Words Planning to Flow Longer Conversations

To build an open-domain multi-turn conversation system is one of the most interesting and challenging tasks in Artificial Intelligence. Many research efforts have been dedicated to building such dialogue systems, yet few shed light on modeling the conversation flow in an ongoing dialogue. Besides, it is common for people to talk about highly relevant aspects during a conversation. And the topics are coherent and drift naturally, which demonstrates the necessity of dialogue flow modeling. To this end, we present the multi-turn cue-words driven conversation system with reinforcement learning method (RLCw), which strives to select an adaptive cue word with the greatest future credit, and therefore improve the quality of generated responses. We introduce a new reward to measure the quality of cue words in terms of effectiveness and relevance. To further optimize the model for long-term conversations, a reinforcement approach is adopted in this paper. Experiments on real-life dataset demonstrate that our model consistently outperforms a set of competitive baselines in terms of simulated turns, diversity and human evaluation.

cs.CL↗

arXiv2018

Low-Rank Embedding of Kernels in Convolutional Neural Networks under Random Shuffling

Although the convolutional neural networks (CNNs) have become popular for various image processing and computer vision task recently, it remains a challenging problem to reduce the storage cost of the parameters for resource-limited platforms. In the previous studies, tensor decomposition (TD) has achieved promising compression performance by embedding the kernel of a convolutional layer into a low-rank subspace. However the employment of TD is naively on the kernel or its specified variants. Unlike the conventional approaches, this paper shows that the kernel can be embedded into more general or even random low-rank subspaces. We demonstrate this by compressing the convolutional layers via randomly-shuffled tensor decomposition (RsTD) for a standard classification task using CIFAR-10. In addition, we analyze how the spatial similarity of the training data influences the low-rank structure of the kernels. The experimental results show that the CNN can be significantly compressed even if the kernels are randomly shuffled. Furthermore, the RsTD-based method yields more stable classification accuracy than the conventional TD-based methods in a large range of compression ratios.

cs.LG↗

arXiv2018

Understanding Phonon Scattering by Nano-Precipitates in Potassium-Doped Lead Chalcogenides

We present a comprehensive experimental and theoretical study of phonon scattering by nanoprecipitates in potassium-doped PbTe, PbSe and PbS. We highlight the role of the precipitate size distribution measured by microscopy, whose tuning allows for thermal conductivities lower than the limit achievable with a single size. The correlation between the size distribution and the contributions to thermal conductivity from phonons in different frequency ranges provides a physical basis to the experimentally measured thermal conductivities, and a criterion to estimate the lowest achievable thermal conductivity. The results have clear implications for efficiency enhancements in nanostructured bulk thermoelectrics.

cond-mat.mtrl-sci↗

arXiv2018

High Performance Visual Tracking with Circular and Structural Operators

In this paper, a novel circular and structural operator tracker (CSOT) is proposed for high performance visual tracking, it not only possesses the powerful discriminative capability of SOSVM but also efficiently inherits the superior computational efficiency of DCF. Based on the proposed circular and structural operators, a set of primal confidence score maps can be obtained by circular correlating feature maps with their corresponding structural correlation filters. Furthermore, an implicit interpolation is applied to convert the multi-resolution feature maps to the continuous domain and make all primal confidence score maps have the same spatial resolution. Then, we exploit an efficient ensemble post-processor based on relative entropy, which can coalesce primal confidence score maps and create an optimal confidence score map for more accurate localization. The target is localized on the peak of the optimal confidence score map. Besides, we introduce a collaborative optimization strategy to update circular and structural operators by iteratively training structural correlation filters, which significantly reduces computational complexity and improves robustness. Experimental results demonstrate that our approach achieves state-of-the-art performance in mean AUC scores of 71.5% and 69.4% on the OTB-2013 and OTB-2015 benchmarks respectively, and obtains a third-best expected average overlap (EAO) score of 29.8% on the VOT-2017 benchmark.

cs.CV↗