arXiv subjects

Li Jing

Publications and source records attributed to Li Jing.

33 records · Page 2Linked to original sources

arXiv2019

Perceptual representations of structural information in images: application to quality assessment of synthesized view in FTV scenario

As the immersive multimedia techniques like Free-viewpoint TV (FTV) develop at an astonishing rate, user's demand for high-quality immersive contents increases dramatically. Unlike traditional uniform artifacts, the distortions within immersive contents could be non-uniform structure-related and thus are challenging for commonly used quality metrics. Recent studies have demonstrated that the representation of visual features can be extracted from multiple levels of the hierarchy. Inspired by the hierarchical representation mechanism in the human visual system (HVS), in this paper, we explore to adopt structural representations to quantitatively measure the impact of such structure-related distortion on perceived quality in FTV scenario. More specifically, a bio-inspired full reference image quality metric is proposed based on 1) low-level contour descriptor; 2) mid-level contour category descriptor; and 3) task-oriented non-natural structure descriptor. The experimental results show that the proposed model outperforms significantly the state-of-the-art metrics.

eess.IV↗

arXiv2019

Migrating Knowledge between Physical Scenarios based on Artificial Neural Networks

Deep learning is known to be data-hungry, which hinders its application in many areas of science when datasets are small. Here, we propose to use transfer learning methods to migrate knowledge between different physical scenarios and significantly improve the prediction accuracy of artificial neural networks trained on a small dataset. This method can help reduce the demand for expensive data by making use of additional inexpensive data. First, we demonstrate that in predicting the transmission from multilayer photonic film, the relative error rate is reduced by 46.8% (26.5%) when the source data comes from 10-layer (8-layer) films and the target data comes from 8-layer (10-layer) films. Second, we show that the relative error rate is decreased by 22% when knowledge is transferred between two very different physical scenarios: transmission from multilayer films and scattering from multilayer nanoparticles. Finally, we propose a multi-task learning method to improve the performance of different physical scenarios simultaneously in which each task only has a small dataset.

cs.CV↗

arXiv2018

WaveletNet: Logarithmic Scale Efficient Convolutional Neural Networks for Edge Devices

We present a logarithmic-scale efficient convolutional neural network architecture for edge devices, named WaveletNet. Our model is based on the well-known depthwise convolution, and on two new layers, which we introduce in this work: a wavelet convolution and a depthwise fast wavelet transform. By breaking the symmetry in channel dimensions and applying a fast algorithm, WaveletNet shrinks the complexity of convolutional blocks by an O(logD/D) factor, where D is the number of channels. Experiments on CIFAR-10 and ImageNet classification show superior and comparable performances of WaveletNet compared to state-of-the-art models such as MobileNetV2.

cs.LG↗

arXiv2017

Rotational Unit of Memory

The concepts of unitary evolution matrices and associative memory have boosted the field of Recurrent Neural Networks (RNN) to state-of-the-art performance in a variety of sequential tasks. However, RNN still have a limited capacity to manipulate long-term memory. To bypass this weakness the most successful applications of RNN use external techniques such as attention mechanisms. In this paper we propose a novel RNN model that unifies the state-of-the-art approaches: Rotational Unit of Memory (RUM). The core of RUM is its rotational operation, which is, naturally, a unitary matrix, providing architectures with the power to learn long-term dependencies by overcoming the vanishing and exploding gradients problem. Moreover, the rotational unit also serves as associative memory. We evaluate our model on synthetic memorization, question answering and language modeling tasks. RUM learns the Copying Memory task completely and improves the state-of-the-art result in the Recall task. RUM's performance in the bAbI Question Answering task is comparable to that of models with attention mechanism. We also improve the state-of-the-art result to 1.189 bits-per-character (BPC) loss in the Character Level Penn Treebank (PTB) task, which is to signify the applications of RUM to real-world sequential data. The universality of our construction, at the core of RNN, establishes RUM as a promising approach to language modeling, speech recognition and machine translation.

cs.LG↗

arXiv2017

Gated Orthogonal Recurrent Units: On Learning to Forget

We present a novel recurrent neural network (RNN) based model that combines the remembering ability of unitary RNNs with the ability of gated RNNs to effectively forget redundant/irrelevant information in its memory. We achieve this by extending unitary RNNs with a gating mechanism. Our model is able to outperform LSTMs, GRUs and Unitary RNNs on several long-term dependency benchmark tasks. We empirically both show the orthogonal/unitary RNNs lack the ability to forget and also the ability of GORU to simultaneously remember long term dependencies while forgetting irrelevant information. This plays an important role in recurrent neural networks. We provide competitive results along with an analysis of our model on many natural sequential tasks including the bAbI Question Answering, TIMIT speech spectrum prediction, Penn TreeBank, and synthetic tasks that involve long-term dependencies such as algorithmic, parenthesis, denoising and copying tasks.

cs.LG↗

arXiv2017

Nanophotonic Particle Simulation and Inverse Design Using Artificial Neural Networks

We propose a method to use artificial neural networks to approximate light scattering by multilayer nanoparticles. We find the network needs to be trained on only a small sampling of the data in order to approximate the simulation to high precision. Once the neural network is trained, it can simulate such optical processes orders of magnitude faster than conventional simulations. Furthermore, the trained neural network can be used solve nanophotonic inverse design problems by using back- propogation - where the gradient is analytical, not numerical.

physics.comp-ph↗

arXiv2017

Tunable Efficient Unitary Neural Networks (EUNN) and their application to RNNs

Using unitary (instead of general) matrices in artificial neural networks (ANNs) is a promising way to solve the gradient explosion/vanishing problem, as well as to enable ANNs to learn long-term correlations in the data. This approach appears particularly promising for Recurrent Neural Networks (RNNs). In this work, we present a new architecture for implementing an Efficient Unitary Neural Network (EUNNs); its main advantages can be summarized as follows. Firstly, the representation capacity of the unitary space in an EUNN is fully tunable, ranging from a subspace of SU(N) to the entire unitary space. Secondly, the computational complexity for training an EUNN is merely $\mathcal{O}(1)$ per parameter. Finally, we test the performance of EUNNs on the standard copying task, the pixel-permuted MNIST digit recognition benchmark as well as the Speech Prediction Test (TIMIT). We find that our architecture significantly outperforms both other state-of-the-art unitary RNNs and the LSTM architecture, in terms of the final performance and/or the wall-clock training speed. EUNNs are thus promising alternatives to RNNs and LSTMs for a wide variety of applications.

cs.LG↗

arXiv2014

Quantum Cloning Machines and the Applications

No-cloning theorem is fundamental for quantum mechanics and for quantum information science that states an unknown quantum state cannot be cloned perfectly. However, we can try to clone a quantum state approximately with the optimal fidelity, or instead, we can try to clone it perfectly with the largest probability. Thus various quantum cloning machines have been designed for different quantum information protocols. Specifically, quantum cloning machines can be designed to analyze the security of quantum key distribution protocols such as BB84 protocol, six-state protocol, B92 protocol and their generalizations. Some well-known quantum cloning machines include universal quantum cloning machine, phase-covariant cloning machine, the asymmetric quantum cloning machine and the probabilistic quantum cloning machine etc. In the past years, much progress has been made in studying quantum cloning machines and their applications and implementations, both theoretically and experimentally. In this review, we will give a complete description of those important developments about quantum cloning and some related topics. On the other hand, this review is self-consistent, and in particular, we try to present some detailed formulations so that further study can be taken based on those results.

quant-ph↗

arXiv2014

Fitting magnetic field gradient with Heisenberg-scaling accuracy

We propose a quantum fitting scheme to estimate the magnetic field gradient with $N$-atom spins preparing in W state, which attains the Heisenberg-scaling accuracy. Our scheme combines the quantum multi-parameter estimation and the least square linear fitting method to achieve the quantum Cramér-Rao bound (QCRB). We show that the estimated quantity achieves the Heisenberg-scaling accuracy. In single parameter estimation with assumption that the magnetic field is strictly linear, two optimal measurements can achieve the identical Heisenberg-scaling accuracy. Proper interpretation of the super-Heisenberg-scaling accuracy is presented. The scheme of quantum metrology combined with data fitting provides a new method in fast high precision measurements.

quant-ph↗

arXiv2013

Quantum key distribution based on a quantum retrodiction protocol

This paper is withdrawn. We study the quantum key distribution (QKD) protocol based on a quantum retrodiction protocol, namely the so-called mean king problem. The security is analyzed by considering the eavesdropping on both the preparation of the entangled pair and the transmission of the quantum state. This QKD protocol can generate efficiently a bit of raw key in every single run. We find that, for qubit system, it is more secure than the QKD Bennett-Brassard 1984 protocol and the six-state protocol. This QKD protocol works also for higher dimensional system.

quant-ph↗

arXiv2012

Quantum network teleportation for quantum information distribution and concentration

We investigate the schemes of quantum network teleportation for quantum information distribution and concentration which are essential in quantum cloud computation and quantum internet. In those schemes, the cloud can send simultaneously identical unknown quantum states to clients located in different places by a network like teleportation with a prior shared multipartite entangled state resource. The cloud first perform the quantum operation, each client can recover their quantum state locally by using the classical information announced by the cloud about the measurement result. The number of clients can be beyond the number of identical quantum states intentionally being sent, this quantum network teleportation can make sure that the retrieved quantum state is optimal. Furthermore, we present a scheme to realize its reverse process, which concentrates the states from the clients to reconstruct the original state of the cloud. These schemes facilitate the quantum information distribution and concentration in quantum networks in the framework of quantum cloud computation. Potential applications in time synchronization are discussed.

quant-ph↗

arXiv2012

Minimal sets determining universal and phase-covariant quantum cloning

We study the minimal input sets which can determine completely the universal and the phase-covariant quantum cloning machines. We find that the universal quantum cloning machine, which can copy arbitrary input qubit equally well, however can be determined completely by only four input states located at the four vertices of a tetrahedron. The phase-covariant quantum cloning machine, which can copy all qubits located on the equator of the Bloch sphere, can be determined by three equatorial qubits with equal angular distance. These results sharpen further the well-known results that BB84 states and six-states used in quantum cryptography can determine completely the phase-covariant and universal quantum cloning machines. This concludes the study of the power of universal and phase-covariant quantum cloning, i.e., from minimal input sets necessarily to full input sets by definition. This can simplify dramatically the testing of whether the quantum clone machines are successful or not, we only need to check that the minimal input sets can be cloned optimally.

quant-ph↗

arXiv2012

General Quantum Key Distribution in Higher Dimension

We study a general quantum key distribution protocol in higher dimension. In this protocol, quantum states in arbitrary $g+1$ ($1\le g\le d$) out of all $d+1$ mutually unbiased bases in a d-dimensional system can be used for the key encoding. This provides a natural generalization of the quantum key distribution in higher dimension and recovers the previously known results for $g=1$ and $d$. In our investigation, we study Eve's attack by two slightly different approaches. One is considering the optimal cloner for Eve, and the other, defined as the optimal attack, is maximizing Eve's information. We derive results for both approaches and show the deviation of the optimal cloner from the optimal attack. With our systematic investigation of the quantum key distribution protocols in higher dimension, one may balance the security gain and the implementation cost by changing the number of bases in the key encoding. As a side product, we also prove the equivalency between the optimal phase covariant quantum cloning machine and the optimal cloner for the $g=d-1$ quantum key distribution.

quant-ph↗

arXiv2011

No-compressing of quantum phase information

We raise a general question of quantum information theory whether the quantum phase information can be compressed and retrieved. A general qubit contains both amplitude and phase information, while an equatorial qubit contains only a phase information. We study whether it is possible to compress the phase information of n equatorial qubits into m general qubits with m being less than n, and still those information can be retrieved perfectly. We prove that this process is not allowed by quantum mechanics.

quant-ph↗

arXiv2011

Unified Universal Quantum Cloning Machine and Fidelities

We present a unified universal quantum cloning machine, which combines several different existing universal cloning machines together including the asymmetric case. In this unified framework, the identical pure states are projected equally into each copy initially constituted by input and one half of the maximally entangled states. We show explicitly that the output states of those universal cloning machines are the same. One importance of this unified cloning machine is that the cloning procession is always the symmetric projection which reduces dramatically the difficulties for implementation. Also it is found that this unified cloning machine can be directly modified to the general asymmetric case. Besides the global fidelity and the single-copy fidelity, we also present all possible arbitrary-copy fidelities.

quant-ph↗