arXiv subjects

Liming Zhao

Publications and source records attributed to Liming Zhao.

28 records · Page 2Linked to original sources

arXiv2019

Variational Quantum Circuit Model for Knowledge Graphs Embedding

In this work, we propose the first quantum Ansätze for the statistical relational learning on knowledge graphs using parametric quantum circuits. We introduce two types of variational quantum circuits for knowledge graph embedding. Inspired by the classical representation learning, we first consider latent features for entities as coefficients of quantum states, while predicates are characterized by parametric gates acting on the quantum states. For the first model, the quantum advantages disappear when it comes to the optimization of this model. Therefore, we introduce a second quantum circuit model where embeddings of entities are generated from parameterized quantum gates acting on the pure quantum state. The benefit of the second method is that the quantum embeddings can be trained efficiently meanwhile preserving the quantum advantages. We show the proposed methods can achieve comparable results to the state-of-the-art classical models, e.g., RESCAL, DistMult. Furthermore, after optimizing the models, the complexity of inductive inference on the knowledge graphs might be reduced with respect to the number of entities.

quant-ph↗

arXiv2017

Measurement-driven quantum computing: Performance of a 3-SAT solver

We investigate the performance of a quantum algorithm for solving classical 3-SAT problems. A cycle of post-selected measurements drives the computer's register monotonically toward a steady state which is correlated to the classical solution(s). An internal parameter $θ$ determines both the degree of correlation and the success probability, thus controlling the algorithm's runtime. Optionally this parameter can be gradually evolved during the algorithm's execution to create a Zeno-like effect; this can be viewed as an adiabatic evolution of a Hamiltonian which remains frustration-free at all points, and we lower-bound the corresponding gap. In exact numerical simulations of small systems up to 34 qubits our approach competes favourably with a high-performing classical 3-SAT solver, which itself outperforms a brute-force application of Grover's search.

quant-ph↗

arXiv2017

Deeply-Learned Part-Aligned Representations for Person Re-Identification

In this paper, we address the problem of person re-identification, which refers to associating the persons captured from different cameras. We propose a simple yet effective human part-aligned representation for handling the body part misalignment problem. Our approach decomposes the human body into regions (parts) which are discriminative for person matching, accordingly computes the representations over the regions, and aggregates the similarities computed between the corresponding regions of a pair of probe and gallery images as the overall matching score. Our formulation, inspired by attention models, is a deep neural network modeling the three steps together, which is learnt through minimizing the triplet loss function without requiring body part labeling information. Unlike most existing deep learning algorithms that learn a global or spatial partition-based local representation, our approach performs human body partition, and thus is more robust to pose changes and various human spatial distributions in the person bounding box. Our approach shows state-of-the-art results over standard datasets, Market-$1501$, CUHK$03$, CUHK$01$ and VIPeR.

cs.CV↗

arXiv2017

Deep Convolutional Neural Networks with Merge-and-Run Mappings

A deep residual network, built by stacking a sequence of residual blocks, is easy to train, because identity mappings skip residual branches and thus improve information flow. To further reduce the training difficulty, we present a simple network architecture, deep merge-and-run neural networks. The novelty lies in a modularized building block, merge-and-run block, which assembles residual branches in parallel through a merge-and-run mapping: Average the inputs of these residual branches (Merge), and add the average to the output of each residual branch as the input of the subsequent residual branch (Run), respectively. We show that the merge-and-run mapping is a linear idempotent function in which the transformation matrix is idempotent, and thus improves information flow, making training easy. In comparison to residual networks, our networks enjoy compelling advantages: they contain much shorter paths, and the width, i.e., the number of channels, is increased. We evaluate the performance on the standard recognition tasks. Our approach demonstrates consistent improvements over ResNets with the comparable setup, and achieves competitive results (e.g., $3.57\%$ testing error on CIFAR-$10$, $19.00\%$ on CIFAR-$100$, $1.51\%$ on SVHN).

cs.CV↗

arXiv2017

A measurement driven analog of adiabatic quantum computation for frustration-free Hamiltonians

The adiabatic quantum algorithm has drawn intense interest as a potential approach to accelerating optimization tasks using quantum computation. The algorithm is most naturally realised in systems which support Hamiltonian evolution, rather than discrete gates. We explore an alternative approach in which slowly varying measurements are used to mimic adiabatic evolution. We show that for certain Hamiltonians, which remain frustration-free all along the adiabatic path, the necessary measurements can be implemented through the measurement of random terms from the Hamiltonian. This offers a new, and potentially more viable, method of realising adiabatic evolution in gate-based quantum computer architectures.

quant-ph↗

arXiv2017

Computing on quantum shared secrets

A (k,n)-threshold secret-sharing scheme allows for a string to be split into n shares in such a way that any subset of at least k shares suffices to recover the secret string, but such that any subset of at most k-1 shares contains no information about the secret. Quantum secret-sharing schemes extend this idea to the sharing of quantum states. Here we propose a method of performing computation on quantum shared secrets. We introduce a (n,n)-quantum secret sharing scheme together with a set of protocols that allow quantum circuits to be evaluated on the shared secret without the need to decode the secret. We consider a multipartite setting, with each participant holding a share of the secret. We show that if there exists at least one honest participant, no group of dishonest participants can recover any information about the shared secret, independent of their deviations from the protocol.

quant-ph↗

arXiv2016

DeepSaliency: Multi-Task Deep Neural Network Model for Salient Object Detection

A key problem in salient object detection is how to effectively model the semantic properties of salient objects in a data-driven manner. In this paper, we propose a multi-task deep saliency model based on a fully convolutional neural network (FCNN) with global input (whole raw images) and global output (whole saliency maps). In principle, the proposed saliency model takes a data-driven strategy for encoding the underlying saliency prior information, and then sets up a multi-task learning scheme for exploring the intrinsic correlations between saliency detection and semantic image segmentation. Through collaborative feature learning from such two correlated tasks, the shared fully convolutional layers produce effective features for object perception. Moreover, it is capable of capturing the semantic information on salient objects across different levels using the fully convolutional layers, which investigate the feature-sharing properties of salient object detection with great feature redundancy reduction. Finally, we present a graph Laplacian regularized nonlinear regression model for saliency refinement. Experimental results demonstrate the effectiveness of our approach in comparison with the state-of-the-art approaches.

cs.CV↗

arXiv2015

Fast graph operations in quantum computation

The connection between certain entangled states and graphs has been heavily studied in the context of measurement-based quantum computation as a tool for understanding entanglement. Here we show that this correspondence can be harnessed in the reverse direction to yield a graph data structure which allows for more efficient manipulation and comparison of graphs than any possible classical structure. We introduce efficient algorithms for many transformation and comparison operations on graphs represented as graph states, and prove that no classical data structure can have similar performance for the full set of operations studied.

quant-ph↗

arXiv2014

Metric Learning Driven Multi-Task Structured Output Optimization for Robust Keypoint Tracking

As an important and challenging problem in computer vision and graphics, keypoint-based object tracking is typically formulated in a spatio-temporal statistical learning framework. However, most existing keypoint trackers are incapable of effectively modeling and balancing the following three aspects in a simultaneous manner: temporal model coherence across frames, spatial model consistency within frames, and discriminative feature construction. To address this issue, we propose a robust keypoint tracker based on spatio-temporal multi-task structured output optimization driven by discriminative metric learning. Consequently, temporal model coherence is characterized by multi-task structured keypoint model learning over several adjacent frames, while spatial model consistency is modeled by solving a geometric verification based structured learning problem. Discriminative feature construction is enabled by metric learning to ensure the intra-class compactness and inter-class separability. Finally, the above three modules are simultaneously optimized in a joint learning scheme. Experimental results have demonstrated the effectiveness of our tracker.

cs.CV↗

arXiv2012

Genuine Correlations of Tripartite System

We define genuine total, classical and quantum correlations in tripartite systems. The measure we propose is based on the idea that genuine tripartite correlation exists if and only if the correlation between any bipartition does not vanish. We find in a symmetrical tripartite state, for total correlation and classical correlation, the genuine tripartite correlations are no less than pair-wise correlations. However, the genuine quantum tripartite correlation can be surpassed by the pair-wise quantum correlations. Analytical expressions for genuine tripartite correlations are obtained for pure states and rank-2 symmetrical states. The genuine correlations in both entangled and separable states are calculated.

quant-ph↗