arXiv subjects

Hao Tang

Publications and source records attributed to Hao Tang.

At least 505 records · Page 28Linked to original sources

arXiv2018

Quantum Topological Boundary States in Quasi-crystal

Topological phase, a novel and fundamental role in matter, displays an extraordinary robustness to smooth changes in material parameters or disorder. A crossover between topological physics and quantum information may lead to inherent fault-tolerant quantum simulations and quantum computing. Quantum features may be preserved by being encoded among topological structures of physical evolution systems. This requires us to stimulate, manipulate and observe topological phenomena at single quantum particle level, which, however, hasn't been realized yet. Here, we address such a question whether the quantum features of single photons can be preserved in topological structures. We experimentally observe the boundary states of single photons and demonstrate the performance of topological phase on protecting the quantum features in quasi-periodic systems. Our work confirms the compatibility between macroscopic topological states and microscopic single photons on a photonic chip. We believe the emerging 'quantum topological photonics' will add entirely new and versatile capacities into quantum technologies.

quant-ph↗

arXiv2018

Integrated Server for Measurement-Device-Independent Quantum Key Distribution Network

Quantum key distribution (QKD), harnessing quantum physics and optoelectronics, may promise unconditionally secure information exchange in theory. Recently, theoretical and experimental advances in measurement-device-independent (MDI-) QKD have successfully closed the physical backdoor in detection terminals. However, the issues of scalability, stability, cost and loss prevent QKD systems from widespread application in practice. Here, we propose and experimentally demonstrate a solution to build a star-topology quantum access network with an integrated server. By using femtosecond laser direct writing technique, we construct integrated circuits for all the elements of Bell state analyzer together and are able to integrate several such analyzers on single photonic chip. The measured high-visibility Bell state analysis suggests integrated server a promising platform for the practical application of MDI-QKD network.

quant-ph↗

arXiv2018

Direct Observation of Broadband Nonclassical States in a Room-temperature Light-matter Interface

Nonclassical state is an essential resource for quantum-enhanced communication, computing and metrology to outperform their classical counterpart. The nonclassical states that can operate at high bandwidth and room temperature while being compatible with quantum memory are highly desirable to enable the scalability of quantum technologies. Here, we present a direct observation of broadband nonclasscal states in a room-temperature light-matter interface, where the atoms can also be controlled to store and interfere with photons. With a single coupling pulse and far off-resonance configuration, we are able to induce a multi-field interference between light and atoms to create the desired nonclassical states by spectrally selecting the two correlated photons out of seven possible emissions. We explicitly confirm the nonclassicality by observing a cross correlation up to 17 and a violation of Cauchy-Schwarz inequality with 568 standard deviations. Our results demonstrate the potential of a state-built-in, broadband and room-temperature light-matter interface for scalable quantum information networks.

quant-ph↗

arXiv2018

Accurate indoor mapping using an autonomous unmanned aerial vehicle (UAV)

An autonomous indoor aerial vehicle requires reliable simul- taneous localization and mapping (SLAM), accurate flight control, and robust path planning for navigation. This paper presents a system level combination of these existing technologies for 2D navigation. An Unmanned aerial vehicle (UAV) called URSA (Unmanned Recon and Safety Aircraft) that can autonomously flight and mapping indoors environments with an accuracy of 2 cm was developed. Performance in indoor environments was assessed in terms of mapping and navigation precision.

cs.RO↗

arXiv2018

Mapping and Measuring Large-scale Photonic Correlation with Single-photon Imaging

Quantum correlation and its measurement are essential in exploring fundamental quantum physics problems and developing quantum enhanced technologies. Quantum correlation may be generated and manipulated in different spaces, which demands different measurement approaches corresponding to position, time, frequency and polarization of quantum particles. In addition, after early proof-of-principle demonstrations, it is of great demand to measure quantum correlation in a Hilbert space large enough for real quantum applications. When the number of modes goes up to several hundreds, it becomes economically unfeasible for single-mode addressing and also extremely challenging for processing correlation events with hardware. Here we present a general and large-scale measurement approach of Correlation on Spatially-mapped Photon-Level Image (COSPLI). The quantum correlations in other spaces are mapped into the position space and are captured by single-photon-sensitive imaging system. Synthetic methods are developed to suppress noises so that single-photon registrations can be faithfully identified in images. We eventually succeed in retrieving all the correlations with big-data technique from tens of millions of images. We demonstrate our COSPLI by measuring the joint spectrum of parametric down-conversion photons. Our approach provides an elegant way to observe the evolution results of large-scale quantum systems, representing an innovative and powerful tool added into the platform for boosting quantum information processing.

quant-ph↗

arXiv2018

Experimental Machine Learning of Quantum States

Quantum information technologies provide promising applications in communication and computation, while machine learning has become a powerful technique for extracting meaningful structures in 'big data'. A crossover between quantum information and machine learning represents a new interdisciplinary area stimulating progresses in both fields. Traditionally, a quantum state is characterized by quantum state tomography, which is a resource-consuming process when scaled up. Here we experimentally demonstrate a machine-learning approach to construct a quantum-state classifier for identifying the separability of quantum states. We show that it is possible to experimentally train an artificial neural network to efficiently learn and classify quantum states, without the need of obtaining the full information of the states. We also show how adding a hidden layer of neurons to the neural network can significantly boost the performance of the state classifier. These results shed new light on how classification of quantum states can be achieved with limited resources, and represent a step towards machine-learning-based applications in quantum information processing.

quant-ph↗

arXiv2018

Sequence Prediction with Neural Segmental Models

Segments that span contiguous parts of inputs, such as phonemes in speech, named entities in sentences, actions in videos, occur frequently in sequence prediction problems. Segmental models, a class of models that explicitly hypothesizes segments, have allowed the exploration of rich segment features for sequence prediction. However, segmental models suffer from slow decoding, hampering the use of computationally expensive features. In this thesis, we introduce discriminative segmental cascades, a multi-pass inference framework that allows us to improve accuracy by adding higher-order features and neural segmental features while maintaining efficiency. We also show that instead of including more features to obtain better accuracy, segmental cascades can be used to speed up training and decoding. Segmental models, similarly to conventional speech recognizers, are typically trained in multiple stages. In the first stage, a frame classifier is trained with manual alignments, and then in the second stage, segmental models are trained with manual alignments and the out- puts of the frame classifier. However, obtaining manual alignments are time-consuming and expensive. We explore end-to-end training for segmental models with various loss functions, and show how end-to-end training with marginal log loss can eliminate the need for detailed manual alignments. We draw the connections between the marginal log loss and a popular end-to-end training approach called connectionist temporal classification. We present a unifying framework for various end-to-end graph search-based models, such as hidden Markov models, connectionist temporal classification, and segmental models. Finally, we discuss possible extensions of segmental models to large-vocabulary sequence prediction tasks.

cs.CL↗

arXiv2018

A Study of Enhancement, Augmentation, and Autoencoder Methods for Domain Adaptation in Distant Speech Recognition

Speech recognizers trained on close-talking speech do not generalize to distant speech and the word error rate degradation can be as large as 40% absolute. Most studies focus on tackling distant speech recognition as a separate problem, leaving little effort to adapting close-talking speech recognizers to distant speech. In this work, we review several approaches from a domain adaptation perspective. These approaches, including speech enhancement, multi-condition training, data augmentation, and autoencoders, all involve a transformation of the data between domains. We conduct experiments on the AMI data set, where these approaches can be realized under the same controlled setting. These approaches lead to different amounts of improvement under their respective assumptions. The purpose of this paper is to quantify and characterize the performance gap between the two domains, setting up the basis for studying adaptation of speech recognizers from close-talking speech to distant speech. Our results also have implications for improving distant speech recognition.

cs.CL↗

arXiv2018

Unsupervised Adaptation with Interpretable Disentangled Representations for Distant Conversational Speech Recognition

The current trend in automatic speech recognition is to leverage large amounts of labeled data to train supervised neural network models. Unfortunately, obtaining data for a wide range of domains to train robust models can be costly. However, it is relatively inexpensive to collect large amounts of unlabeled data from domains that we want the models to generalize to. In this paper, we propose a novel unsupervised adaptation method that learns to synthesize labeled data for the target domain from unlabeled in-domain data and labeled out-of-domain data. We first learn without supervision an interpretable latent representation of speech that encodes linguistic and nuisance factors (e.g., speaker and channel) using different latent variables. To transform a labeled out-of-domain utterance without altering its transcript, we transform the latent nuisance variables while maintaining the linguistic variables. To demonstrate our approach, we focus on a channel mismatch setting, where the domain of interest is distant conversational speech, and labels are only available for close-talking speech. Our proposed method is evaluated on the AMI dataset, outperforming all baselines and bridging the gap between unadapted and in-domain models by over 77% without using any parallel data.

cs.CL↗

arXiv2018

Experimental Two-dimensional Quantum Walk on a Photonic Chip

Quantum walks, in virtue of the coherent superposition and quantum interference, possess exponential superiority over its classical counterpart in applications of quantum searching and quantum simulation. The quantum enhanced power is highly related to the state space of quantum walks, which can be expanded by enlarging the photon number and/or the dimensions of the evolution network, but the former is considerably challenging due to probabilistic generation of single photons and multiplicative loss. Here we demonstrate a two-dimensional continuous-time quantum walk by using the external geometry of photonic waveguide arrays, rather than the inner degree of freedoms of photons. Using femtosecond laser direct writing, we construct a large-scale three-dimensional structure which forms a two-dimensional lattice with up to 49X49 nodes on a photonic chip. We demonstrate spatial two-dimensional quantum walks using heralded single photons and single-photon-level imaging. We analyze the quantum transport properties via observing the ballistic evolution pattern and the variance profile, which agree well with simulation results. We further reveal the transient nature that is the unique feature for quantum walks of beyond one dimension. An architecture that allows a walk to freely evolve in all directions and a large scale, combining with defect and disorder control, may bring up powerful and versatile quantum walk machines for classically intractable problems.

quant-ph↗

arXiv2018

Entropy of Vaidya Black Hole on Apparent Horizon with Minimal length Revisited

By considering the generalized uncertainty principle, the degrees of freedom near the apparent horizon of Vaidya black hole are calculated with thin film model. The result shows that a cut-off can be introduced naturally rather than taking by hand. Furthermore, if the minimal length is chosen to be a specific value, the statistical entropy will satisfy the conventional area law at the horizon, which might reveal some deep things of the minimal length.

gr-qc↗

arXiv2018

Structured Attention Guided Convolutional Neural Fields for Monocular Depth Estimation

Recent works have shown the benefit of integrating Conditional Random Fields (CRFs) models into deep architectures for improving pixel-level prediction tasks. Following this line of research, in this paper we introduce a novel approach for monocular depth estimation. Similarly to previous works, our method employs a continuous CRF to fuse multi-scale information derived from different layers of a front-end Convolutional Neural Network (CNN). Differently from past works, our approach benefits from a structured attention model which automatically regulates the amount of information transferred between corresponding features at different scales. Importantly, the proposed attention model is seamlessly integrated into the CRF, allowing end-to-end training of the entire architecture. Our extensive experimental evaluation demonstrates the effectiveness of the proposed method which is competitive with previous methods on the KITTI benchmark and outperforms the state of the art on the NYU Depth V2 dataset.

cs.CV↗

arXiv2017

Activation of entanglement from quantum coherence and superposition

Quantum entanglement and coherence are two fundamental features of nature, arising from the superposition principle of quantum mechanics. While considered as puzzling phenomena in the early days of quantum theory, it is only very recently that entanglement and coherence have been recognized as resources for the emerging quantum technologies, including quantum metrology, quantum communication, and quantum computing. In this work we study the limitations for the interconversion between coherence and entanglement. We prove a fundamental no-go theorem, stating that a general resource theory of superposition does not allow for entanglement activation. By constructing a CNOT gate as a free operation, we experimentally show that such activation is possible within the more constrained framework of quantum coherence. Our results provide new insights into the interplay between coherence and entanglement, representing a substantial step forward for solving longstanding open questions in quantum information science.

quant-ph↗

arXiv2017

End-to-End Neural Segmental Models for Speech Recognition

Segmental models are an alternative to frame-based models for sequence prediction, where hypothesized path weights are based on entire segment scores rather than a single frame at a time. Neural segmental models are segmental models that use neural network-based weight functions. Neural segmental models have achieved competitive results for speech recognition, and their end-to-end training has been explored in several studies. In this work, we review neural segmental models, which can be viewed as consisting of a neural network-based acoustic encoder and a finite-state transducer decoder. We study end-to-end segmental models with different weight functions, including ones based on frame-level neural classifiers and on segmental recurrent neural networks. We study how reducing the search space size impacts performance under different weight functions. We also compare several loss functions for end-to-end training. Finally, we explore training approaches, including multi-stage vs. end-to-end training and multitask training that combines segmental and frame-level losses.

cs.CL↗

arXiv2017

Multitask Learning with Low-Level Auxiliary Tasks for Encoder-Decoder Based Speech Recognition

End-to-end training of deep learning-based models allows for implicit learning of intermediate representations based on the final task loss. However, the end-to-end approach ignores the useful domain knowledge encoded in explicit intermediate-level supervision. We hypothesize that using intermediate representations as auxiliary supervision at lower levels of deep networks may be a good way of combining the advantages of end-to-end training and more traditional pipeline approaches. We present experiments on conversational speech recognition where we use lower-level tasks, such as phoneme recognition, in a multitask training approach with an encoder-decoder model for direct character transcription. We compare multiple types of lower-level tasks and analyze the effects of the auxiliary tasks. Our results on the Switchboard corpus show that this approach improves recognition accuracy over a standard encoder-decoder model on the Eval2000 test set.

cs.CL↗

arXiv2016

End-to-End Training Approaches for Discriminative Segmental Models

Recent work on discriminative segmental models has shown that they can achieve competitive speech recognition performance, using features based on deep neural frame classifiers. However, segmental models can be more challenging to train than standard frame-based approaches. While some segmental models have been successfully trained end to end, there is a lack of understanding of their training under different settings and with different losses. We investigate a model class based on recent successful approaches, consisting of a linear model that combines segmental features based on an LSTM frame classifier. Similarly to hybrid HMM-neural network models, segmental models of this class can be trained in two stages (frame classifier training followed by linear segmental model weight training), end to end (joint training of both frame classifier and linear weights), or with end-to-end fine-tuning after two-stage training. We study segmental models trained end to end with hinge loss, log loss, latent hinge loss, and marginal log loss. We consider several losses for the case where training alignments are available as well as where they are not. We find that in general, marginal log loss provides the most consistent strong performance without requiring ground-truth alignments. We also find that training with dropout is very important in obtaining good performance with end-to-end training. Finally, the best results are typically obtained by a combination of two-stage training and fine-tuning.

cs.CL↗

arXiv2016

Lexicon-Free Fingerspelling Recognition from Video: Data, Models, and Signer Adaptation

We study the problem of recognizing video sequences of fingerspelled letters in American Sign Language (ASL). Fingerspelling comprises a significant but relatively understudied part of ASL. Recognizing fingerspelling is challenging for a number of reasons: It involves quick, small motions that are often highly coarticulated; it exhibits significant variation between signers; and there has been a dearth of continuous fingerspelling data collected. In this work we collect and annotate a new data set of continuous fingerspelling videos, compare several types of recognizers, and explore the problem of signer variation. Our best-performing models are segmental (semi-Markov) conditional random fields using deep neural network-based features. In the signer-dependent setting, our recognizers achieve up to about 92% letter accuracy. The multi-signer setting is much more challenging, but with neural network adaptation we achieve up to 83% letter accuracies in this setting.

cs.CL↗

arXiv2016

Discriminative Segmental Cascades for Feature-Rich Phone Recognition

Discriminative segmental models, such as segmental conditional random fields (SCRFs) and segmental structured support vector machines (SSVMs), have had success in speech recognition via both lattice rescoring and first-pass decoding. However, such models suffer from slow decoding, hampering the use of computationally expensive features, such as segment neural networks or other high-order features. A typical solution is to use approximate decoding, either by beam pruning in a single pass or by beam pruning to generate a lattice followed by a second pass. In this work, we study discriminative segmental models trained with a hinge loss (i.e., segmental structured SVMs). We show that beam search is not suitable for learning rescoring models in this approach, though it gives good approximate decoding performance when the model is already well-trained. Instead, we consider an approach inspired by structured prediction cascades, which use max-marginal pruning to generate lattices. We obtain a high-accuracy phonetic recognition system with several expensive feature types: a segment neural network, a second-order language model, and second-order phone boundary features.

cs.CL↗