arXiv subjects

Hang Li

Publications and source records attributed to Hang Li.

At least 325 records · Page 18Linked to original sources

arXiv2016

Neural Machine Translation Advised by Statistical Machine Translation

Neural Machine Translation (NMT) is a new approach to machine translation that has made great progress in recent years. However, recent studies show that NMT generally produces fluent but inadequate translations (Tu et al. 2016b; Tu et al. 2016a; He et al. 2016; Tu et al. 2017). This is in contrast to conventional Statistical Machine Translation (SMT), which usually yields adequate but non-fluent translations. It is natural, therefore, to leverage the advantages of both models for better translations, and in this work we propose to incorporate SMT model into NMT framework. More specifically, at each decoding step, SMT offers additional recommendations of generated words based on the decoding information from NMT (e.g., the generated partial translation and attention history). Then we employ an auxiliary classifier to score the SMT recommendations and a gating function to combine the SMT recommendations with NMT generations, both of which are jointly trained within the NMT architecture in an end-to-end manner. Experimental results on Chinese-English translation show that the proposed approach achieves significant and consistent improvements over state-of-the-art NMT and SMT systems on multiple NIST test sets.

cs.CL↗

arXiv2016

Experimental Study of Forrelation in Nuclear Spins

Correlation functions are often employed to quantify the relationships among interdependent variables or sets of data. Recently, a new class of correlation functions, called Forrelation, has been introduced by Aaronson and Ambainis for studying the query complexity of quantum devices. It was found that there exists a quantum query algorithm solving 2-fold Forrelation problems with an exponential quantum speedup over all possible classical means, which represents essentially the largest possible separation between quantum and classical query complexities. Here we report an experimental study probing the 2-fold and 3-fold Forrelations encoded in nuclear spins. The major experimental challenge is to control the spin fluctuation to within a threshold value, which is achieved by developing a set of optimized GRAPE pulse sequences. Overall, our small-scale implementation indicates that the quantum query algorithm is capable of determine the values of Forrelations within an acceptable accuracy required for demonstrating quantum supremacy, given the current technology and in the presence of experimental noise.

quant-ph↗

arXiv2016

Neural Machine Translation with Reconstruction

Although end-to-end Neural Machine Translation (NMT) has achieved remarkable progress in the past two years, it suffers from a major drawback: translations generated by NMT systems often lack of adequacy. It has been widely observed that NMT tends to repeatedly translate some source words while mistakenly ignoring other words. To alleviate this problem, we propose a novel encoder-decoder-reconstructor framework for NMT. The reconstructor, incorporated into the NMT model, manages to reconstruct the input source sentence from the hidden layer of the output target sentence, to ensure that the information in the source side is transformed to the target side as much as possible. Experiments show that the proposed framework significantly improves the adequacy of NMT output and achieves superior translation result over state-of-the-art NMT and statistical MT systems.

cs.CL↗

arXiv2016

Correlation between non-centrosymmetry and superconductivity in quasi-one-dimensional compounds A2Cr3As3 (A=K, Rb)

Non-centrosymmetric superconductors, whose crystal structure is absent of inversion symmetry, have recently received special attentions due to the expectation of unconventional pairings and exotic physics associated with such pairings. The newly discovered superconductors A2Cr3As3 (A=K, Rb), featured by the quasi-one dimensional structure with conducting CrAs chains, belongs to such kind of superconductor. In this study, we are the first to report the finding that the superconductivity of A2Cr3As3 (A=K, Rb) has a positive correlation with the extent of non-centrosymmetry. Our in-situ high pressure ac susceptibility and synchrotron x-ray diffraction measurements reveal that the larger bond angle of As-Cr-As in the CrAs chains can be taken as a key factor controlling superconductivity. While the smaller bond angle and the distance between the CrAs chains also affect the superconductivity due to their structural connections with the angle. We find that the larger value of the difference between the larger and samller angles, which is associated with the extent of the non-centrosymmetry of the lattice structure, is in favor of superconductivity. These results are expected to shed a new light on the underlying mechanism of the superconductivity in these Q1D superconductors and also to provide new perspective in understanding other non-centrosymmetric superconductors.

cond-mat.supr-con↗

arXiv2016

Interactive Attention for Neural Machine Translation

Conventional attention-based Neural Machine Translation (NMT) conducts dynamic alignment in generating the target sentence. By repeatedly reading the representation of source sentence, which keeps fixed after generated by the encoder (Bahdanau et al., 2015), the attention mechanism has greatly enhanced state-of-the-art NMT. In this paper, we propose a new attention mechanism, called INTERACTIVE ATTENTION, which models the interaction between the decoder and the representation of source sentence during translation by both reading and writing operations. INTERACTIVE ATTENTION can keep track of the interaction history and therefore improve the translation performance. Experiments on NIST Chinese-English translation task show that INTERACTIVE ATTENTION can achieve significant improvements over both the previous attention-based NMT baseline and some state-of-the-art variants of attention-based NMT (i.e., coverage models (Tu et al., 2016)). And neural machine translator with our INTERACTIVE ATTENTION can outperform the open source attention-based NMT system Groundhog by 4.22 BLEU points and the open source phrase-based system Moses by 3.94 BLEU points averagely on multiple test sets.

cs.CL↗

arXiv2016

Cosmological constraints on coupled dark energy

The coupled dark energy model provides a possible approach to mitigate the coincidence problem of cosmological standard model. Here, the coupling term is assumed as $\bar{Q}=3Hξ_x\barρ_x$, which is related to the interaction rate and energy density of dark energy. We derive the background and perturbation evolution equations for several coupled models. Then, we test these models by currently available cosmic observations which include cosmic microwave background radiation from Planck 2015, baryon acoustic oscillation, type Ia supernovae, $fσ_8(z)$ data points from redshift-space distortions, and weak gravitational lensing. The constraint results tell us there is no evidence of interaction at 2$σ$ level, it is very hard to distinguish different coupled models from other ones.

astro-ph.CO↗

arXiv2016

Modeling Coverage for Neural Machine Translation

Attention mechanism has enhanced state-of-the-art Neural Machine Translation (NMT) by jointly learning to align and translate. It tends to ignore past alignment information, however, which often leads to over-translation and under-translation. To address this problem, we propose coverage-based NMT in this paper. We maintain a coverage vector to keep track of the attention history. The coverage vector is fed to the attention model to help adjust future attention, which lets NMT system to consider more about untranslated source words. Experiments show that the proposed approach significantly improves both translation quality and alignment quality over standard attention-based NMT.

cs.CL↗

arXiv2016

Multiuser Energy Diversity in Energy Harvesting Wireless Communications

Energy harvesting communication has raised great research interests due to its wide applications and feasibility of commercialization. In this paper, we investigate the multiuser energy diversity. Specifically, we reveal the throughput gain coming from the increase of total available energy harvested over time/space and from the combined dynamics of batteries. Considering both centralized and distributed access schemes, the scaling of the average throughput over the number of transmitters is studied, along with the scaling of corresponding available energy in the batteries.

cs.IT↗

arXiv2016

Incorporating Copying Mechanism in Sequence-to-Sequence Learning

We address an important problem in sequence-to-sequence (Seq2Seq) learning referred to as copying, in which certain segments in the input sequence are selectively replicated in the output sequence. A similar phenomenon is observable in human language communication. For example, humans tend to repeat entity names or even long phrases in conversation. The challenge with regard to copying in Seq2Seq is that new machinery is needed to decide when to perform the operation. In this paper, we incorporate copying into neural network-based Seq2Seq learning and propose a new model called CopyNet with encoder-decoder structure. CopyNet can nicely integrate the regular way of word generation in the decoder with the new copying mechanism which can choose sub-sequences in the input sequence and put them at proper places in the output sequence. Our empirical study on both synthetic data sets and real world data sets demonstrates the efficacy of CopyNet. For example, CopyNet can outperform regular RNN-based model with remarkable margins on text summarization tasks.

cs.CL↗

arXiv2016

Memory-enhanced Decoder for Neural Machine Translation

We propose to enhance the RNN decoder in a neural machine translator (NMT) with external memory, as a natural but powerful extension to the state in the decoding RNN. This memory-enhanced RNN decoder is called \textsc{MemDec}. At each time during decoding, \textsc{MemDec} will read from this memory and write to this memory once, both with content-based addressing. Unlike the unbounded memory in previous work\cite{RNNsearch} to store the representation of source sentence, the memory in \textsc{MemDec} is a matrix with pre-determined size designed to better capture the information important for the decoding process at each time step. Our empirical study on Chinese-English translation shows that it can improve by $4.8$ BLEU upon Groundhog and $5.3$ BLEU upon on Moses, yielding the best performance achieved with the same training set.

cs.CL↗

arXiv2016

Neural Machine Translation with External Phrase Memory

In this paper, we propose phraseNet, a neural machine translator with a phrase memory which stores phrase pairs in symbolic form, mined from corpus or specified by human experts. For any given source sentence, phraseNet scans the phrase memory to determine the candidate phrase pairs and integrates tagging information in the representation of source sentence accordingly. The decoder utilizes a mixture of word-generating component and phrase-generating component, with a specifically designed strategy to generate a sequence of multiple words all at once. The phraseNet not only approaches one step towards incorporating external knowledge into neural machine translation, but also makes an effort to extend the word-by-word generation mechanism of recurrent neural network. Our empirical study on Chinese-to-English translation shows that, with carefully-chosen phrase table in memory, phraseNet yields 3.45 BLEU improvement over the generic neural machine translator.

cs.CL↗

arXiv2016

A novel three-axis cylindrical hohlraum designed for inertial confinement fusion ignition

A novel ignition hohlraum for indirect-drive inertial confinement fusion is proposed, which is named as three-axis cylindrical hohlraum (TACH). TACH is a kind of 6 laser entrance holes (LEHs) hohlraum, which is made of three cylindrical hohlraums orthogonally jointed. Laser beams are injected through every entrance hole with the same incident angle of 55°. The view-factor simulation result shows that the time-varying drive asymmetry of TACH is no more than 1.0% in the whole drive pulse period without any supplementary technology such as beam phasing etc. Its coupling efficiency of TACH is close to that of 6 LEHs spherical hohlraum with corresponding size. Its plasma-filling time is close to typical cylindrical ignition hohlraum. Its laser plasma interaction has as low backscattering as the outer cone of the cylindrical ignition hohlraum. Therefore, the proposed hohlraum provides a competitive candidate for ignition hohlraum.

physics.plasm-ph↗

arXiv2016

Tunable spin-charge conversion through topological phase transitions in zigzag nanoribbons

We study spin-orbit torques and charge pumping in magnetic quasi-one dimensional zigzag nanoribbons with hexagonal lattice, in the presence of large intrinsic spin-orbit coupling. Such a system experiences topological phase transition from a trivial band insulator to a quantum spin Hall insulator either by tuning the magnetization direction or the intrinsic spin-orbit coupling. We find that spin-charge conversion efficiency (i.e. spin-orbit torque and charge pumping) is dramatically enhanced at the topological transition, displaying a substantial angular anisotropy.

cond-mat.mes-hall↗

arXiv2016

Neural Generative Question Answering

This paper presents an end-to-end neural network model, named Neural Generative Question Answering (GENQA), that can generate answers to simple factoid questions, based on the facts in a knowledge-base. More specifically, the model is built on the encoder-decoder framework for sequence-to-sequence learning, while equipped with the ability to enquire the knowledge-base, and is trained on a corpus of question-answer pairs, with their associated triples in the knowledge-base. Empirical study shows the proposed model can effectively deal with the variations of questions and answers, and generate right and natural answers by referring to the facts in the knowledge-base. The experiment on question answering demonstrates that the proposed model can outperform an embedding-based QA model as well as a neural dialogue model trained on the same data.

cs.CL↗

arXiv2016

Incorporating Semantic Knowledge into Latent Matching Model in Search

The relevance between a query and a document in search can be represented as matching degree between the two objects. Latent space models have been proven to be effective for the task, which are often trained with click-through data. One technical challenge with the approach is that it is hard to train a model for tail queries and tail documents for which there are not enough clicks. In this paper, we propose to address the challenge by learning a latent matching model, using not only click-through data but also semantic knowledge. The semantic knowledge can be categories of queries and documents as well as synonyms of words, manually or automatically created. Specifically, we incorporate semantic knowledge into the objective function by including regularization terms. We develop two methods to solve the learning task on the basis of coordinate descent and gradient descent respectively, which can be employed in different settings. Experimental results on two datasets from an app search engine demonstrate that our model can make effective use of semantic knowledge, and thus can significantly enhance the accuracies of latent matching models, particularly for tail queries.

cs.IR↗

arXiv2016

A Novel Approach to Dropped Pronoun Translation

Dropped Pronouns (DP) in which pronouns are frequently dropped in the source language but should be retained in the target language are challenge in machine translation. In response to this problem, we propose a semi-supervised approach to recall possibly missing pronouns in the translation. Firstly, we build training data for DP generation in which the DPs are automatically labelled according to the alignment information from a parallel corpus. Secondly, we build a deep learning-based DP generator for input sentences in decoding when no corresponding references exist. More specifically, the generation is two-phase: (1) DP position detection, which is modeled as a sequential labelling task with recurrent neural networks; and (2) DP prediction, which employs a multilayer perceptron with rich features. Finally, we integrate the above outputs into our translation system to recall missing pronouns by both extracting rules from the DP-labelled training data and translating the DP-generated input sentences. Experimental results show that our approach achieves a significant improvement of 1.58 BLEU points in translation performance with 66% F-score for DP generation accuracy.

cs.CL↗

arXiv2016

Chiral Quantum Walks

Given its importance to many other areas of physics, from condensed matter physics to thermodynamics, time-reversal symmetry has had relatively little influence on quantum information science. Here we develop a network-based picture of time-reversal theory, classifying Hamiltonians and quantum circuits as time-symmetric or not in terms of the elements and geometries of their underlying networks. Many of the typical circuits of quantum information science are found to exhibit time-asymmetry. Moreover, we show that time-asymmetry in circuits can be controlled using local gates only, and can simulate time-asymmetry in Hamiltonian evolution. We experimentally implement a fundamental example in which controlled time-reversal asymmetry in a palindromic quantum circuit leads to near-perfect transport. Our results pave the way for using time-symmetry breaking to control coherent transport, and imply that time-asymmetry represents an omnipresent yet poorly understood effect in quantum information science.

quant-ph↗

arXiv2016

Energy-Efficient Localization and Tracking of Mobile Devices in Wireless Sensor Networks

Wireless sensor networks (WSNs) are effective for locating and tracking people and objects in various industrial environments. Since energy consumption is critical to prolonging the lifespan of WSNs, we propose an energy-efficient LOcalization and Tracking} (eLOT) system, using low-cost and portable hardware to enable highly accurate tracking of targets. Various fingerprint-based approaches for localization and tracking are implemented in eLOT. In order to achieve high energy efficiency, a network-level scheme coordinating collision and interference is proposed. On the other hand, based on the location information, mobile devices in eLOT can quickly associate with the specific channel in a given area, while saving energy through avoiding unnecessary transmission. Finally, a platform based on TI CC2530 and the Linux operating system is built to demonstrate the effectiveness of our proposed scheme in terms of localization accuracy and energy efficiency.

cs.IT↗