arXiv subjects

Qian Chen

Publications and source records attributed to Qian Chen.

At least 325 records · Page 18Linked to original sources

arXiv2021

Discriminative Self-training for Punctuation Prediction

Punctuation prediction for automatic speech recognition (ASR) output transcripts plays a crucial role for improving the readability of the ASR transcripts and for improving the performance of downstream natural language processing applications. However, achieving good performance on punctuation prediction often requires large amounts of labeled speech transcripts, which is expensive and laborious. In this paper, we propose a Discriminative Self-Training approach with weighted loss and discriminative label smoothing to exploit unlabeled speech transcripts. Experimental results on the English IWSLT2011 benchmark test set and an internal Chinese spoken language dataset demonstrate that the proposed approach achieves significant improvement on punctuation prediction accuracy over strong baselines including BERT, RoBERTa, and ELECTRA models. The proposed Discriminative Self-Training approach outperforms the vanilla self-training approach. We establish a new state-of-the-art (SOTA) on the IWSLT2011 test set, outperforming the current SOTA model by 1.3% absolute gain on F$_1$.

cs.CL↗

arXiv2021

High speed and reconfigurable optronic neural network with digital nonlinear activation

With its unique parallel processing capability, optical neural network has shown low-power consumption in image recognition and speech processing. At present, the manufacturing technology of programmable photonic chip is not mature, and the realization of optical neural network in free-space is still a hot spot of intelligent optical computing. In this article, based on MNIST datasets and 4f system, three-layer optical neural networks are constructed, whose recognition accuracy can reach 93.66%. Our network is programmable, high speed, reconfigurable and is better than the existing free-space optical neural network in terms of spatial complexity.

cs.ET↗

arXiv2021

Loop Quantum Gravity's Boundary Maps

In canonical quantum gravity, the presence of spatial boundaries naturally leads to a boundary quantum states, representing quantum boundary conditions for the bulk fields. As a consequence, quantum states of the bulk geometry needs to be upgraded to wave-functions valued in the boundary Hilbert space: the bulk become quantum operator acting on boundary states. We apply this to loop quantum gravity and describe spin networks with 2d boundary as wave-functions mapping bulk holonomies to spin states on the boundary. This sets the bulk-boundary relation in a clear mathematical framework, which allows to define the boundary density matrix induced by a bulk spin network states after tracing out the bulk degrees of freedom. We ask the question of the bulk reconstruction and prove a boundary-to-bulk universal reconstruction procedure, to be understood as a purification of the mixed boundary state into a pure bulk state. We further perform a first investigation in the algebraic structure of induced boundary density matrices and show how correlations between bulk excitations, i.e. quanta of 3d geometry, get reflected into the boundary density matrix.

gr-qc↗

arXiv2021

RGB-D Salient Object Detection via 3D Convolutional Neural Networks

RGB-D salient object detection (SOD) recently has attracted increasing research interest and many deep learning methods based on encoder-decoder architectures have emerged. However, most existing RGB-D SOD models conduct feature fusion either in the single encoder or the decoder stage, which hardly guarantees sufficient cross-modal fusion ability. In this paper, we make the first attempt in addressing RGB-D SOD through 3D convolutional neural networks. The proposed model, named RD3D, aims at pre-fusion in the encoder stage and in-depth fusion in the decoder stage to effectively promote the full integration of RGB and depth streams. Specifically, RD3D first conducts pre-fusion across RGB and depth modalities through an inflated 3D encoder, and later provides in-depth feature fusion by designing a 3D decoder equipped with rich back-projection paths (RBPP) for leveraging the extensive aggregation ability of 3D convolutions. With such a progressive fusion strategy involving both the encoder and decoder, effective and thorough interaction between the two modalities can be exploited and boost the detection accuracy. Extensive experiments on six widely used benchmark datasets demonstrate that RD3D performs favorably against 14 state-of-the-art RGB-D SOD approaches in terms of four key evaluation metrics. Our code will be made publicly available: https://github.com/PPOLYpubki/RD3D.

cs.CV↗

arXiv2021

A Semi-parametric Realized Joint Value-at-Risk and Expected Shortfall Regression Framework

A new realized conditional autoregressive Value-at-Risk (VaR) framework is proposed, through incorporating a measurement equation into the original quantile regression model. The framework is further extended by employing various Expected Shortfall (ES) components, to jointly estimate and forecast VaR and ES. The measurement equation models the contemporaneous dependence between the realized measure (i.e., Realized Variance and Realized Range) and the latent conditional ES. An adaptive Bayesian Markov Chain Monte Carlo method is employed for estimation and forecasting, the properties of which are assessed and compared with maximum likelihood through a simulation study. In a comprehensive forecasting study on 1% and 2.5 % quantile levels, the proposed models are compared to a range of parametric, non-parametric and semi-parametric models, based on 7 market indices and 7 individual assets. One-day-ahead VaR and ES forecasting results favor the proposed models, especially when incorporating the sub-sampled Realized Variance and the sub-sampled Realized Range in the model.

q-fin.RM↗

arXiv2020

Multi-feature driven active contour segmentation model for infrared image with intensity inhomogeneity

Infrared (IR) image segmentation is essential in many urban defence applications, such as pedestrian surveillance, vehicle counting, security monitoring, etc. Active contour model (ACM) is one of the most widely used image segmentation tools at present, but the existing methods only utilize the local or global single feature information of image to minimize the energy function, which is easy to cause false segmentations in IR images. In this paper, we propose a multi-feature driven active contour segmentation model to handle IR images with intensity inhomogeneity. Firstly, an especially-designed signed pressure force (SPF) function is constructed by combining the global information calculated by global average gray information and the local multi-feature information calculated by local entropy, local standard deviation and gradient information. Then, we draw upon adaptive weight coefficient calculated by local range to adjust the afore-mentioned global term and local term. Next, the SPF function is substituted into the level set formulation (LSF) for further evolution. Finally, the LSF converges after a finite number of iterations, and the IR image segmentation result is obtained from the corresponding convergence result. Experimental results demonstrate that the presented method outperforms the state-of-the-art models in terms of precision rate and overlapping rate in IR test images.

cs.CV↗

arXiv2020

Infrared target tracking based on proximal robust principal component analysis method

Infrared target tracking plays an important role in both civil and military fields. The main challenges in designing a robust and high-precision tracker for infrared sequences include overlap, occlusion and appearance change. To this end, this paper proposes an infrared target tracker based on proximal robust principal component analysis method. Firstly, the observation matrix is decomposed into a sparse occlusion matrix and a low-rank target matrix, and the constraint optimization is carried out with an approaching proximal norm which is better than L1-norm. To solve this convex optimization problem, Alternating Direction Method of Multipliers (ADMM) is employed to estimate the variables alternately. Finally, the framework of particle filter with model update strategy is exploited to locate the target. Through a series of experiments on real infrared target sequences, the effectiveness and robustness of our algorithm are proved.

cs.CV↗

arXiv2020

T3: Tree-Autoencoder Constrained Adversarial Text Generation for Targeted Attack

Adversarial attacks against natural language processing systems, which perform seemingly innocuous modifications to inputs, can induce arbitrary mistakes to the target models. Though raised great concerns, such adversarial attacks can be leveraged to estimate the robustness of NLP models. Compared with the adversarial example generation in continuous data domain (e.g., image), generating adversarial text that preserves the original meaning is challenging since the text space is discrete and non-differentiable. To handle these challenges, we propose a target-controllable adversarial attack framework T3, which is applicable to a range of NLP tasks. In particular, we propose a tree-based autoencoder to embed the discrete text data into a continuous representation space, upon which we optimize the adversarial perturbation. A novel tree-based decoder is then applied to regularize the syntactic correctness of the generated text and manipulate it on either sentence (T3(Sent)) or word (T3(Word)) level. We consider two most representative NLP tasks: sentiment analysis and question answering (QA). Extensive experimental results and human studies show that T3 generated adversarial texts can successfully manipulate the NLP models to output the targeted incorrect answer without misleading the human. Moreover, we show that the generated adversarial texts have high transferability which enables the black-box attacks in practice. Our work sheds light on an effective and general way to examine the robustness of NLP models. Our code is publicly available at https://github.com/AI-secure/T3/.

cs.CL↗

arXiv2020

Align, Mask and Select: A Simple Method for Incorporating Commonsense Knowledge into Language Representation Models

The state-of-the-art pre-trained language representation models, such as Bidirectional Encoder Representations from Transformers (BERT), rarely incorporate commonsense knowledge or other knowledge explicitly. We propose a pre-training approach for incorporating commonsense knowledge into language representation models. We construct a commonsense-related multi-choice question answering dataset for pre-training a neural language representation model. The dataset is created automatically by our proposed "align, mask, and select" (AMS) method. We also investigate different pre-training tasks. Experimental results demonstrate that pre-training models using the proposed approach followed by fine-tuning achieve significant improvements over previous state-of-the-art models on two commonsense-related benchmarks, including CommonsenseQA and Winograd Schema Challenge. We also observe that fine-tuned models after the proposed pre-training approach maintain comparable performance on other NLP tasks, such as sentence classification and natural language inference tasks, compared to the original BERT models. These results verify that the proposed approach, while significantly improving commonsense-related NLP tasks, does not degrade the general language representation capabilities.

cs.CL↗

arXiv2020

Deep-learning-enabled geometric constraints and phase unwrapping for single-shot absolute 3D shape measurement

Fringe projection profilometry (FPP) is one of the most popular three-dimensional (3D) shape measurement techniques, and has becoming more prevalently adopted in intelligent manufacturing, defect detection and some other important applications. In FPP, how to efficiently recover the absolute phase has always been a great challenge. The stereo phase unwrapping (SPU) technologies based on geometric constraints can eliminate phase ambiguity without projecting any additional fringe patterns, which maximizes the efficiency of the retrieval of absolute phase. Inspired by the recent success of deep learning technologies for phase analysis, we demonstrate that deep learning can be an effective tool that organically unifies the phase retrieval, geometric constraints, and phase unwrapping steps into a comprehensive framework. Driven by extensive training dataset, the neutral network can gradually "learn" how to transfer one high-frequency fringe pattern into the "physically meaningful", and "most likely" absolute phase, instead of "step by step" as in convention approaches. Based on the properly trained framework, high-quality phase retrieval and robust phase ambiguity removal can be achieved based on only single-frame projection. Experimental results demonstrate that compared with traditional SPU, our method can more efficiently and stably unwrap the phase of dense fringe images in a larger measurement volume with fewer camera views. Limitations about the proposed approach are also discussed. We believe the proposed approach represents an important step forward in high-speed, high-accuracy, motion-artifacts-free absolute 3D shape measurement for complicated object from a single fringe pattern.

eess.IV↗

arXiv2020

Controllable Time-Delay Transformer for Real-Time Punctuation Prediction and Disfluency Detection

With the increased applications of automatic speech recognition (ASR) in recent years, it is essential to automatically insert punctuation marks and remove disfluencies in transcripts, to improve the readability of the transcripts as well as the performance of subsequent applications, such as machine translation, dialogue systems, and so forth. In this paper, we propose a Controllable Time-delay Transformer (CT-Transformer) model that jointly completes the punctuation prediction and disfluency detection tasks in real time. The CT-Transformer model facilitates freezing partial outputs with controllable time delay to fulfill the real-time constraints in partial decoding required by subsequent applications. We further propose a fast decoding strategy to minimize latency while maintaining competitive performance. Experimental results on the IWSLT2011 benchmark dataset and an in-house Chinese annotated dataset demonstrate that the proposed approach outperforms the previous state-of-the-art models on F-scores and achieves a competitive inference speed.

cs.CL↗

arXiv2020

Sequential Neural Networks for Noetic End-to-End Response Selection

The noetic end-to-end response selection challenge as one track in the 7th Dialog System Technology Challenges (DSTC7) aims to push the state of the art of utterance classification for real world goal-oriented dialog systems, for which participants need to select the correct next utterances from a set of candidates for the multi-turn context. This paper presents our systems that are ranked top 1 on both datasets under this challenge, one focused and small (Advising) and the other more diverse and large (Ubuntu). Previous state-of-the-art models use hierarchy-based (utterance-level and token-level) neural networks to explicitly model the interactions among different turns' utterances for context modeling. In this paper, we investigate a sequential matching model based only on chain sequence for multi-turn response selection. Our results demonstrate that the potentials of sequential matching approaches have not yet been fully exploited in the past for multi-turn response selection. In addition to ranking top 1 in the challenge, the proposed model outperforms all previous models, including state-of-the-art hierarchy-based models, on two large-scale public multi-turn response selection benchmark datasets.

cs.CL↗

arXiv2020

Transfer Learning for Context-Aware Spoken Language Understanding

Spoken language understanding (SLU) is a key component of task-oriented dialogue systems. SLU parses natural language user utterances into semantic frames. Previous work has shown that incorporating context information significantly improves SLU performance for multi-turn dialogues. However, collecting a large-scale human-labeled multi-turn dialogue corpus for the target domains is complex and costly. To reduce dependency on the collection and annotation effort, we propose a Context Encoding Language Transformer (CELT) model facilitating exploiting various context information for SLU. We explore different transfer learning approaches to reduce dependency on data collection and annotation. In addition to unsupervised pre-training using large-scale general purpose unlabeled corpora, such as Wikipedia, we explore unsupervised and supervised adaptive training approaches for transfer learning to benefit from other in-domain and out-of-domain dialogue corpora. Experimental results demonstrate that the proposed model with the proposed transfer learning approaches achieves significant improvement on the SLU performance over state-of-the-art models on two large-scale single-turn dialogue benchmarks and one large-scale multi-turn dialogue benchmark.

cs.CL↗

arXiv2020

On a universal solution to the transport-of-intensity equation

Transport-of-intensity equation (TIE) is one of the most well-known approaches for phase retrieval and quantitative phase imaging. It directly recovers the quantitative phase distribution of an optical field by through-focus intensity measurements in a noninterferometic, deterministic manner. Nevertheless, the accuracy and validity of state-of-the-art TIE solvers depend on restrictive preknowledge or assumptions, including appropriate boundary conditions, a well-defined closed region, and quasi-uniform in-focus intensity distribution, which, however, cannot be strictly satisfied simultaneously under practical experimental conditions. In this Letter, we propose a universal solution to TIE with the advantages of high accuracy, convergence guarantee, applicability to arbitrarily-shaped regions, and simplified implementation and computation. With the "maximum intensity assumption", we firstly simplified TIE as a standard Possion equation to get an initial guess of the solution. Then the initial solution is further refined iteratively by solving the same Possion equation, and thus, the instability associated with the division by zero/small intensity values and large intensity variations can be effectively bypassed. Simulations and experiments with arbitrary phase, arbitrary aperture shapes, and nonuniform intensity distributions verify the effectiveness and universality of the proposed method.

eess.IV↗

arXiv2020

Dual Adversarial Domain Adaptation

Unsupervised domain adaptation aims at transferring knowledge from the labeled source domain to the unlabeled target domain. Previous adversarial domain adaptation methods mostly adopt the discriminator with binary or $K$-dimensional output to perform marginal or conditional alignment independently. Recent experiments have shown that when the discriminator is provided with domain information in both domains and label information in the source domain, it is able to preserve the complex multimodal information and high semantic information in both domains. Following this idea, we adopt a discriminator with $2K$-dimensional output to perform both domain-level and class-level alignments simultaneously in a single discriminator. However, a single discriminator can not capture all the useful information across domains and the relationships between the examples and the decision boundary are rarely explored before. Inspired by multi-view learning and latest advances in domain adaptation, besides the adversarial process between the discriminator and the feature extractor, we also design a novel mechanism to make two discriminators pit against each other, so that they can provide diverse information for each other and avoid generating target features outside the support of the source domain. To the best of our knowledge, it is the first time to explore a dual adversarial strategy in domain adaptation. Moreover, we also use the semi-supervised learning regularization to make the representations more discriminative. Comprehensive experiments on two real-world datasets verify that our method outperforms several state-of-the-art domain adaptation methods.

cs.LG↗

arXiv2019

Homogeneous Online Transfer Learning with Online Distribution Discrepancy Minimization

Transfer learning has been demonstrated to be successful and essential in diverse applications, which transfers knowledge from related but different source domains to the target domain. Online transfer learning(OTL) is a more challenging problem where the target data arrive in an online manner. Most OTL methods combine source classifier and target classifier directly by assigning a weight to each classifier, and adjust the weights constantly. However, these methods pay little attention to reducing the distribution discrepancy between domains. In this paper, we propose a novel online transfer learning method which seeks to find a new feature representation, so that the marginal distribution and conditional distribution discrepancy can be online reduced simultaneously. We focus on online transfer learning with multiple source domains and use the Hedge strategy to leverage knowledge from source domains. We analyze the theoretical properties of the proposed algorithm and provide an upper mistake bound. Comprehensive experiments on two real-world datasets show that our method outperforms state-of-the-art methods by a large margin.

cs.LG↗

arXiv2019

High-speed in vitro intensity diffraction tomography

We demonstrate a label-free, scan-free {\it intensity} diffraction tomography technique utilizing annular illumination (aIDT) to rapidly characterize large-volume 3D refractive index distributions in vitro. By optimally matching the illumination geometry to the microscope pupil, our technique reduces the data requirement by 60$\times$ to achieve high-speed 10 Hz volume rates. Using 8 intensity images, we recover $\sim350\times100\times20μ$m$^3$ volumes with near diffraction-limited lateral resolution of 487 nm and axial resolution of 3.4 $μ$m. Our technique's large volume rate and high resolution enables 3D quantitative phase imaging of complex living biological samples across multiple length scales. We demonstrate aIDT's capabilities on unicellular diatom microalgae, epithelial buccal cell clusters with native bacteria, and live \emph{Caenorhabditis elegans} specimens. Within these samples, we recover macro-scale cellular structures, subcellular organelles, and dynamic micro-organism tissues with minimal motion artifacts. Quantifying such features has significant utility in oncology, immunology, and cellular pathophysiology, where these morphological features are evaluated for changes in the presence of disease, parasites, and new drug treatments. Finally, we simulate our aIDT system to highlight the accuracy and sensitivity of our technique. aIDT shows promise as a powerful high-speed, label-free computational microscopy technique applications where natural imaging is required to evaluate environmental effects on a sample in real-time. We provide example datasets and an open source implementation of aIDT at \href{https://github.com/bu-cisl/IDT-using-Annular-Illumination}{https://github.com/bu-cisl/IDT-using-Annular-Illumination}.

physics.optics↗

arXiv2019

Sequential Attention-based Network for Noetic End-to-End Response Selection

The noetic end-to-end response selection challenge as one track in Dialog System Technology Challenges 7 (DSTC7) aims to push the state of the art of utterance classification for real world goal-oriented dialog systems, for which participants need to select the correct next utterances from a set of candidates for the multi-turn context. This paper describes our systems that are ranked the top on both datasets under this challenge, one focused and small (Advising) and the other more diverse and large (Ubuntu). Previous state-of-the-art models use hierarchy-based (utterance-level and token-level) neural networks to explicitly model the interactions among different turns' utterances for context modeling. In this paper, we investigate a sequential matching model based only on chain sequence for multi-turn response selection. Our results demonstrate that the potentials of sequential matching approaches have not yet been fully exploited in the past for multi-turn response selection. In addition to ranking the top in the challenge, the proposed model outperforms all previous models, including state-of-the-art hierarchy-based models, and achieves new state-of-the-art performances on two large-scale public multi-turn response selection benchmark datasets.

cs.CL↗