arXiv subjects

Xiaodong He

Publications and source records attributed to Xiaodong He.

At least 91 records · Page 5Linked to original sources

arXiv2021

ViDA-MAN: Visual Dialog with Digital Humans

We demonstrate ViDA-MAN, a digital-human agent for multi-modal interaction, which offers realtime audio-visual responses to instant speech inquiries. Compared to traditional text or voice-based system, ViDA-MAN offers human-like interactions (e.g, vivid voice, natural facial expression and body gestures). Given a speech request, the demonstration is able to response with high quality videos in sub-second latency. To deliver immersive user experience, ViDA-MAN seamlessly integrates multi-modal techniques including Acoustic Speech Recognition (ASR), multi-turn dialog, Text To Speech (TTS), talking heads video generation. Backed with large knowledge base, ViDA-MAN is able to chat with users on a number of topics including chit-chat, weather, device control, News recommendations, booking hotels, as well as answering questions via structured knowledge.

cs.CV↗

arXiv2021

SGG: Learning to Select, Guide, and Generate for Keyphrase Generation

Keyphrases, that concisely summarize the high-level topics discussed in a document, can be categorized into present keyphrase which explicitly appears in the source text, and absent keyphrase which does not match any contiguous subsequence but is highly semantically related to the source. Most existing keyphrase generation approaches synchronously generate present and absent keyphrases without explicitly distinguishing these two categories. In this paper, a Select-Guide-Generate (SGG) approach is proposed to deal with present and absent keyphrase generation separately with different mechanisms. Specifically, SGG is a hierarchical neural network which consists of a pointing-based selector at low layer concentrated on present keyphrase generation, a selection-guided generator at high layer dedicated to absent keyphrase generation, and a guider in the middle to transfer information from selector to generator. Experimental results on four keyphrase generation benchmarks demonstrate the effectiveness of our model, which significantly outperforms the strong baselines for both present and absent keyphrases generation. Furthermore, we extend SGG to a title generation task which indicates its extensibility in natural language generation tasks.

cs.CL↗

arXiv2021

K-PLUG: Knowledge-injected Pre-trained Language Model for Natural Language Understanding and Generation in E-Commerce

Existing pre-trained language models (PLMs) have demonstrated the effectiveness of self-supervised learning for a broad range of natural language processing (NLP) tasks. However, most of them are not explicitly aware of domain-specific knowledge, which is essential for downstream tasks in many domains, such as tasks in e-commerce scenarios. In this paper, we propose K-PLUG, a knowledge-injected pre-trained language model based on the encoder-decoder transformer that can be transferred to both natural language understanding and generation tasks. We verify our method in a diverse range of e-commerce scenarios that require domain-specific knowledge. Specifically, we propose five knowledge-aware self-supervised pre-training objectives to formulate the learning of domain-specific knowledge, including e-commerce domain-specific knowledge-bases, aspects of product entities, categories of product entities, and unique selling propositions of product entities. K-PLUG achieves new state-of-the-art results on a suite of domain-specific NLP tasks, including product knowledge base completion, abstractive product summarization, and multi-turn dialogue, significantly outperforms baselines across the board, which demonstrates that the proposed method effectively learns a diverse set of domain-specific knowledge for both language understanding and generation tasks.

cs.CL↗

arXiv2021

The JDDC 2.0 Corpus: A Large-Scale Multimodal Multi-Turn Chinese Dialogue Dataset for E-commerce Customer Service

With the development of the Internet, more and more people get accustomed to online shopping. When communicating with customer service, users may express their requirements by means of text, images, and videos, which precipitates the need for understanding these multimodal information for automatic customer service systems. Images usually act as discriminators for product models, or indicators of product failures, which play important roles in the E-commerce scenario. On the other hand, detailed information provided by the images is limited, and typically, customer service systems cannot understand the intents of users without the input text. Thus, bridging the gap of the image and text is crucial for the multimodal dialogue task. To handle this problem, we construct JDDC 2.0, a large-scale multimodal multi-turn dialogue dataset collected from a mainstream Chinese E-commerce platform (JD.com), containing about 246 thousand dialogue sessions, 3 million utterances, and 507 thousand images, along with product knowledge bases and image category annotations. We present the solutions of top-5 teams participating in the JDDC multimodal dialogue challenge based on this dataset, which provides valuable insights for further researches on the multimodal dialogue task.

cs.CL↗

arXiv2021

Incremental Learning for End-to-End Automatic Speech Recognition

In this paper, we propose an incremental learning method for end-to-end Automatic Speech Recognition (ASR) which enables an ASR system to perform well on new tasks while maintaining the performance on its originally learned ones. To mitigate catastrophic forgetting during incremental learning, we design a novel explainability-based knowledge distillation for ASR models, which is combined with a response-based knowledge distillation to maintain the original model's predictions and the "reason" for the predictions. Our method works without access to the training data of original tasks, which addresses the cases where the previous data is no longer available or joint training is costly. Results on a multi-stage sequential training task show that our method outperforms existing ones in mitigating forgetting. Furthermore, in two practical scenarios, compared to the target-reference joint training method, the performance drop of our method is 0.02% Character Error Rate (CER), which is 97% smaller than the drops of the baseline methods.

eess.AS↗

arXiv2021

RoR: Read-over-Read for Long Document Machine Reading Comprehension

Transformer-based pre-trained models, such as BERT, have achieved remarkable results on machine reading comprehension. However, due to the constraint of encoding length (e.g., 512 WordPiece tokens), a long document is usually split into multiple chunks that are independently read. It results in the reading field being limited to individual chunks without information collaboration for long document machine reading comprehension. To address this problem, we propose RoR, a read-over-read method, which expands the reading field from chunk to document. Specifically, RoR includes a chunk reader and a document reader. The former first predicts a set of regional answers for each chunk, which are then compacted into a highly-condensed version of the original document, guaranteeing to be encoded once. The latter further predicts the global answers from this condensed document. Eventually, a voting strategy is utilized to aggregate and rerank the regional and global answers for final prediction. Extensive experiments on two benchmarks QuAC and TriviaQA demonstrate the effectiveness of RoR for long document reading. Notably, RoR ranks 1st place on the QuAC leaderboard (https://quac.ai/) at the time of submission (May 17th, 2021).

cs.CL↗

arXiv2021

High fidelity entanglement of neutral atoms via a Rydberg-mediated single-modulated-pulse controlled-PHASE gate

Neutral atom platform has become an attractive choice to study the science of quantum information and quantum simulation, where intense efforts have been devoted to the entangling processes between individual atoms. For the development of this area, two-qubit controlled-PHASE gate via Rydberg blockade is one of the most essential elements. Recent theoretical studies have suggested the advantages of introducing non-trivial waveform modulation into the gate protocol, which is anticipated to improve its performance towards the next stage. We report our recent experimental results in realizing a two-qubit controlled-PHASE($C_Z$) gate via off-resonant modulated driving(ORMD) embedded in two-photon transition for Rb atoms. It relies upon a single modulated driving pulse with a carefully calculated smooth waveform to gain the appropriate phase accumulations required by the two-qubit gate. Combining this $C_Z$ gate with global microwave pulses, two-atom entanglement is generated with the raw fidelity of 0.945(6). Accounting for state preparation and measurement (SPAM) errors, we extract the entanglement operation fidelity to be 0.980(7). Our work features completing the $C_Z$ gate operation within a single pulse to avoid shelved Rydberg population, thus demonstrate another promising route for realizing high-fidelity two-qubit gate for neutral atom platform.

quant-ph↗

arXiv2021

CUSTOM: Aspect-Oriented Product Summarization for E-Commerce

Product summarization aims to automatically generate product descriptions, which is of great commercial potential. Considering the customer preferences on different product aspects, it would benefit from generating aspect-oriented customized summaries. However, conventional systems typically focus on providing general product summaries, which may miss the opportunity to match products with customer interests. To address the problem, we propose CUSTOM, aspect-oriented product summarization for e-commerce, which generates diverse and controllable summaries towards different product aspects. To support the study of CUSTOM and further this line of research, we construct two Chinese datasets, i.e., SMARTPHONE and COMPUTER, including 76,279 / 49,280 short summaries for 12,118 / 11,497 real-world commercial products, respectively. Furthermore, we introduce EXT, an extraction-enhanced generation framework for CUSTOM, where two famous sequence-to-sequence models are implemented in this paper. We conduct extensive experiments on the two proposed datasets for CUSTOM and show results of two famous baseline models and EXT, which indicates that EXT can generate diverse, high-quality, and consistent summaries.

cs.CL↗

arXiv2021

Defect-free arbitrary-geometry assembly of mixed-species atom arrays

Optically trapped mixed-species single atom arrays with arbitrary geometries are an attractive and promising platform for various applications, because tunable quantum systems with multiple components provide extra degrees of freedom for experimental control. Here, we report the first demonstration of two-dimensional $6\times4$ dual-species atom assembly with a filling fraction of 0.88 (0.89) for $^{85}$Rb ($^{87}$Rb) atoms. This mixed-species atomic synthetic is achieved via rearranging initially randomly distributed atoms using a sorting algorithm (heuristic heteronuclear algorithm) which is proposed for bottom-up atom assembly with both user-defined geometries and two-species atom number ratios. Our fully tunable hybrid-atom system of scalable advantages is a good starting point for high-fidelity quantum logic, many-body quantum simulation and forming defect-free single molecule arrays.

physics.atom-ph↗

arXiv2021

Joint System-Wise Optimization for Pipeline Goal-Oriented Dialog System

Recent work (Takanobu et al., 2020) proposed the system-wise evaluation on dialog systems and found that improvement on individual components (e.g., NLU, policy) in prior work may not necessarily bring benefit to pipeline systems in system-wise evaluation. To improve the system-wise performance, in this paper, we propose new joint system-wise optimization techniques for the pipeline dialog system. First, we propose a new data augmentation approach which automates the labeling process for NLU training. Second, we propose a novel stochastic policy parameterization with Poisson distribution that enables better exploration and offers a principled way to compute policy gradient. Third, we propose a reward bonus to help policy explore successful dialogs. Our approaches outperform the competitive pipeline systems from Takanobu et al. (2020) by big margins of 12% success rate in automatic system-wise evaluation and of 16% success rate in human evaluation on the standard multi-domain benchmark dataset MultiWOZ 2.1, and also outperform the recent state-of-the-art end-to-end trained model from DSTC9.

cs.CL↗

arXiv2021

RevCore: Review-augmented Conversational Recommendation

Existing conversational recommendation (CR) systems usually suffer from insufficient item information when conducted on short dialogue history and unfamiliar items. Incorporating external information (e.g., reviews) is a potential solution to alleviate this problem. Given that reviews often provide a rich and detailed user experience on different interests, they are potential ideal resources for providing high-quality recommendations within an informative conversation. In this paper, we design a novel end-to-end framework, namely, Review-augmented Conversational Recommender (RevCore), where reviews are seamlessly incorporated to enrich item information and assist in generating both coherent and informative responses. In detail, we extract sentiment-consistent reviews, perform review-enriched and entity-based recommendations for item suggestions, as well as use a review-attentive encoder-decoder for response generation. Experimental results demonstrate the superiority of our approach in yielding better performance on both recommendation and conversation responding.

cs.CL↗

arXiv2021

Neural Kalman Filtering for Speech Enhancement

Statistical signal processing based speech enhancement methods adopt expert knowledge to design the statistical models and linear filters, which is complementary to the deep neural network (DNN) based methods which are data-driven. In this paper, by using expert knowledge from statistical signal processing for network design and optimization, we extend the conventional Kalman filtering (KF) to the supervised learning scheme, and propose the neural Kalman filtering (NKF) for speech enhancement. Two intermediate clean speech estimates are first produced from recurrent neural networks (RNN) and linear Wiener filtering (WF) separately and are then linearly combined by a learned NKF gain to yield the NKF output. Supervised joint training is applied to NKF to learn to automatically trade-off between the instantaneous linear estimation made by the WF and the long-term non-linear estimation made by the RNN. The NKF method can be seen as using expert knowledge from WF to regularize the RNN estimations to improve its generalization ability to the noise conditions unseen in the training. Experiments in different noisy conditions show that the proposed method outperforms the baseline methods both in terms of objective evaluation metrics and automatic speech recognition (ASR) word error rates (WERs).

eess.AS↗

arXiv2021

Graph Ensemble Learning over Multiple Dependency Trees for Aspect-level Sentiment Classification

Recent work on aspect-level sentiment classification has demonstrated the efficacy of incorporating syntactic structures such as dependency trees with graph neural networks(GNN), but these approaches are usually vulnerable to parsing errors. To better leverage syntactic information in the face of unavoidable errors, we propose a simple yet effective graph ensemble technique, GraphMerge, to make use of the predictions from differ-ent parsers. Instead of assigning one set of model parameters to each dependency tree, we first combine the dependency relations from different parses before applying GNNs over the resulting graph. This allows GNN mod-els to be robust to parse errors at no additional computational cost, and helps avoid overparameterization and overfitting from GNN layer stacking by introducing more connectivity into the ensemble graph. Our experiments on the SemEval 2014 Task 4 and ACL 14 Twitter datasets show that our GraphMerge model not only outperforms models with single dependency tree, but also beats other ensemble mod-els without adding model parameters.

cs.CL↗

arXiv2021

Hydrophobic interaction determines docking affinity of SARS CoV 2 variants with antibodies

Preliminary epidemiologic, phylogenetic and clinical findings suggest that several novel severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) variants have increased transmissibility and decreased efficacy of several existing vaccines. Four mutations in the receptor-binding domain (RBD) of the spike protein that are reported to contribute to increased transmission. Understanding physical mechanism responsible for the affinity enhancement between the SARS-CoV-2 variants and ACE2 is the "urgent challenge" for developing blockers, vaccines and therapeutic antibodies against the coronavirus disease 2019 (COVID-19) pandemic. Based on a hydrophobic-interaction-based protein docking mechanism, this study reveals that the mutation N501Y obviously increased the hydrophobic attraction and decrease hydrophilic repulsion between the RBD and ACE2 that most likely caused the transmissibility increment of the variants. By analyzing the mutation-induced hydrophobic surface changes in the attraction and repulsion at the binding site of the complexes of the SARS-CoV-2 variants and antibodies, we found out that all the mutations of N501Y, E484K, K417N and L452R can selectively decrease or increase their binding affinity with some antibodies.

q-bio.BM↗

arXiv2021

Conversational Query Rewriting with Self-supervised Learning

Context modeling plays a critical role in building multi-turn dialogue systems. Conversational Query Rewriting (CQR) aims to simplify the multi-turn dialogue modeling into a single-turn problem by explicitly rewriting the conversational query into a self-contained utterance. However, existing approaches rely on massive supervised training data, which is labor-intensive to annotate. And the detection of the omitted important information from context can be further improved. Besides, intent consistency constraint between contextual query and rewritten query is also ignored. To tackle these issues, we first propose to construct a large-scale CQR dataset automatically via self-supervised learning, which does not need human annotation. Then we introduce a novel CQR model Teresa based on Transformer, which is enhanced by self-attentive keywords detection and intent consistency constraint. Finally, we conduct extensive experiments on two public datasets. Experimental results demonstrate that our proposed model outperforms existing CQR baselines significantly, and also prove the effectiveness of self-supervised learning on improving the CQR performance.

cs.CL↗

arXiv2021

Selective Attention Based Graph Convolutional Networks for Aspect-Level Sentiment Classification

Aspect-level sentiment classification aims to identify the sentiment polarity towards a specific aspect term in a sentence. Most current approaches mainly consider the semantic information by utilizing attention mechanisms to capture the interactions between the context and the aspect term. In this paper, we propose to employ graph convolutional networks (GCNs) on the dependency tree to learn syntax-aware representations of aspect terms. GCNs often show the best performance with two layers, and deeper GCNs do not bring additional gain due to over-smoothing problem. However, in some cases, important context words cannot be reached within two hops on the dependency tree. Therefore we design a selective attention based GCN block (SA-GCN) to find the most important context words, and directly aggregate these information into the aspect-term representation. We conduct experiments on the SemEval 2014 Task 4 datasets. Our experimental results show that our model outperforms the current state-of-the-art.

cs.CL↗

arXiv2020

Influence of ground-Rydberg coherence in two-qubit gate based on Rydberg blockade

For neutral atom qubits, the two-qubit gate is typically realized via the Rydberg blockade effect, which hints about the special status of the Rydberg level besides the regular qubit register states. Here, we carry out experimental and theoretical studies to reveal how the ground-Rydberg coherence of the control qubit atom affects the process of two-qubit Controlled-Z ($C_Z$) gate, such as the commonly used ground-Rydberg $π$-gap-$π$ pulse sequence originally proposed in Phys. Rev. Lett. \textbf{85}, 2208 (2000). We measure the decoherence of the control qubit atom after the $π$-gap-$π$ pulses and make a direct comparison with the typical decoherence time $τ_{gr}$ extracted from Ramsey fringes of the ground-Rydberg transition. In particular, we observe that the control qubit atom subject to such pulse sequences experiences a process which is essentially similar to the ground-Rydberg Ramsey interference. Furthermore, we build a straightforward theoretical model to link the decoherence process of control qubit subject to $C_Z$ gate $π$-gap-$π$ pulse sequence and the $τ_{gr}$, and also analyze the typical origins of decoherence effects. Finally, we discuss the $C_Z$ gate fidelity loss due to the limits imposed by the ground-Rydberg coherence properties and prospective for improving fidelity with new gate protocols.

quant-ph↗

arXiv2020

Efficient preparation of 2D defect-free atom arrays with near-fewest sorting-atom moves

Sorting atoms stochastically loaded in optical tweezer arrays via an auxiliary mobile tweezer is an efficient approach to preparing intermediate-scale defect-free atom arrays in arbitrary geometries. However, high filling fraction of atom-by-atom assemblers is impeded by redundant sorting moves with imperfect atom transport, especially for scaling the system size to larger atom numbers. Here, we propose a new sorting algorithm (heuristic cluster algorithm, HCA) which provides near-fewest moves in our tailored atom assembler scheme and experimentally demonstrate a $5\times6$ defect-free atom array with 98.4(7)$\%$ filling fraction for one rearrangement cycle. The feature of HCA that the number of moves $N_{m}\approx N$ ($N$ is the number of defect sites to be filled) makes the filling fraction uniform as the size of atom assembler enlarged. Our method is essential to scale hundreds of assembled atoms for bottom-up quantum computation, quantum simulation and precision measurement.

quant-ph↗