arXiv subjects

Wei Wu

Publications and source records attributed to Wei Wu.

At least 433 records · Page 24Linked to original sources

arXiv2020

Learning a Simple and Effective Model for Multi-turn Response Generation with Auxiliary Tasks

We study multi-turn response generation for open-domain dialogues. The existing state-of-the-art addresses the problem with deep neural architectures. While these models improved response quality, their complexity also hinders the application of the models in real systems. In this work, we pursue a model that has a simple structure yet can effectively leverage conversation contexts for response generation. To this end, we propose four auxiliary tasks including word order recovery, utterance order recovery, masked word recovery, and masked utterance recovery, and optimize the objectives of these tasks together with maximizing the likelihood of generation. By this means, the auxiliary tasks that relate to context understanding can guide the learning of the generation model to achieve a better local optimum. Empirical studies with three benchmarks indicate that our model can significantly outperform state-of-the-art generation models in terms of response quality on both automatic evaluation and human judgment, and at the same time enjoys a much faster decoding process.

cs.CL↗

arXiv2020

Retrieve, Program, Repeat: Complex Knowledge Base Question Answering via Alternate Meta-learning

A compelling approach to complex question answering is to convert the question to a sequence of actions, which can then be executed on the knowledge base to yield the answer, aka the programmer-interpreter approach. Use similar training questions to the test question, meta-learning enables the programmer to adapt to unseen questions to tackle potential distributional biases quickly. However, this comes at the cost of manually labeling similar questions to learn a retrieval model, which is tedious and expensive. In this paper, we present a novel method that automatically learns a retrieval model alternately with the programmer from weak supervision, i.e., the system's performance with respect to the produced answers. To the best of our knowledge, this is the first attempt to train the retrieval model with the programmer jointly. Our system leads to state-of-the-art performance on a large-scale task for complex question answering over knowledge bases. We have released our code at https://github.com/DevinJake/MARL.

cs.AI↗

arXiv2020

Less is More: Data-Efficient Complex Question Answering over Knowledge Bases

Question answering is an effective method for obtaining information from knowledge bases (KB). In this paper, we propose the Neural-Symbolic Complex Question Answering (NS-CQA) model, a data-efficient reinforcement learning framework for complex question answering by using only a modest number of training samples. Our framework consists of a neural generator and a symbolic executor that, respectively, transforms a natural-language question into a sequence of primitive actions, and executes them over the knowledge base to compute the answer. We carefully formulate a set of primitive symbolic actions that allows us to not only simplify our neural network design but also accelerate model convergence. To reduce search space, we employ the copy and masking mechanisms in our encoder-decoder architecture to drastically reduce the decoder output vocabulary and improve model generalizability. We equip our model with a memory buffer that stores high-reward promising programs. Besides, we propose an adaptive reward function. By comparing the generated trial with the trials stored in the memory buffer, we derive the curriculum-guided reward bonus, i.e., the proximity and the novelty. To mitigate the sparse reward problem, we combine the adaptive reward and the reward bonus, reshaping the sparse reward into dense feedback. Also, we encourage the model to generate new trials to avoid imitating the spurious trials while making the model remember the past high-reward trials to improve data efficiency. Our NS-CQA model is evaluated on two datasets: CQA, a recent large-scale complex question answering dataset, and WebQuestionsSP, a multi-hop question answering dataset. On both datasets, our model outperforms the state-of-the-art models. Notably, on CQA, NS-CQA performs well on questions with higher complexity, while only using approximately 1% of the total training samples.

cs.CL↗

arXiv2020

Resonantly tunable second harmonic generation from lithium niobate metasurfaces

Second harmonic generation (SHG) is a coherent nonlinear phenomenon that plays an important role in laser color conversion. Lithium niobate (LN), which features both a large band gap and outstanding second-order nonlinearities, acts as an important optical material for nonlinear frequency conversion covering a wide spectral range from ultraviolet to mid-infrared. Here we experimentally demonstrate LN metasurfaces with controllable SHG properties. Distinct enhancements for the SHG efficiency are observed at Mie-resonances. And by changing the geometric parameters thus the resonances of the metasurfaces, we manage to selectively boost the SHG efficiency at different wavelengths. Our results would pave a way for developing with high flexibility the novel compact nonlinear light sources for applications, such as biosensing and optical communications.

physics.optics↗

arXiv2020

Knowledge-Grounded Dialogue Generation with Pre-trained Language Models

We study knowledge-grounded dialogue generation with pre-trained language models. To leverage the redundant external knowledge under capacity constraint, we propose equipping response generation defined by a pre-trained language model with a knowledge selection module, and an unsupervised approach to jointly optimizing knowledge selection and response generation with unlabeled dialogues. Empirical results on two benchmarks indicate that our model can significantly outperform state-of-the-art methods in both automatic evaluation and human judgment.

cs.CL↗

arXiv2020

StyleDGPT: Stylized Response Generation with Pre-trained Language Models

Generating responses following a desired style has great potentials to extend applications of open-domain dialogue systems, yet is refrained by lacking of parallel data for training. In this work, we explore the challenging task with pre-trained language models that have brought breakthrough to various natural language tasks. To this end, we introduce a KL loss and a style classifier to the fine-tuning step in order to steer response generation towards the target style in both a word-level and a sentence-level. Comprehensive empirical studies with two public datasets indicate that our model can significantly outperform state-of-the-art methods in terms of both style consistency and contextual coherence.

cs.CL↗

arXiv2020

Behaviour Prediction of Closed-loop HTS coils in Non-Uniform AC fields

Field decay rate is the key characteristic of the superconducting magnets based on closed-loop coils. However, in Maglev trains or rotating machines, closed-loop magnets work in external AC fields and will exhibit an evidently accelerated field decay resulting from dynamic resistances, which are usually much larger than joint resistance. Nevertheless, there has not been a numerical model capable of systematically studying this behaviour, which is the main topic of this work. The field decay curves of a closed-loop high-temperature-superconducting (HTS) coil in various AC fields are simulated based on H-formulation. A non-uniform external field generated by armature coils is considered. Reasonable consistence is found between experimental and simulation results. In our numerical model, the impact of current relaxation, which is a historical challenge, is analysed and subsequently eliminated with acceptable precision. Our simulation results suggest that most proportion of the field decay rate is from the innermost and outermost turns. Based on this observation, a magnetic shielding pattern is designed to reduce the field decay rate efficiently. This work has provided magnet designers an effective method to predict the field decay rate of closed-loop HTS coils in external AC fields, and explore various shielding designs.

physics.app-ph↗

arXiv2020

SAMOT: Switcher-Aware Multi-Object Tracking and Still Another MOT Measure

Multi-Object Tracking (MOT) is a popular topic in computer vision. However, identity issue, i.e., an object is wrongly associated with another object of a different identity, still remains to be a challenging problem. To address it, switchers, i.e., confusing targets thatmay cause identity issues, should be focused. Based on this motivation,this paper proposes a novel switcher-aware framework for multi-object tracking, which consists of Spatial Conflict Graph model (SCG) and Switcher-Aware Association (SAA). The SCG eliminates spatial switch-ers within one frame by building a conflict graph and working out the optimal subgraph. The SAA utilizes additional information from potential temporal switcher across frames, enabling more accurate data association. Besides, we propose a new MOT evaluation measure, Still Another IDF score (SAIDF), aiming to focus more on identity issues.This new measure may overcome some problems of the previous measures and provide a better insight for identity issues in MOT. Finally,the proposed framework is tested under both the traditional measures and the new measure we proposed. Extensive experiments show that ourmethod achieves competitive results on all measure.

cs.CV↗

arXiv2020

Collaborative Distillation in the Parameter and Spectrum Domains for Video Action Recognition

Recent years have witnessed the significant progress of action recognition task with deep networks. However, most of current video networks require large memory and computational resources, which hinders their applications in practice. Existing knowledge distillation methods are limited to the image-level spatial domain, ignoring the temporal and frequency information which provide structural knowledge and are important for video analysis. This paper explores how to train small and efficient networks for action recognition. Specifically, we propose two distillation strategies in the frequency domain, namely the feature spectrum and parameter distribution distillations respectively. Our insight is that appealing performance of action recognition requires \textit{explicitly} modeling the temporal frequency spectrum of video features. Therefore, we introduce a spectrum loss that enforces the student network to mimic the temporal frequency spectrum from the teacher network, instead of \textit{implicitly} distilling features as many previous works. Second, the parameter frequency distribution is further adopted to guide the student network to learn the appearance modeling process from the teacher. Besides, a collaborative learning strategy is presented to optimize the training process from a probabilistic view. Extensive experiments are conducted on several action recognition benchmarks, such as Kinetics, Something-Something, and Jester, which consistently verify effectiveness of our approach, and demonstrate that our method can achieve higher performance than state-of-the-art methods with the same backbone.

cs.CV↗

arXiv2020

Excited states of a phosphorus pair in silicon: Effects of valley-orbital interaction and electron-electron interactions

Excitations of impurity complexes in semiconductors can not only provide a route to fill the terahertz gap in optical technologies, but can also connect local quantum bits to scale up solid-state quantum-computing devices. However, taking into account both the interactions among electrons/holes, and the host band structures, is challenging. Here we combine first-principles band-structure calculations with quantum-chemistry methodology to evaluate the ground and excited states of a pair of phosphorous donors in silicon within s single framework. We use a broken-symmetry Hartree-Fock approach, followed by a time-dependent Hartree-Fock method to compute the excited states. Our Hamiltonian for each valley includes an anisotropic kinetic energy term, which splits the 2p_0 and 2p_+- transitions of isolated donors by ~4 meV, in good agreement with experiments. Our single-valley calculations show the optical response is a strong function of the optical polarisation, and suggest the use of valley polarisation to control optics and reduce oscillations in exchange interactions. When taking into account all valleys, including valley-orbital interactions, we find a gap opens between the 1s to 2p transition and low-energy charge-transfer states within 1s manifolds (which become optically allowed because of inter-donor interactions). In contrast to the single-valley case, we find charge-transfer excited states also in the triplet sector, thanks to the valley degrees of freedom. These states have a qualitatively correct energy as compared with the previous experiments; additionally, we predict new excitations below 20 meV that have not been analysed previously. A statistical average of nearest-neighbour pairs at different separations suggests that THz radiation could be used to excite pairs spin-selectively. Our approach can readily be extended to other donors and to other semiconducting hosts.

cond-mat.mes-hall↗

arXiv2020

Controllable dynamics of a dissipative two-level system

We propose a strategy to modulate the decoherence dynamics of a two-level system, which interacts with a dissipative bosonic environment, by introducing an assisted degree of freedom. It is revealed that the decay rate of the two-level system can be significantly suppressed under suitable steers of the assisted degree of freedom. Our result provides an alternative way to fight against decoherence and realize a controllable dissipative dynamics.

quant-ph↗

arXiv2020

Complementary Boundary Generator with Scale-Invariant Relation Modeling for Temporal Action Localization: Submission to ActivityNet Challenge 2020

This technical report presents an overview of our solution used in the submission to ActivityNet Challenge 2020 Task 1 (\textbf{temporal action localization/detection}). Temporal action localization requires to not only precisely locate the temporal boundaries of action instances, but also accurately classify the untrimmed videos into specific categories. In this paper, we decouple the temporal action localization task into two stages (i.e. proposal generation and classification) and enrich the proposal diversity through exhaustively exploring the influences of multiple components from different but complementary perspectives. Specifically, in order to generate high-quality proposals, we consider several factors including the video feature encoder, the proposal generator, the proposal-proposal relations, the scale imbalance, and ensemble strategy. Finally, in order to obtain accurate detections, we need to further train an optimal video classifier to recognize the generated proposals. Our proposed scheme achieves the state-of-the-art performance on the temporal action localization task with \textbf{42.26} average mAP on the challenge testing set.

cs.CV↗

arXiv2020

Class-wise Dynamic Graph Convolution for Semantic Segmentation

Recent works have made great progress in semantic segmentation by exploiting contextual information in a local or global manner with dilated convolutions, pyramid pooling or self-attention mechanism. In order to avoid potential misleading contextual information aggregation in previous works, we propose a class-wise dynamic graph convolution (CDGC) module to adaptively propagate information. The graph reasoning is performed among pixels in the same class. Based on the proposed CDGC module, we further introduce the Class-wise Dynamic Graph Convolution Network(CDGCNet), which consists of two main parts including the CDGC module and a basic segmentation network, forming a coarse-to-fine paradigm. Specifically, the CDGC module takes the coarse segmentation result as class mask to extract node features for graph construction and performs dynamic graph convolutions on the constructed graph to learn the feature aggregation and weight allocation. Then the refined feature and the original feature are fused to get the final prediction. We conduct extensive experiments on three popular semantic segmentation benchmarks including Cityscapes, PASCAL VOC 2012 and COCO Stuff, and achieve state-of-the-art performance on all three benchmarks.

cs.CV↗

arXiv2020

Coreference Resolution as Query-based Span Prediction

In this paper, we present an accurate and extensible approach for the coreference resolution task. We formulate the problem as a span prediction task, like in machine reading comprehension (MRC): A query is generated for each candidate mention using its surrounding context, and a span prediction module is employed to extract the text spans of the coreferences within the document using the generated query. This formulation comes with the following key advantages: (1) The span prediction strategy provides the flexibility of retrieving mentions left out at the mention proposal stage; (2) In the MRC framework, encoding the mention and its context explicitly in a query makes it possible to have a deep and thorough examination of cues embedded in the context of coreferent mentions; and (3) A plethora of existing MRC datasets can be used for data augmentation to improve the model's generalization capability. Experiments demonstrate significant performance boost over previous models, with 87.5 (+2.5) F1 score on the GAP benchmark and 83.1 (+3.5) F1 score on the CoNLL-2012 benchmark.

cs.CL↗

arXiv2020

Convenient Real-Time Monitoring of the Contamination of Surface Ion Trap

Recent studies indicated that contamination by adatoms on the surface ion trap can generate contact potential, leading to fluctuations in patch potential. By investigating contamination induced by surface adatoms during a loading process, a direct physical image of the contamination process and the relationship between the capacitance change and the contamination from surface adatoms is examined theoretically and experimentally. From the relationship, the contamination by surface adatoms and the effect of in situ treatment process can be monitored by the capacitance between electrodes in real time. This study is foundational to further research on anomalous heating with practical applications in quantum information processing from surface ion traps.

physics.app-ph↗

arXiv2020

Deep learning to estimate the physical proportion of infected region of lung for COVID-19 pneumonia with CT image set

Utilizing computed tomography (CT) images to quickly estimate the severity of cases with COVID-19 is one of the most straightforward and efficacious methods. Two tasks were studied in this present paper. One was to segment the mask of intact lung in case of pneumonia. Another was to generate the masks of regions infected by COVID-19. The masks of these two parts of images then were converted to corresponding volumes to calculate the physical proportion of infected region of lung. A total of 129 CT image set were herein collected and studied. The intrinsic Hounsfiled value of CT images was firstly utilized to generate the initial dirty version of labeled masks both for intact lung and infected regions. Then, the samples were carefully adjusted and improved by two professional radiologists to generate the final training set and test benchmark. Two deep learning models were evaluated: UNet and 2.5D UNet. For the segment of infected regions, a deep learning based classifier was followed to remove unrelated blur-edged regions that were wrongly segmented out such as air tube and blood vessel tissue etc. For the segmented masks of intact lung and infected regions, the best method could achieve 0.972 and 0.757 measure in mean Dice similarity coefficient on our test benchmark. As the overall proportion of infected region of lung, the final result showed 0.961 (Pearson's correlation coefficient) and 11.7% (mean absolute percent error). The instant proportion of infected regions of lung could be used as a visual evidence to assist clinical physician to determine the severity of the case. Furthermore, a quantified report of infected regions can help predict the prognosis for COVID-19 cases which were scanned periodically within the treatment cycle.

eess.IV↗

arXiv2020

Description Based Text Classification with Reinforcement Learning

The task of text classification is usually divided into two stages: {\it text feature extraction} and {\it classification}. In this standard formalization categories are merely represented as indexes in the label vocabulary, and the model lacks for explicit instructions on what to classify. Inspired by the current trend of formalizing NLP problems as question answering tasks, we propose a new framework for text classification, in which each category label is associated with a category description. Descriptions are generated by hand-crafted templates or using abstractive/extractive models from reinforcement learning. The concatenation of the description and the text is fed to the classifier to decide whether or not the current label should be assigned to the text. The proposed strategy forces the model to attend to the most salient texts with respect to the label, which can be regarded as a hard version of attention, leading to better performances. We observe significant performance boosts over strong baselines on a wide range of text classification tasks including single-label classification, multi-label classification and multi-aspect sentiment analysis.

cs.CL↗

arXiv2020

Population transfer via a finite temperature state

We study quantum population transfer via a common intermediate state initially in thermal equilibrium with a finite temperature $T$, exhibiting a multi-level Stimulated Raman adiabatic passage structure. We consider two situations for the common intermediate state, namely a discrete two-level spin and a bosonic continuum. In both cases we show that the finite temperature strongly affects the efficiency of the population transfer. We also show in the discrete case that strong coupling with the intermediate state, or a longer duration of the controlled pulse would suppress the effect of finite temperature. In the continuous case, we adapt the thermofield-based chain-mapping matrix product states algorithm to study the time evolution of the system plus the continuum under time-dependent controlled pulses, which shows a great potential to be used to solve open quantum system problems in quantum optics.

quant-ph↗