Searcharxiv⌕ Search

arXiv subjects

Zhigang Chen

Publications and source records attributed to Zhigang Chen.

At least 73 records · Page 4Linked to original sources

Nonlinear guiding of light in tunable food-dye solutions

We provide the first demonstration of nonlinear self-trapping of light in solutions of food coloring dyes, leading to dirt cheap soliton waveguides. The optical self-focusing nonlinearity observed in such solutions is attributed to the optical absorption force acting on dye molecules, as the thermal effect at relatively higher power results in a self-defocusing nonlinearity. We found the nonlinear response increases when the laser wavelength is switched to near the peak absorption for all food dyes tested. Furthermore, the absorption-dependent nonlinearity can be tuned by making purposive samples of different food dyes. In particular, with appropriately mixing of dye solutions, nonlinear response arises for otherwise inactive wavelengths, leading to interaction between two beams of different wavelengths and formation of a coupled soliton-like pair. At milliwatt low power levels, self-trapping achieved at one wavelength in food-dye solutions can serve as low-cost waveguides for optical beams at other wavelengths.

physics.optics↗

Knowledge Enhanced Sports Game Summarization

Sports game summarization aims at generating sports news from live commentaries. However, existing datasets are all constructed through automated collection and cleaning processes, resulting in a lot of noise. Besides, current works neglect the knowledge gap between live commentaries and sports news, which limits the performance of sports game summarization. In this paper, we introduce K-SportsSum, a new dataset with two characteristics: (1) K-SportsSum collects a large amount of data from massive games. It has 7,854 commentary-news pairs. To improve the quality, K-SportsSum employs a manual cleaning process; (2) Different from existing datasets, to narrow the knowledge gap, K-SportsSum further provides a large-scale knowledge corpus that contains the information of 523 sports teams and 14,724 sports players. Additionally, we also introduce a knowledge-enhanced summarizer that utilizes both live commentaries and the knowledge to generate sports news. Extensive experiments on K-SportsSum and SportsSum datasets show that our model achieves new state-of-the-art performances. Qualitative analysis and human study further verify that our model generates more informative sports news.

cs.CL↗

Topologically tuned terahertz confinement in a nonlinear photonic chip

Compact terahertz (THz) functional devices are greatly sought after for high-speed wireless communication, biochemical sensing, and non-destructive inspection. However, conventional devices to generate and guide THz waves are afflicted with diffraction loss and disorder due to inevitable fabrication defects. Here, based on the topological protection of electromagnetic waves, we demonstrate nonlinear generation and topologically tuned confinement of THz waves in a judiciously-patterned lithium niobate chip forming a wedge-shaped Su-Schrieffer-Heeger lattice. Experimentally measured band structures provide direct visualization of the generated THz waves in momentum space, and their robustness to chiral perturbation is also analyzed and compared between topologically trivial and nontrivial regimes. Such chip-scale control of THz waves may bring about new possibilities for THz integrated topological circuits, promising for advanced photonic applications.

physics.optics↗

Airy beams and accelerating waves: An overview of recent advances

Over the last dozen of years, the area of accelerating waves has made considerable advances not only in terms of fundamentals and experimental demonstrations but also in connection to a wide range of applications. Starting from the prototypical Airy beam that was proposed and observed in 2007, new families of accelerating waves have been identified in the paraxial and nonparaxial domains in space and/or time, with different methods developed to control at will their trajectory, amplitude, and beam width. Accelerating optical waves exhibit a number of highly desirable attributes. They move along a curved or accelerating trajectory while being resilient to perturbations (self-healing), and, are diffraction-free. It is because of these particular features that accelerating waves have been utilized in a variety of applications in the areas of filamentation, beam focusing, particle manipulation, biomedical imaging, plasmons, and material processing among others.

physics.optics↗

Affective Decoding for Empathetic Response Generation

Understanding speaker's feelings and producing appropriate responses with emotion connection is a key communicative skill for empathetic dialogue systems. In this paper, we propose a simple technique called Affective Decoding for empathetic response generation. Our method can effectively incorporate emotion signals during each decoding step, and can additionally be augmented with an auxiliary dual emotion encoder, which learns separate embeddings for the speaker and listener given the emotion base of the dialogue. Extensive empirical studies show that our models are perceived to be more empathetic by human evaluations, in comparison to several strong mainstream methods for empathetic responding.

cs.CL↗

SportsSum2.0: Generating High-Quality Sports News from Live Text Commentary

Sports game summarization aims to generate news articles from live text commentaries. A recent state-of-the-art work, SportsSum, not only constructs a large benchmark dataset, but also proposes a two-step framework. Despite its great contributions, the work has three main drawbacks: 1) the noise existed in SportsSum dataset degrades the summarization performance; 2) the neglect of lexical overlap between news and commentaries results in low-quality pseudo-labeling algorithm; 3) the usage of directly concatenating rewritten sentences to form news limits its practicability. In this paper, we publish a new benchmark dataset SportsSum2.0, together with a modified summarization framework. In particular, to obtain a clean dataset, we employ crowd workers to manually clean the original dataset. Moreover, the degree of lexical overlap is incorporated into the generation of pseudo labels. Further, we introduce a reranker-enhanced summarizer to take into account the fluency and expressiveness of the summarized news. Extensive experiments show that our model outperforms the state-of-the-art baseline.

cs.CL↗

Realization of second-order photonic square-root topological insulators

Square-root higher-order topological insulators (HOTIs) are recently discovered new topological phases, with intriguing topological properties inherited from a parent lattice Hamiltonian. Different from conventional HOTIs, the square-root HOTIs typically manifest two paired non-zero energy corner states. In this work, we experimentally demonstrate the second-order square-root HOTIs in photonics for the first time to our knowledge, thereby unveiling such distinct corner states. The specific platform is a laser-written decorated honeycomb lattice (HCL), for which the squared Hamiltonian represents a direct sum of the underlying HCL and breathing Kagome lattice. The localized corner states residing in different bandgaps are observed with characteristic phase structures, in sharp contrast to discrete diffraction in a topologically trivial structure. Our work illustrates a scheme to study fundamental topological phenomena in systems with coexistence of spin-1/2 and spin-1 Dirac-Weyl fermions, and may bring about new possibilities in topology-driven photonic devices.

physics.optics↗

Program Enhanced Fact Verification with Verbalization and Graph Attention Network

Performing fact verification based on structured data is important for many real-life applications and is a challenging research problem, particularly when it involves both symbolic operations and informal inference based on language understanding. In this paper, we present a Program-enhanced Verbalization and Graph Attention Network (ProgVGAT) to integrate programs and execution into textual inference models. Specifically, a verbalization with program execution model is proposed to accumulate evidences that are embedded in operations over the tables. Built on that, we construct the graph attention verification networks, which are designed to fuse different sources of evidences from verbalized program execution, program structures, and the original statements and tables, to make the final verification decision. To support the above framework, we propose a program selection module optimized with a new training strategy based on margin loss, to produce more accurate programs, which is shown to be effective in enhancing the final verification results. Experimental results show that the proposed framework achieves the new state-of-the-art performance, a 74.4% accuracy, on the benchmark dataset TABFACT.

cs.AI↗

Detecting Speaker Personas from Conversational Texts

Personas are useful for dialogue response prediction. However, the personas used in current studies are pre-defined and hard to obtain before a conversation. To tackle this issue, we study a new task, named Speaker Persona Detection (SPD), which aims to detect speaker personas based on the plain conversational text. In this task, a best-matched persona is searched out from candidates given the conversational text. This is a many-to-many semantic matching task because both contexts and personas in SPD are composed of multiple sentences. The long-term dependency and the dynamic redundancy among these sentences increase the difficulty of this task. We build a dataset for SPD, dubbed as Persona Match on Persona-Chat (PMPC). Furthermore, we evaluate several baseline models and propose utterance-to-profile (U2P) matching networks for this task. The U2P models operate at a fine granularity which treat both contexts and personas as sets of multiple sequences. Then, each sequence pair is scored and an interpretable overall score is obtained for a context-persona pair through aggregation. Evaluation results show that the U2P models outperform their baseline counterparts significantly.

cs.CL↗

Higher-order exceptional point and Landau-Zener Bloch oscillations in driven non-Hermitian photonic Lieb lattices

We propose a scheme to realize parity-time (PT) symmetric photonic Lieb lattices of ribbon shape and complex couplings, thereby demonstrating the higher-order exceptional point (EP) and Landau-Zener Bloch (LZB) oscillations in presence of a refractive index gradient. Quite different from non-Hermitian flatband lattices with on-site gain/loss, which undergo thresholdless PT symmetry breaking, the spectrum for such quasi-one-dimensional Lieb lattices has completely real values when the index gradient is applied perpendicular to the ribbon, and a triply degenerated (third-order) EP with coalesced eigenvalues and eigenvectors emerges only when the amplitude of gain/loss ratio reaches a certain threshold value. When the index gradient is applied parallel to the ribbon, the LZB oscillations exhibit intriguing characteristics including asymmetric energy transition and pseudo-Hermitian propagation as the flatband is excited. Meanwhile, a secondary emission occurs each time when the oscillatory motion passes through the EP, leading to distinct energy distribution in the flatband when a dispersive band is excited. Such novel phenomena may appear in other non-Hermitian flatband systems. Our work may also bring insight and suggest a photonic platform to study the symmetry and topological characterization of higher-order EPs that may find unique applications in for example enhancing sensitivity.

physics.optics↗

Fractal-like photonic lattices and localized states arising from singular and nonsingular flatbands

We realize fractal-like photonic lattices using cw-laser-writing technique, thereby observe distinct compact localized states (CLSs) associated with different flatbands in the same lattice setting. Such triangle-shaped lattices, akin to the first generation Sierpinski lattices, possess a band structure where singular non-degenerate and nonsingular degenerate flatbands coexist. By proper phase modulation of an input excitation beam, we demonstrate experimentally not only the simplest CLSs but also their superimposition into other complex mode structures. Furthermore, we show by numerical simulation a dynamical oscillation of the flatband states due to beating of the CLSs that have different eigenenergies. These results may provide inspiration for exploring fundamental phenomena arising from fractal structure, flatband singularity, and real-space topology.

physics.optics↗

Multi-Modal Chorus Recognition for Improving Song Search

We discuss a novel task, Chorus Recognition, which could potentially benefit downstream tasks such as song search and music summarization. Different from the existing tasks such as music summarization or lyrics summarization relying on single-modal information, this paper models chorus recognition as a multi-modal one by utilizing both the lyrics and the tune information of songs. We propose a multi-modal Chorus Recognition model that considers diverse features. Besides, we also create and publish the first Chorus Recognition dataset containing 627 songs for public use. Our empirical study performed on the dataset demonstrates that our approach outperforms several baselines in chorus recognition. In addition, our approach also helps to improve the accuracy of its downstream task - song search by more than 10.6%.

cs.IR↗

Empower Distantly Supervised Relation Extraction with Collaborative Adversarial Training

With recent advances in distantly supervised (DS) relation extraction (RE), considerable attention is attracted to leverage multi-instance learning (MIL) to distill high-quality supervision from the noisy DS. Here, we go beyond label noise and identify the key bottleneck of DS-MIL to be its low data utilization: as high-quality supervision being refined by MIL, MIL abandons a large amount of training instances, which leads to a low data utilization and hinders model training from having abundant supervision. In this paper, we propose collaborative adversarial training to improve the data utilization, which coordinates virtual adversarial training (VAT) and adversarial training (AT) at different levels. Specifically, since VAT is label-free, we employ the instance-level VAT to recycle instances abandoned by MIL. Besides, we deploy AT at the bag-level to unleash the full potential of the high-quality supervision got by MIL. Our proposed method brings consistent improvements (~ 5 absolute AUC score) to the previous state of the art, which verifies the importance of the data utilization issue and the effectiveness of our method.

cs.CL↗

CIL: Contrastive Instance Learning Framework for Distantly Supervised Relation Extraction

The journey of reducing noise from distant supervision (DS) generated training data has been started since the DS was first introduced into the relation extraction (RE) task. For the past decade, researchers apply the multi-instance learning (MIL) framework to find the most reliable feature from a bag of sentences. Although the pattern of MIL bags can greatly reduce DS noise, it fails to represent many other useful sentence features in the datasets. In many cases, these sentence features can only be acquired by extra sentence-level human annotation with heavy costs. Therefore, the performance of distantly supervised RE models is bounded. In this paper, we go beyond typical MIL framework and propose a novel contrastive instance learning (CIL) framework. Specifically, we regard the initial MIL as the relational triple encoder and constraint positive pairs against negative pairs for each instance. Experiments demonstrate the effectiveness of our proposed framework, with significant improvements over the previous methods on NYT10, GDS and KBP.

cs.CL↗

Nonlinear control of photonic higher-order topological bound states in the continuum

Higher-order topological insulators (HOTIs) are recently discovered topological phases, possessing symmetry-protected corner states with fractional charges. An unexpected connection between these states and the seemingly unrelated phenomenon of bound states in the continuum (BICs) was recently unveiled. When nonlinearity is added to a HOTI system, a number of fundamentally important questions arise. For example, how does nonlinearity couple higher-order topological BICs with the rest of the system, including continuum states? In fact, thus far BICs in nonlinear HOTIs have remained unexplored. Here, we demonstrate the interplay of nonlinearity, higher-order topology, and BICs in a photonic platform. We observe topological corner states which, serendipitously, are also BICs in a laser-written second-order topological lattice. We further demonstrate nonlinear coupling with edge states at a low nonlinearity, transitioning to solitons at a high nonlinearity. Theoretically, we calculate the analog of the Zak phase in the nonlinear regime, illustrating that a topological BIC can be actively tuned by both focusing and defocusing nonlinearities. Our studies are applicable to other nonlinear HOTI systems, with promising applications in emerging topology-driven devices.

physics.optics↗

Partner Matters! An Empirical Study on Fusing Personas for Personalized Response Selection in Retrieval-Based Chatbots

Persona can function as the prior knowledge for maintaining the consistency of dialogue systems. Most of previous studies adopted the self persona in dialogue whose response was about to be selected from a set of candidates or directly generated, but few have noticed the role of partner in dialogue. This paper makes an attempt to thoroughly explore the impact of utilizing personas that describe either self or partner speakers on the task of response selection in retrieval-based chatbots. Four persona fusion strategies are designed, which assume personas interact with contexts or responses in different ways. These strategies are implemented into three representative models for response selection, which are based on the Hierarchical Recurrent Encoder (HRE), Interactive Matching Network (IMN) and Bidirectional Encoder Representations from Transformers (BERT) respectively. Empirical studies on the Persona-Chat dataset show that the partner personas neglected in previous studies can improve the accuracy of response selection in the IMN- and BERT-based models. Besides, our BERT-based model implemented with the context-response-aware persona fusion strategy outperforms previous methods by margins larger than 2.7% on original personas and 4.6% on revised personas in terms of hits@1 (top-1 accuracy), achieving a new state-of-the-art performance on the Persona-Chat dataset.

cs.CL↗

Giant Enhancement of Nonlinear Harmonic Generation in a Silicon Topological Photonic Crystal Nanocavity Chain

Strongly enhanced third-harmonic generation (THG) by the topological localization of an edge mode in a Su-Schrieffer-Heeger (SSH) chain of silicon photonic crystal nanocavities is demonstrated. The edge mode of the nanocavity chain not only naturally inherits resonant properties of the single nanocavity, but also exhibits the topological feature with mode robustness extending well beyond individual nanocavity. By engineering the SSH nanocavities with alternating strong and weak coupling strengths on a silicon slab, we observe the edge mode formation that entails a THG signal with three orders of magnitude enhancement compared with that in a trivial SSH structure. Our results indicate that the photonic crystal nanocavity chain could provide a promising on-chip platform for topology-driven nonlinear photonics.

physics.optics↗

ExpMRC: Explainability Evaluation for Machine Reading Comprehension

Achieving human-level performance on some of Machine Reading Comprehension (MRC) datasets is no longer challenging with the help of powerful Pre-trained Language Models (PLMs). However, it is necessary to provide both answer prediction and its explanation to further improve the MRC system's reliability, especially for real-life applications. In this paper, we propose a new benchmark called ExpMRC for evaluating the explainability of the MRC systems. ExpMRC contains four subsets, including SQuAD, CMRC 2018, RACE$^+$, and C$^3$ with additional annotations of the answer's evidence. The MRC systems are required to give not only the correct answer but also its explanation. We use state-of-the-art pre-trained language models to build baseline systems and adopt various unsupervised approaches to extract evidence without a human-annotated training set. The experimental results show that these models are still far from human performance, suggesting that the ExpMRC is challenging. Resources will be available through https://github.com/ymcui/expmrc

cs.CL↗