arXiv subjects

Yue Hu

Publications and source records attributed to Yue Hu.

At least 253 records · Page 14Linked to original sources

arXiv2022

CLSEG: Contrastive Learning of Story Ending Generation

Story Ending Generation (SEG) is a challenging task in natural language generation. Recently, methods based on Pre-trained Language Models (PLM) have achieved great prosperity, which can produce fluent and coherent story endings. However, the pre-training objective of PLM-based methods is unable to model the consistency between story context and ending. The goal of this paper is to adopt contrastive learning to generate endings more consistent with story context, while there are two main challenges in contrastive learning of SEG. First is the negative sampling of wrong endings inconsistent with story contexts. The second challenge is the adaptation of contrastive learning for SEG. To address these two issues, we propose a novel Contrastive Learning framework for Story Ending Generation (CLSEG), which has two steps: multi-aspect sampling and story-specific contrastive learning. Particularly, for the first issue, we utilize novel multi-aspect sampling mechanisms to obtain wrong endings considering the consistency of order, causality, and sentiment. To solve the second issue, we well-design a story-specific contrastive training strategy that is adapted for SEG. Experiments show that CLSEG outperforms baselines and can produce story endings with stronger consistency and rationality.

cs.CL↗

arXiv2022

Modeling Intention, Emotion and External World in Dialogue Systems

Intention, emotion and action are important elements in human activities. Modeling the interaction process between individuals by analyzing the relationships between these elements is a challenging task. However, previous work mainly focused on modeling intention and emotion independently, and neglected of exploring the mutual relationships between intention and emotion. In this paper, we propose a RelAtion Interaction Network (RAIN), consisting of Intention Relation Module and Emotion Relation Module, to jointly model mutual relationships and explicitly integrate historical intention information. The experiments on the dataset show that our model can take full advantage of the intention, emotion and action between individuals and achieve a remarkable improvement over BERT-style baselines. Qualitative analysis verifies the importance of the mutual interaction between the intention and emotion.

cs.CL↗

arXiv2022

Superdiffusion of Cosmic Rays in Compressible Magnetized Turbulence

Owing to the complexity of turbulent magnetic fields, modeling the diffusion of cosmic rays is challenging. Based on the current understanding of anisotropic magnetohydrodynamic (MHD) turbulence, we use test particles to examine the cosmic rays' superdiffusion in the direction perpendicular to the mean magnetic field. By changing Alfven Mach number $M_A$ and sonic Mach number $M_S$ of compressible MHD simulations, our study covers a wide range of astrophysical conditions including subsonic warm gas phase and supersonic cold molecular gas. We show that freely streaming cosmic rays' perpendicular displacement increases as 3/2 to the power of the time traveled along local magnetic field lines. This power-law index changes to 3/4 if the parallel propagation is diffusive. We find that the cosmic rays' parallel mean free path decreases in a power-law relation of $M_A^{-2}$ in supersonic turbulence. We investigate the energy fraction of slow, fast, and Alfvenic modes and confirm the dominance of Alfvenic modes in the perpendicular superdiffusion. In particular, the energy fraction of fast mode, which is the main agent for pitch-angle scattering, increases with $M_A$ but is insensitive to $M_S \ge 2$. Accordingly, our results suggest that the suppressed diffusion in supersonic molecular clouds arises primarily due to the variations of $M_A$ instead of $M_S$.

astro-ph.GA↗

arXiv2022

Multi-scale Magnetic Fields in the Central Molecular Zone: Inference from the Gradient Technique

The central molecular zone (CMZ) plays an essential role in regulating the nuclear ecosystem of our Galaxy. To get an insight into the magnetic fields of the CMZ, we employ the Gradient Technique (GT), which is rooted in the anisotropy of magnetohydrodynamic turbulence. Our analysis is based on the data of multiple wavelengths, including molecular emission lines, radio 1.4 GHz continuum image, and Herschel 70 $μ$m image, as well as ionized [Ne II] and Paschen-alpha emissions. The results are compared with the observations of Planck 353 GHz and High-resolution Airborne Wideband Camera Plus (HWAC+) 53 $μ$m polarized dust emissions. We map the orientation of the magnetic field at multiple wavelengths across the central molecular zone, including close-ups of the Radio Arc and Sagittarius A West regions, on multi scales from $\sim$ 0.1 pc to 10 pc. The magnetic fields towards the central molecular zone traced by GT are globally compatible with the polarization measurements, accounting for the contribution from the galactic foreground and background. This correspondence suggests that the magnetic field and turbulence are dynamically crucial in the galactic center. We find that the magnetic fields associated with the Arched filaments and the thermal components of the Radio Arc are in good agreement with the HAWC+ polarization. Our measurement towards the non-thermal Radio Arc reveals the poloidal magnetic field components in the galactic center. For Sagittarius A West region, we find a great agreement between the GT measurement using [Ne II] emission and HWAC+ 53 $μ$m observation. We use GT to predict the magnetic fields associated with ionized Paschen-alpha gas down to scales of 0.1 pc.

astro-ph.GA↗

arXiv2021

Velocity Gradients: Magnetic Field Tomography towards the Supernova Remnant W44

As a novel approach for tracing interstellar magnetic fields, the Velocity Gradient Technique (VGT) has been proven to be effective for probing magnetic fields in the diffuse interstellar medium (ISM). In this work, we verify the VGT in a broader context by applying the technique to a molecular cloud interacting with the supernovae remnant (SNR) W44. We probe the magnetic fields with the VGT using CO, $\rm HCO^+$, and H I emission lines and make a comparison with the Planck 353 GHZ dust polarization. We show that the VGT gives an accurate measurement that coheres with the Planck polarization especially in intense molecular gas emission regions. We further study the foreground's contribution to the polarization that results in a misalignment between the VGT and the Planck measurements in low-intensity molecular gas areas. We advance the VGT to achieve magnetic field tomography by decomposing the W44 into various velocity components. We show that W44's velocity component at $v\sim45$ km s$^{-1}$ exhibits the largest coverage and gives the best agreement with Planck polarization in terms of magnetic field orientation.

astro-ph.GA↗

arXiv2021

Who Are the `Silent Spreaders'?: Contact Tracing in Spatio-Temporal Memory Models

The COVID-19 epidemic has swept the world for over a year. However, a large number of infectious asymptomatic COVID-19 cases (\textit{ACC}s) are still making the breaking up of the transmission chains very difficult. Efforts by epidemiological researchers in many countries have thrown light on the clinical features of ACCs, but there is still a lack of practical approaches to detect ACCs so as to help contain the pandemic. To address the issue of ACCs, this paper presents a neural network model called Spatio-Temporal Episodic Memory for COVID-19 (\textit{STEM-COVID}) to identify ACCs from contact tracing data. Based on the fusion Adaptive Resonance Theory (\textit{ART}), the model encodes a collective spatio-temporal episodic memory of individuals and incorporates an effective mechanism of parallel searches for ACCs. Specifically, the episodic traces of the identified positive cases are used to map out the episodic traces of suspected ACCs using a weighted evidence pooling method. To evaluate the efficacy of STEM-COVID, a realistic agent based simulation model for COVID-19 spreading is implemented based on the recent epidemiological findings on ACCs. The experiments based on rigorous simulation scenarios, manifesting the current situation of COVID-19 spread, show that the STEM-COVID model with weighted evidence pooling has a higher level of accuracy and efficiency for identifying ACCs when compared with several baselines. Moreover, the model displays strong robustness against noisy data and different ACC proportions, which partially reflects the effect of breakthrough infections after vaccination on the virus transmission.

cs.SI↗

arXiv2021

Optimized phonon band discretization scheme for efficiently solving the non-gray Boltzmann transport equation

Phonon Boltzmann transport equation (BTE) is an important tool for studying the nanoscale thermal transport. Because phonons have a large spread in their properties, the non-gray (i.e. considering different phonon bands) phonon BTE is needed to accurately capture the nanoscale transport phenomena. However, BTE solvers generally require large computational cost. Non-gray modeling imposes significant additional complexity to the numerical simulations, which hinders the modeling of real nanoscale systems. In this work, we address this issue by a systematic investigation on the phonon band discretization scheme using real material properties of four representative materials, including silicon, gallium arsenide, diamond, and lead telluride. We find that the schemes used in previous studies require at least a few tens of bands to ensure the accuracy, which requires large computational costs. We then propose an improved band discretization scheme, in which we divide the mean free path domain into two subdomains, one on either side of the inflection point of the mean free path accumulated thermal conductivity and adopt the Gauss-Legendre quadrature for each subdomain. With this scheme, the solution of phonon BTE converges (error < 1%) with less than 10 phonon bands for all these materials. The proposed scheme allows to significantly reduce the time and memory consumption of the numerical BTE solver, which is an important step towards large-scale phonon BTE simulation for real materials.

physics.comp-ph↗

arXiv2021

Streaming data preprocessing via online tensor recovery for large environmental sensor networks

Measuring the built and natural environment at a fine-grained scale is now possible with low-cost urban environmental sensor networks. However, fine-grained city-scale data analysis is complicated by tedious data cleaning including removing outliers and imputing missing data. While many methods exist to automatically correct anomalies and impute missing entries, challenges still exist on data with large spatial-temporal scales and shifting patterns. To address these challenges, we propose an online robust tensor recovery (OLRTR) method to preprocess streaming high-dimensional urban environmental datasets. A small-sized dictionary that captures the underlying patterns of the data is computed and constantly updated with new data. OLRTR enables online recovery for large-scale sensor networks that provide continuous data streams, with a lower computational memory usage compared to offline batch counterparts. In addition, we formulate the objective function so that OLRTR can detect structured outliers, such as faulty readings over a long period of time. We validate OLRTR on a synthetically degraded National Oceanic and Atmospheric Administration temperature dataset, with a recovery error of 0.05, and apply it to the Array of Things city-scale sensor network in Chicago, IL, showing superior results compared with several established online and batch-based low rank decomposition methods.

cs.LG↗

arXiv2021

Know Deeper: Knowledge-Conversation Cyclic Utilization Mechanism for Open-domain Dialogue Generation

End-to-End intelligent neural dialogue systems suffer from the problems of generating inconsistent and repetitive responses. Existing dialogue models pay attention to unilaterally incorporating personal knowledge into the dialog while ignoring the fact that incorporating the personality-related conversation information into personal knowledge taken as the bilateral information flow boosts the quality of the subsequent conversation. Besides, it is indispensable to control personal knowledge utilization over the conversation level. In this paper, we propose a conversation-adaption multi-view persona aware response generation model that aims at enhancing conversation consistency and alleviating the repetition from two folds. First, we consider conversation consistency from multiple views. From the view of the persona profile, we design a novel interaction module that not only iteratively incorporates personalized knowledge into each turn conversation but also captures the personality-related information from conversation to enhance personalized knowledge semantic representation. From the view of speaking style, we introduce the speaking style vector and feed it into the decoder to keep the speaking style consistency. To avoid conversation repetition, we devise a coverage mechanism to keep track of the activation of personal knowledge utilization. Experiments on both automatic and human evaluation verify the superiority of our model over previous models.

cs.CL↗

arXiv2021

Coarse-to-Careful: Seeking Semantic-related Knowledge for Open-domain Commonsense Question Answering

It is prevalent to utilize external knowledge to help machine answer questions that need background commonsense, which faces a problem that unlimited knowledge will transmit noisy and misleading information. Towards the issue of introducing related knowledge, we propose a semantic-driven knowledge-aware QA framework, which controls the knowledge injection in a coarse-to-careful fashion. We devise a tailoring strategy to filter extracted knowledge under monitoring of the coarse semantic of question on the knowledge extraction stage. And we develop a semantic-aware knowledge fetching module that engages structural knowledge information and fuses proper knowledge according to the careful semantic of questions in a hierarchical way. Experiments demonstrate that the proposed approach promotes the performance on the CommonsenseQA dataset comparing with strong baselines.

cs.CL↗

arXiv2021

CogTree: Cognition Tree Loss for Unbiased Scene Graph Generation

Scene graphs are semantic abstraction of images that encourage visual understanding and reasoning. However, the performance of Scene Graph Generation (SGG) is unsatisfactory when faced with biased data in real-world scenarios. Conventional debiasing research mainly studies from the view of balancing data distribution or learning unbiased models and representations, ignoring the correlations among the biased classes. In this work, we analyze this problem from a novel cognition perspective: automatically building a hierarchical cognitive structure from the biased predictions and navigating that hierarchy to locate the relationships, making the tail relationships receive more attention in a coarse-to-fine mode. To this end, we propose a novel debiasing Cognition Tree (CogTree) loss for unbiased SGG. We first build a cognitive structure CogTree to organize the relationships based on the prediction of a biased SGG model. The CogTree distinguishes remarkably different relationships at first and then focuses on a small portion of easily confused ones. Then, we propose a debiasing loss specially for this cognitive structure, which supports coarse-to-fine distinction for the correct relationships. The loss is model-agnostic and consistently boosting the performance of several state-of-the-art models. The code is available at: https://github.com/CYVincent/Scene-Graph-Transformer-CogTree.

cs.CV↗

arXiv2021

MCR-Net: A Multi-Step Co-Interactive Relation Network for Unanswerable Questions on Machine Reading Comprehension

Question answering systems usually use keyword searches to retrieve potential passages related to a question, and then extract the answer from passages with the machine reading comprehension methods. However, many questions tend to be unanswerable in the real world. In this case, it is significant and challenging how the model determines when no answer is supported by the passage and abstains from answering. Most of the existing systems design a simple classifier to determine answerability implicitly without explicitly modeling mutual interaction and relation between the question and passage, leading to the poor performance for determining the unanswerable questions. To tackle this problem, we propose a Multi-Step Co-Interactive Relation Network (MCR-Net) to explicitly model the mutual interaction and locate key clues from coarse to fine by introducing a co-interactive relation module. The co-interactive relation module contains a stack of interaction and fusion blocks to continuously integrate and fuse history-guided and current-query-guided clues in an explicit way. Experiments on the SQuAD 2.0 and DuReader datasets show that our model achieves a remarkable improvement, outperforming the BERT-style baselines in literature. Visualization analysis also verifies the importance of the mutual interaction between the question and passage.

cs.CL↗

arXiv2021

Anisotropic Turbulence in Position-Position-Velocity Space: Probing Three-Dimensional Magnetic Fields

Direct measurements of three-dimensional magnetic fields in the interstellar medium (ISM) are not achievable. However, the anisotropic nature of magnetohydrodynamic (MHD) turbulence provides a novel way of tracing the magnetic fields. Guided by the advanced understanding of turbulence's anisotropy in the Position-Position-Velocity (PPV) space, we extend the Structure-Function Analysis (SFA) to measure both the three-dimensional magnetic field orientation and Alfven Mach number $M_A$, which provides the information on magnetic field strength. Following the theoretical framework developed in Kandel et al. (2016), we find that the anisotropy in a given velocity channel is affected by the inclination angle between the 3D magnetic field direction and the line-of-sight as well as media magnetization. We analyze the synthetic PPV cubes generated by incompressible and compressible MHD simulations. We confirm that the PPV channel's intensity fluctuations measured in various position angles reveal plane-of-the-sky magnetic field orientation. We show that by varying the channel width, the anisotropies of the intensity fluctuations in PPV space can be used to simultaneously estimate both magnetic field inclination angle and strength of total magnetic fields.

astro-ph.GA↗

arXiv2021

Super-exponential scrambling of Out-of-time-ordered correlators

Out-of-time-ordered correlators (OTOCs) are an effective tool in characterizing black hole chaos, many-body thermalization and quantum dynamics instability. Previous research findings have shown that the OTOCs' exponential growth (EG) marks the limit for quantum systems. However, we report in this letter a periodically-modulated nonlinear Schrödinger system, in which we interestingly find a novel way of information scrambling: super-EG. We show that the quantum OTOCs' growth, which stems from the quantum chaotic dynamics, will increase in a super-exponential way. We also find that in the classical limit, the hyper-chaos revealed by a linearly-increasing Lyapunov exponent actually triggers the super-EG of classical OTOCs. The results in this paper break the restraints of EG as the limit for quantum systems, which give us new insight into the nature of information scrambling in various fields of physics from black hole to many-body system.

quant-ph↗

arXiv2021

On Learning Universal Representations Across Languages

Recent studies have demonstrated the overwhelming advantage of cross-lingual pre-trained models (PTMs), such as multilingual BERT and XLM, on cross-lingual NLP tasks. However, existing approaches essentially capture the co-occurrence among tokens through involving the masked language model (MLM) objective with token-level cross entropy. In this work, we extend these approaches to learn sentence-level representations and show the effectiveness on cross-lingual understanding and generation. Specifically, we propose a Hierarchical Contrastive Learning (HiCTL) method to (1) learn universal representations for parallel sentences distributed in one or multiple languages and (2) distinguish the semantically-related words from a shared cross-lingual vocabulary for each sentence. We conduct evaluations on two challenging cross-lingual tasks, XTREME and machine translation. Experimental results show that the HiCTL outperforms the state-of-the-art XLM-R by an absolute gain of 4.2% accuracy on the XTREME benchmark as well as achieves substantial improvements on both of the high-resource and low-resource English-to-X translation tasks over strong baselines.

cs.CL↗

arXiv2021

Revealing Gravitational Collapse in Serpens G3-G6 Molecular Cloud using Velocity Gradients

The relative role of turbulence, magnetic fields, self-gravity in star formation is a subject of intensive debate. We present IRAM 30m telescope observations of the $^{13}$CO (1-0) emission in the Serpens G3-G6 molecular cloud and apply to the data a set of statistical methods. Those include the probability density functions (PDFs) of column density and the Velocity Gradients Technique (VGT). We combine our data with the Planck 353 GHz polarized dust emission observations, Hershel H$_2$ column density. We suggest that the Serpens G3-G6 south clump is undergoing a gravitational collapse. Our analysis reveals that the gravitational collapse happens at volume density $n\ge10^3$ $\rm cm^{-3}$. We estimate the plane-of-the-sky magnetic field strength of approximately 120 $μG$ using the traditional Davis-Chandrasekhar-Fermi method and 100 $μG$ using a new technique proposed in Lazarian et al.(2020). We find the Serpens G3-G6 south clump's total magnetic field energy significantly surpasses kinetic energy and gravitational energy. We conclude that the gravitational collapse could be successfully triggered in a supersonic and sub-Alfvénic cloud.

astro-ph.GA↗

arXiv2021

End-to-End Human Object Interaction Detection with HOI Transformer

We propose HOI Transformer to tackle human object interaction (HOI) detection in an end-to-end manner. Current approaches either decouple HOI task into separated stages of object detection and interaction classification or introduce surrogate interaction problem. In contrast, our method, named HOI Transformer, streamlines the HOI pipeline by eliminating the need for many hand-designed components. HOI Transformer reasons about the relations of objects and humans from global image context and directly predicts HOI instances in parallel. A quintuple matching loss is introduced to force HOI predictions in a unified way. Our method is conceptually much simpler and demonstrates improved accuracy. Without bells and whistles, HOI Transformer achieves $26.61\% $ $ AP $ on HICO-DET and $52.9\%$ $AP_{role}$ on V-COCO, surpassing previous methods with the advantage of being much simpler. We hope our approach will serve as a simple and effective alternative for HOI tasks. Code is available at https://github.com/bbepoch/HoiTransformer .

cs.CV↗

arXiv2021

IIE-NLP-Eyas at SemEval-2021 Task 4: Enhancing PLM for ReCAM with Special Tokens, Re-Ranking, Siamese Encoders and Back Translation

This paper introduces our systems for all three subtasks of SemEval-2021 Task 4: Reading Comprehension of Abstract Meaning. To help our model better represent and understand abstract concepts in natural language, we well-design many simple and effective approaches adapted to the backbone model (RoBERTa). Specifically, we formalize the subtasks into the multiple-choice question answering format and add special tokens to abstract concepts, then, the final prediction of question answering is considered as the result of subtasks. Additionally, we employ many finetuning tricks to improve the performance. Experimental results show that our approaches achieve significant performance compared with the baseline systems. Our approaches achieve eighth rank on subtask-1 and tenth rank on subtask-2.

cs.CL↗