no code implementations • EMNLP 2021 • Shuqun Li, Liang Yang, Weidong He, Shiqi Zhang, Jingjie Zeng, Hongfei Lin
At the sentence level, we leverage the metaphor information of words that except the target word in the sentence to strengthen the reasoning ability of our model via a novel label-enhanced contextualized representation.
no code implementations • CCL 2021 • Tongyue Zhang, Shaowu Zhang, Bo Xu, Liang Yang, Hongfei Lin
“幽默在人类交流中扮演着重要角色, 并大量存在于情景喜剧中。笑点(punchline)是情景喜剧实现幽默效果的形式之一, 在情景喜剧笑点识别任务中, 每条句子的标签代表该句是否为笑点, 但是以往的笑点识别工作通常只通过建模上下文语义关系识别笑点, 对标签的利用并不充分。为了充分利用标签序列中的信息, 本文提出了一种新的识别方法, 即结合条件随机场的单词级-句子级多任务学习模型, 该模型在两方面进行了改进, 首先将标签序列中相邻两个标签之间的转移关系看作幽默理论中不一致性的一种体现, 并使用条件随机场学习这种转移关系, 其次由于学习相邻标签之间的转移关系以及上下文语义关系均能够学习到铺垫和笑点之间的不一致性, 两者之间存在相关性, 为了使模型通过利用这种相关性提高笑点识别的效果, 该模型引入了多任务学习方法, 使用多任务学习方法同时学习每条句子的句义、组成每条句子的所有字符的词义, 单词级别的标签转移关系以及句子级别的标签转移关系。本文在CCL2020“小牛杯”幽默计算—情景喜剧笑点识别评测任务的英文数据集上进行实验, 结果表明, 本文提出的方法比目前最好的方法提高了3. 2%, 在情景喜剧幽默笑点识别任务上取得了最好的效果, 并通过消融实验证明了上述两方面改进的有效性。”
no code implementations • CCL 2020 • Jinhui Zhang, Shaowu Zhang, Xiaochao Fan, Liang Yang, Hongfei Lin
幽默在人们日常交流中发挥着重要作用。随着人工智能的快速发展, 幽默等级识别成为自然语言处理领域的热点研究问题之一。已有的幽默等级识别研究往往将幽默文本看作一个整体, 忽视了幽默文本内部的语义关系。本文将幽默等级识别视为自然语言推理任务, 将幽默文本划分为“铺垫”和“笑点”两个部分, 分别对其语义和语义关系进行建模, 提出了一种多粒度语义交互理解网络, 从单词和子句两个粒度捕获幽默文本中语义的关联和交互。本文在Reddit公开幽默数据集上进行了实验, 相比之前最优结果, 模型在语料上的准确率提升了1. 3%。实验表明, 引入幽默内部的语义关系信息可以提高模型幽默识别的性能, 而本文提出的模型也可以很好地建模这种语义关系。
no code implementations • Findings (EMNLP) 2021 • Changrong Min, Yonghe Chu, Liang Yang, Bo Xu, Hongfei Lin
Thus, cosine similarity cannot approximate distances on the manifold.
no code implementations • CCL 2021 • Dongzhen Wen, Fan Zhang, Xiao Zhang, Liang Yang, Yuan Lin, Bo Xu, Hongfei Lin
“软件源代码的理解则是软件协同开发与维护的核心, 而源代码中占半数以上的标识符的理解则在软件理解中起到重要作用, 传统软件工程主要研究通过命名规范限制标识符的命名过程以构造更易理解和交流的标识符。本文则在梳理分析常见编程语言命名规范的基础上, 提出一种全新的标识符可理解性评价标准。具体而言, 本文首先总结梳理了常见主流编程语言中的命名规范并类比自然语言语素概念本文提出基于软件语素的标识符构成过程, 即标识符的构成可被视为软件语素的生成、排列和连接过程。在此基础上, 本文提出一种结合自然语料库的软件标识符规范性评价方法, 用来衡量软件标识符是否易于理解。最后, 本文通过源代码理解数据集和乇乩乴乨乵乢平台中开源项目对规范性指标进行了验证性实验, 结果表明本文提出的规范性分数能够很好衡量软件项目的可理解性。”
no code implementations • CCL 2021 • Chenguang Wang, Hongfei Lin, Liang Yang
“对话风格能够反映对话者的属性, 例如情感、性别和教育背景等。在对话系统中, 通过理解用户的对话风格, 能够更好地对用户进行建模。同样的, 面对不同背景的用户, 对话机器人也应该使用不同的语言风格与之交流。语言表达风格是文本的内在属性, 然而现有的大多数文本风格迁移研究, 集中在英文领域, 在中文领域则研究较少。本文构建了三个可用于中文文本风格迁移研究的数据集, 并将多种已有的文本风格迁移方法应用于该数据集。同时, 本文提出了基于DeepStyle算法与Transformer的风格迁移模型, 通过预训练可以获得不同风格的隐层向量表示。并基于Transformer构建生成端模型, 在解码阶段, 通过重建源文本的方式, 保留生成文本的内容信息, 并且引入对立风格的嵌入表示, 使得模型能够生成不同风格的文本。实验结果表明, 本文提出的模型在构建的中文数据集上均优于现有模型。”
no code implementations • 23 Apr 2024 • Haohao Zhu, Xiaokun Zhang, Junyu Lu, Youlin Wu, Zewen Bai, Changrong Min, Liang Yang, Bo Xu, Dongyu Zhang, Hongfei Lin
This limitation hinders a comprehensive understanding of individuals' personalities, as both stable traits and dynamic states are vital.
no code implementations • 2 Apr 2024 • Jesudara Omidokun, Darlington Egeonu, Bochen Jia, Liang Yang
Statistical analyses consistently support the framework's reliability, with joint angle estimations showing less than a 5-degree difference for hip flexion, elbow flexion, and knee angle methods.
no code implementations • 17 Jan 2024 • Junfu Wang, Yuanfang Guo, Liang Yang, Yunhong Wang
Firstly, we show that by applying a GC operation, the separability gains are determined by two factors, i. e., the Euclidean distance of the neighborhood distributions and $\sqrt{\mathbb{E}\left[\operatorname{deg}\right]}$, where $\mathbb{E}\left[\operatorname{deg}\right]$ is the averaged node degree.
1 code implementation • 17 Nov 2023 • Zhimin Chen, Yingwei Li, Longlong Jing, Liang Yang, Bing Li
However, a notable limitation of these approaches is that they do not fully utilize the multi-view attributes inherent in 3D point clouds, which is crucial for a deeper understanding of 3D structures.
no code implementations • 20 Oct 2023 • Ze Gao, Xiang Li, Changkun Liu, Xian Wang, Anqi Wang, Liang Yang, Yuyang Wang, Pan Hui, Tristan Braud
We present VR PreM+, an innovative VR system designed to enhance web exploration beyond traditional computer screens.
1 code implementation • 29 Sep 2023 • Xiaokun Zhang, Bo Xu, Fenglong Ma, Chenliang Li, Liang Yang, Hongfei Lin
(2) How to fuse these heterogeneous descriptive information to comprehensively infer user interests?
no code implementations • 31 May 2023 • Yan Wang, Feng Shu, Zhihong Zhuang, Rongen Dong, Qi Zhang, Di wu, Liang Yang, Jiangzhou Wang
Numerical simulation results show that a 3-bit discrete phase shifter is required to achieve a trivial performance loss for a large-scale active IRS.
1 code implementation • 8 May 2023 • Junyu Lu, Bo Xu, Xiaokun Zhang, Changrong Min, Liang Yang, Hongfei Lin
In addition, it is crucial to introduce lexical knowledge to detect the toxicity of posts, which has been a challenge for researchers.
1 code implementation • 7 May 2023 • Yuhan Chen, Yihong Luo, Jing Tang, Liang Yang, Siya Qiu, Chuan Wang, Xiaochun Cao
Motivated by it, we propose to use the local similarity (LocalSim) to learn node-level weighted fusion, which can also serve as a plug-and-play module.
no code implementations • ICCV 2023 • Ziyue Feng, Liang Yang, Pengsheng Guo, Bing Li
Recent advances in neural reconstruction using posed image sequences have made remarkable progress.
1 code implementation • ICCV 2023 • Noah Stier, Anurag Ranjan, Alex Colburn, Yajie Yan, Liang Yang, Fangchang Ma, Baptiste Angles
Recent works on 3D reconstruction from posed images have demonstrated that direct inference of scene-level 3D geometry without test-time optimization is feasible using deep neural networks, showing remarkable promise and high efficiency.
1 code implementation • ICCV 2023 • Noah Stier, Baptiste Angles, Liang Yang, Yajie Yan, Alex Colburn, Ming Chuang
Dense 3D reconstruction from RGB images traditionally assumes static camera pose estimates.
no code implementations • 7 Feb 2023 • Junfu Wang, Yuanfang Guo, Liang Yang, Yunhong Wang
In this paper, we firstly propose a heterophily-aware attention scheme and reveal the benefits of modeling the edge heterophily, i. e., if a GNN assigns different weights to edges according to different heterophilic types, it can learn effective local attention patterns, which enable nodes to acquire appropriate information from distinct neighbors.
1 code implementation • 24 Oct 2022 • Junfu Wang, Yuanfang Guo, Liang Yang, Yunhong Wang
The current success of Graph Neural Networks (GNNs) usually relies on loading the entire attributed graph for processing, which may not be satisfied with limited memory resources, especially when the attributed graph is large.
1 code implementation • 18 Oct 2022 • Zhimin Chen, Longlong Jing, Liang Yang, Yingwei Li, Bing Li
Firstly, a dynamic thresholding strategy is proposed to utilize more unlabeled data, especially for low learning status classes.
no code implementations • 23 Sep 2022 • Junfu Wang, Yuanfang Guo, Liang Yang, Yunhong Wang
Extensive experiments demonstrate that our RE-GNN can effectively and efficiently handle the heterogeneous graphs and can be applied to various homogeneous GNNs.
1 code implementation • 9 May 2022 • Xiaokun Zhang, Bo Xu, Liang Yang, Chenliang Li, Fenglong Ma, Haifeng Liu, Hongfei Lin
Finally, we predict user actions based on item features and users' price and interest preferences.
1 code implementation • 29 Mar 2022 • Ziyue Feng, Liang Yang, Longlong Jing, HaiYan Wang, YingLi Tian, Bing Li
Conventional self-supervised monocular depth prediction methods are based on a static environment assumption, which leads to accuracy degradation in dynamic scenes due to the mismatch and occlusion problems introduced by object motions.
no code implementations • NeurIPS 2021 • Liang Yang, Mengzhe Li, Liyang Liu, bingxin niu, Chuan Wang, Xiaochun Cao, Yuanfang Guo
Based on this attribute homophily rate, we propose a Diverse Message Passing (DMP) framework, which specifies every attribute propagation weight on each edge.
no code implementations • 25 Oct 2021 • Jinglun Feng, Hua Xiao, Ejup Hoxha, Yifeng Song, Liang Yang, Jizhong Xiao
Impact sounding signal has been shown to contain information about structural integrity flaws and subsurface objects from previous research.
no code implementations • ACL 2021 • Dongyu Zhang, Minghao Zhang, Heting Zhang, Liang Yang, Hongfei Lin
Metaphor involves not only a linguistic phenomenon, but also a cognitive phenomenon structuring human thought, which makes understanding it challenging.
1 code implementation • ACL 2021 • Xianbing Zhou, Yang Yong, Xiaochao Fan, Ge Ren, Yunfeng Song, Yufeng Diao, Liang Yang, Hongfei Lin
The wanton spread of hate speech on the internet brings great harm to society and families.
no code implementations • 12 Jul 2021 • Pengsheng Guo, Miguel Angel Bautista, Alex Colburn, Liang Yang, Daniel Ulbricht, Joshua M. Susskind, Qi Shan
We study the problem of novel view synthesis from sparse source observations of a scene comprised of 3D objects.
no code implementations • 3 Jun 2021 • Jinglun Feng, Liang Yang, Ejup Hoxha, Jiang Biao, Jizhong Xiao
Ground Penetrating Radar (GPR) is a very useful non-destructive evaluation (NDE) device for locating and mapping underground assets prior to digging and trenching efforts in construction.
no code implementations • 27 May 2021 • Arthur S. de Sena, Pedro H. J. Nardelli, Daniel B. da Costa, F. Rafael M. Lima, Liang Yang, Petar Popovski, Zhiguo Ding, Constantinos B. Papadias
In this paper, the appealing features of a dual-polarized intelligent reflecting surface (IRS) are exploited to improve the performance of dual-polarized massive multiple-input multiple-output (MIMO) with non-orthogonal multiple access (NOMA) under imperfect successive interference cancellation (SIC).
no code implementations • 21 Apr 2021 • Jinglun Feng, Liang Yang, Jizhong Xiao
Ground Penetrating Radar (GPR) is one of the most important non-destructive evaluation (NDE) devices to detect subsurface objects (i. e., rebars, utility pipes) and reveal the underground scene.
no code implementations • 7 Dec 2020 • Arthur S. de Sena, Pedro H. J. Nardelli, Daniel B. da Costa, F. Rafael M. Lima, Liang Yang, Petar Popovski, Zhiguo Ding, Constantinos B. Papadias
By considering the downlink of a multi-cluster scenario, the IRSs assist the base station (BS) to multiplex subsets of users in the polarization domain.
no code implementations • 5 Nov 2020 • Jinglun Feng, Liang Yang, Ejup Hoxha, Diar Sanakov, Stanislav Sotnikov, Jizhong Xiao
In this paper, both the quantitative and qualitative experiment results verify our method that can generate a dense and complete point cloud model of pipe-shaped utilities based on a sparse input, i. e., GPR raw data incompleteness and various noise.
1 code implementation • CVPR 2021 • Junfu Wang, Yunhong Wang, Zhen Yang, Liang Yang, Yuanfang Guo
Graph Neural Networks (GNNs) have achieved tremendous success in graph representation learning.
no code implementations • 20 Aug 2020 • Jinglun Feng, Liang Yang, HaiYan Wang, Yifeng Song, Jizhong Xiao
This system is composed of three modules: 1) visual inertial fusion (VIF) module to generate the pose information of GPR device, 2) deep neural network module (i. e., DepthNet) which detects B-scan of GPR image, extracts hyperbola features to remove the noise in B-scan data and predicts dielectric to determine the depth of the objects, 3) 3D GPR migration module which synchronizes the pose information with GPR scan data processed by DepthNet to reconstruct and visualize the 3D underground targets.
no code implementations • 31 Jul 2020 • Liang Yang, Yongjie Yuan
In this paper, the physical layer security (PLS) for a novel reconfigurable intelligent surface (RIS)-assisted non-orthogonal multiple access (NOMA) system in a multi-user scenario is investigated, where we consider the worst case that the eavesdropper also utilizes the advantage of the RISs.
no code implementations • WS 2020 • Shuqun Li, Jingjie Zeng, Jinhui Zhang, Tao Peng, Liang Yang, Hongfei Lin
In our daily life, metaphor is a common way of expression.
no code implementations • 26 Apr 2020 • Hai-Yan Wang, Xuejian Rong, Liang Yang, Jinglun Feng, Jizhong Xiao, YingLi Tian
The deficiency of 3D segmentation labels is one of the main obstacles to effective point cloud segmentation, especially for scenes in the wild with varieties of different objects.
no code implementations • NeurIPS 2019 • Shen-Huan Lyu, Liang Yang, Zhi-Hua Zhou
In this paper, we formulate the forest representation learning approach called \textsc{CasDF} as an additive model which boosts the augmented feature instead of the prediction.
no code implementations • 15 Nov 2019 • Liang Yang, Xi-Zhu Wu, Yuan Jiang, Zhi-Hua Zhou
In multi-label learning, each instance is associated with multiple labels and the crucial task is how to leverage label correlations in building models.
no code implementations • 7 May 2019 • Shen-Huan Lv, Liang Yang, Zhi-Hua Zhou
In this paper, we reformulate the forest representation learning approach as an additive model which boosts the augmented feature instead of the prediction.
no code implementations • 2 Dec 2018 • Liang Yang, Hao Jiang, Jizhong Xiao, Zhouyuan Huo
To provide a possible solution to this problem, this paper proposes a camera system with both ego-downward and third-static view to perform localization and tracking in a learning approach.
no code implementations • EMNLP 2018 • Yufeng Diao, Hongfei Lin, Di wu, Liang Yang, Kan Xu, Zhihao Yang, Jian Wang, Shaowu Zhang, Bo Xu, Dongyu Zhang
In this work, we first use WordNet to understand and expand word embedding for settling the polysemy of homographic puns, and then propose a WordNet-Encoded Collocation-Attention network model (WECA) which combined with the context weights for recognizing the puns.
no code implementations • ACL 2018 • Dongyu Zhang, Hongfei Lin, Liang Yang, Shaowu Zhang, Bo Xu
However, there is little research on the construction of metaphor corpora annotated with emotion for the analysis of emotionality of metaphorical expressions.