Added author page lei-wang-xjipc

data/xml/2020.acl.xml

@@ -4878,7 +4878,7 @@
     <paper id="362">
       <title>Graph-to-Tree Learning for Solving Math Word Problems</title>
       <author><first>Jipeng</first><last>Zhang</last></author>
-      <author><first>Lei</first><last>Wang</last></author>
+      <author id="lei-wang"><first>Lei</first><last>Wang</last></author>
       <author><first>Roy Ka-Wei</first><last>Lee</last></author>
       <author><first>Yi</first><last>Bin</last></author>
       <author><first>Yan</first><last>Wang</last></author>

data/xml/2020.ccl.xml

@@ -821,7 +821,7 @@
     <paper id="69">
       <title>基于预训练语言模型的案件要素识别方法(A Method for Case Factor Recognition Based on Pre-trained Language Models)</title>
       <author><first>Haishun</first><last>Liu</last><variant script="hani"><first>海顺</first><last>刘</last></variant></author>
-      <author><first>Lei</first><last>Wang</last><variant script="hani"><first>雷</first><last>王</last></variant></author>
+      <author id="lei-wang"><first>Lei</first><last>Wang</last><variant script="hani"><first>雷</first><last>王</last></variant></author>
       <author><first>Yanguang</first><last>Chen</last><variant script="hani"><first>彦光</first><last>陈</last></variant></author>
       <author><first>Shuchen</first><last>Zhang</last><variant script="hani"><first>书晨</first><last>张</last></variant></author>
       <author><first>Yuanyuan</first><last>Sun</last><variant script="hani"><first>媛媛</first><last>孙</last></variant></author>

data/xml/2021.findings.xml

@@ -10337,7 +10337,7 @@
     <paper id="350">
       <title><fixed-case>NOAHQA</fixed-case>: Numerical Reasoning with Interpretable Graph Question Answering Dataset</title>
       <author><first>Qiyuan</first><last>Zhang</last></author>
-      <author><first>Lei</first><last>Wang</last></author>
+      <author id="lei-wang"><first>Lei</first><last>Wang</last></author>
       <author><first>Sicheng</first><last>Yu</last></author>
       <author><first>Shuohang</first><last>Wang</last></author>
       <author><first>Yang</first><last>Wang</last></author>

data/xml/2022.coling.xml

@@ -5051,7 +5051,7 @@
     <paper id="436">
       <title><fixed-case>R</fixed-case>otate<fixed-case>CT</fixed-case>: Knowledge Graph Embedding by Rotation and Coordinate Transformation in Complex Space</title>
       <author><first>Yao</first><last>Dong</last></author>
-      <author><first>Lei</first><last>Wang</last></author>
+      <author id="lei-wang"><first>Lei</first><last>Wang</last></author>
       <author><first>Ji</first><last>Xiang</last></author>
       <author><first>Xiaobo</first><last>Guo</last></author>
       <author><first>Yuqiang</first><last>Xie</last></author>

data/xml/2022.emnlp.xml

@@ -1597,7 +1597,7 @@
       <author><first>Zheng</first><last>Lin</last><affiliation>Institute of Information Engineering, Chinese Academy of Sciences</affiliation></author>
       <author><first>Yuanxin</first><last>Liu</last><affiliation>Institute of Information Engineering, Chinese Academy of Sciences</affiliation></author>
       <author><first>Zhengxiao</first><last>Liu</last><affiliation>Institute of Information Engineering, Chinese Academy of Sciences</affiliation></author>
-      <author><first>Lei</first><last>Wang</last><affiliation>Institute of Information Engineering, Chinese Academy of Sciences</affiliation></author>
+      <author id="lei-wang"><first>Lei</first><last>Wang</last><affiliation>Institute of Information Engineering, Chinese Academy of Sciences</affiliation></author>
       <author><first>Weiping</first><last>Wang</last><affiliation>Institute of Information Engineering, Chinese Academy of Sciences</affiliation></author>
       <pages>1719-1730</pages>
       <abstract>Transformer-based pre-trained language models (PLMs) mostly suffer from excessive overhead despite their advanced capacity. For resource-constrained devices, there is an urgent need for a spatially and temporally efficient model which retains the major capacity of PLMs. However, existing statically compressed models are unaware of the diverse complexities between input instances, potentially resulting in redundancy and inadequacy for simple and complex inputs. Also, miniature models with early exiting encounter challenges in the trade-off between making predictions and serving the deeper layers. Motivated by such considerations, we propose a collaborative optimization for PLMs that integrates static model compression and dynamic inference acceleration. Specifically, the PLM is slenderized in width while the depth remains intact, complementing layer-wise early exiting to speed up inference dynamically. To address the trade-off of early exiting, we propose a joint training approach that calibrates slenderization and preserves contributive structures to each exit instead of only the final layer. Experiments are conducted on GLUE benchmark and the results verify the Pareto optimality of our approach at high compression and acceleration rate with 1/8 parameters and 1/19 FLOPs of BERT.</abstract>
@@ -4379,7 +4379,7 @@
       <author><first>Jiayu</first><last>Xu</last><affiliation>Beijing Baidu Netcom Science and Technology Co., Ltd.</affiliation></author>
       <author><first>Chao</first><last>Lu</last><affiliation>Baidu Inc</affiliation></author>
       <author><first>Yehui</first><last>Yang</last><affiliation>Baidu.Inc.</affiliation></author>
-      <author><first>Lei</first><last>Wang</last><affiliation>Baidu</affiliation></author>
+      <author id="lei-wang"><first>Lei</first><last>Wang</last><affiliation>Baidu</affiliation></author>
       <author><first>Haifeng</first><last>Huang</last><affiliation>Baidu Inc</affiliation></author>
       <author><first>Xia</first><last>Zhang</last><affiliation>Neusoft Corporation</affiliation></author>
       <author><first>Junwei</first><last>Liu</last><affiliation>baidu</affiliation></author>

data/xml/2022.findings.xml

@@ -2627,7 +2627,7 @@
       <author><first>Yating</first><last>Yang</last></author>
       <author><first>Zhou</first><last>Xi</last></author>
       <author><first>Bo</first><last>Ma</last></author>
-      <author><first>Lei</first><last>Wang</last></author>
+      <author id="lei-wang-xjipc"><first>Lei</first><last>Wang</last></author>
       <author><first>Rui</first><last>Dong</last></author>
       <author><first>Azmat</first><last>Anwar</last></author>
       <pages>2455-2469</pages>
@@ -5520,7 +5520,7 @@
       <title><fixed-case>MWP</fixed-case>-<fixed-case>BERT</fixed-case>: Numeracy-Augmented Pre-training for Math Word Problem Solving</title>
       <author><first>Zhenwen</first><last>Liang</last></author>
       <author><first>Jipeng</first><last>Zhang</last></author>
-      <author><first>Lei</first><last>Wang</last></author>
+      <author id="lei-wang"><first>Lei</first><last>Wang</last></author>
       <author><first>Wei</first><last>Qin</last></author>
       <author orcid="0000-0002-0192-8498"><first>Yunshi</first><last>Lan</last></author>
       <author><first>Jie</first><last>Shao</last></author>

data/xml/2022.mathnlp.xml

@@ -35,7 +35,7 @@
     <paper id="2">
       <title>Investigating Math Word Problems using Pretrained Multilingual Language Models</title>
       <author><first>Minghuan</first><last>Tan</last></author>
-      <author><first>Lei</first><last>Wang</last></author>
+      <author id="lei-wang"><first>Lei</first><last>Wang</last></author>
       <author><first>Lingxiao</first><last>Jiang</last></author>
       <author><first>Jing</first><last>Jiang</last></author>
       <pages>7-16</pages>

data/xml/2023.acl.xml

@@ -2070,7 +2070,7 @@
     </paper>
     <paper id="147">
       <title>Plan-and-Solve Prompting: Improving Zero-Shot Chain-of-Thought Reasoning by Large Language Models</title>
-      <author orcid="0000-0003-1228-6758"><first>Lei</first><last>Wang</last><affiliation>Singapore Management University</affiliation></author>
+      <author orcid="0000-0003-1228-6758" id="lei-wang"><first>Lei</first><last>Wang</last><affiliation>Singapore Management University</affiliation></author>
       <author><first>Wanyu</first><last>Xu</last><affiliation>Southwest Jiaotong University</affiliation></author>
       <author><first>Yihuai</first><last>Lan</last><affiliation>Xihua University</affiliation></author>
       <author><first>Zhiqiang</first><last>Hu</last><affiliation>Singapore University of Technology and Design</affiliation></author>

data/xml/2023.emnlp.xml

@@ -2801,7 +2801,7 @@
       <author><first>Yi</first><last>Bin</last></author>
       <author><first>Mengqun</first><last>Han</last></author>
       <author><first>Wenhao</first><last>Shi</last></author>
-      <author><first>Lei</first><last>Wang</last></author>
+      <author id="lei-wang"><first>Lei</first><last>Wang</last></author>
       <author><first>Yang</first><last>Yang</last></author>
       <author><first>See-Kiong</first><last>Ng</last></author>
       <author><first>Heng</first><last>Shen</last></author>
@@ -3702,7 +3702,7 @@
       <author><first>Yuanyuan</first><last>Liang</last></author>
       <author><first>Jianing</first><last>Wang</last></author>
       <author><first>Hanlun</first><last>Zhu</last></author>
-      <author><first>Lei</first><last>Wang</last></author>
+      <author id="lei-wang"><first>Lei</first><last>Wang</last></author>
       <author><first>Weining</first><last>Qian</last></author>
       <author><first>Yunshi</first><last>Lan</last></author>
       <pages>4329-4343</pages>
@@ -4485,7 +4485,7 @@
     <paper id="319">
       <title><fixed-case>LLM</fixed-case>-Adapters: An Adapter Family for Parameter-Efficient Fine-Tuning of Large Language Models</title>
       <author><first>Zhiqiang</first><last>Hu</last></author>
-      <author><first>Lei</first><last>Wang</last></author>
+      <author id="lei-wang"><first>Lei</first><last>Wang</last></author>
       <author><first>Yihuai</first><last>Lan</last></author>
       <author><first>Wanyu</first><last>Xu</last></author>
       <author><first>Ee-Peng</first><last>Lim</last></author>
@@ -16515,7 +16515,7 @@ The experiments were repeated and the tables and figures were updated. Changes a
     </paper>
     <paper id="64">
       <title><fixed-case>LLM</fixed-case>4<fixed-case>V</fixed-case>is: Explainable Visualization Recommendation using <fixed-case>C</fixed-case>hat<fixed-case>GPT</fixed-case></title>
-      <author><first>Lei</first><last>Wang</last></author>
+      <author id="lei-wang"><first>Lei</first><last>Wang</last></author>
       <author><first>Songheng</first><last>Zhang</last></author>
       <author><first>Yun</first><last>Wang</last></author>
       <author><first>Ee-Peng</first><last>Lim</last></author>

data/xml/2023.findings.xml

@@ -16126,7 +16126,7 @@
       <title>R<tex-math>^3</tex-math> Prompting: Review, Rephrase and Resolve for Chain-of-Thought Reasoning in Large Language Models under Noisy Context</title>
       <author><first>Qingyuan</first><last>Tian</last></author>
       <author><first>Hanlun</first><last>Zhu</last></author>
-      <author><first>Lei</first><last>Wang</last></author>
+      <author id="lei-wang"><first>Lei</first><last>Wang</last></author>
       <author><first>Yang</first><last>Li</last></author>
       <author><first>Yunshi</first><last>Lan</last></author>
       <pages>1670-1685</pages>
@@ -25197,7 +25197,7 @@
       <author><first>Baokui</first><last>Li</last></author>
       <author><first>Kun</first><last>Kuang</last></author>
       <author><first>Yating</first><last>Zhang</last></author>
-      <author><first>Lei</first><last>Wang</last></author>
+      <author id="lei-wang"><first>Lei</first><last>Wang</last></author>
       <author><first>Anh</first><last>Luu</last></author>
       <author><first>Yi</first><last>Yang</last></author>
       <author><first>Fei</first><last>Wu</last></author>

data/xml/2024.emnlp.xml

@@ -109,7 +109,7 @@
       <title><fixed-case>LLM</fixed-case>-Based Agent Society Investigation: Collaboration and Confrontation in Avalon Gameplay</title>
       <author><first>Yihuai</first><last>Lan</last></author>
       <author><first>Zhiqiang</first><last>Hu</last><affiliation>Singapore University of Technology and Design</affiliation></author>
-      <author><first>Lei</first><last>Wang</last><affiliation>SalesForce</affiliation></author>
+      <author id="lei-wang"><first>Lei</first><last>Wang</last><affiliation>SalesForce</affiliation></author>
       <author><first>Yang</first><last>Wang</last></author>
       <author orcid="0000-0002-1754-1837"><first>Deheng</first><last>Ye</last><affiliation>Tencent and Tencent</affiliation></author>
       <author><first>Peilin</first><last>Zhao</last><affiliation>Tencent AI Lab</affiliation></author>

data/xml/2024.findings.xml

@@ -2742,7 +2742,7 @@
     </paper>
     <paper id="56">
       <title>The Whole is Better than the Sum: Using Aggregated Demonstrations in In-Context Learning for Sequential Recommendation</title>
-      <author><first>Lei</first><last>Wang</last><affiliation>SalesForce</affiliation></author>
+      <author id="lei-wang"><first>Lei</first><last>Wang</last><affiliation>SalesForce</affiliation></author>
       <author><first>Ee-Peng</first><last>Lim</last><affiliation>Singapore Management University</affiliation></author>
       <pages>876-895</pages>
       <abstract>Large language models (LLMs) have shown excellent performance on various NLP tasks. To use LLMs as strong sequential recommenders, we explore the in-context learning approach to sequential recommendation. We investigate the effects of instruction format, task consistency, demonstration selection, and number of demonstrations. As increasing the number of demonstrations in ICL does not improve accuracy despite using a long prompt, we propose a novel method called LLMSRec-Syn that incorporates multiple demonstration users into one aggregated demonstration. Our experiments on three recommendation datasets show that LLMSRec-Syn outperforms state-of-the-art LLM-based sequential recommendation methods. In some cases, LLMSRec-Syn can perform on par with or even better than supervised learning methods. Our code is publicly available at https://github.com/demoleiwang/LLMSRec_Syn.</abstract>
@@ -4351,7 +4351,7 @@
       <title>Compensate Quantization Errors: Make Weights Hierarchical to Compensate Each Other</title>
       <author><first>Yifei</first><last>Gao</last></author>
       <author><first>Jie</first><last>Ou</last></author>
-      <author><first>Lei</first><last>Wang</last><affiliation>Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, Chinese Academy of Sciences</affiliation></author>
+      <author id="lei-wang"><first>Lei</first><last>Wang</last><affiliation>Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, Chinese Academy of Sciences</affiliation></author>
       <author><first>Yuting</first><last>Xiao</last></author>
       <author><first>Xiangzhiyuan</first><last>Xiangzhiyuan</last></author>
       <author><first>Ruiting</first><last>Dai</last></author>
@@ -17661,7 +17661,7 @@
       <title><fixed-case>NUMC</fixed-case>o<fixed-case>T</fixed-case>: Numerals and Units of Measurement in Chain-of-Thought Reasoning using Large Language Models</title>
       <author><first>Ancheng</first><last>Xu</last></author>
       <author orcid="0000-0001-8287-0453"><first>Minghuan</first><last>Tan</last><affiliation>Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences, Chinese Academy of Sciences</affiliation></author>
-      <author><first>Lei</first><last>Wang</last><affiliation>SalesForce</affiliation></author>
+      <author id="lei-wang"><first>Lei</first><last>Wang</last><affiliation>SalesForce</affiliation></author>
       <author><first>Min</first><last>Yang</last><affiliation>Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, Chinese Academy of Sciences</affiliation></author>
       <author><first>Ruifeng</first><last>Xu</last><affiliation>Harbin Institute of Technology</affiliation></author>
       <pages>14268-14290</pages>
@@ -32389,7 +32389,7 @@ hai-coaching/</abstract>
     </paper>
     <paper id="952">
       <title>The Overlooked Repetitive Lengthening Form in Sentiment Analysis</title>
-      <author><first>Lei</first><last>Wang</last></author>
+      <author id="lei-wang"><first>Lei</first><last>Wang</last></author>
       <author><first>Eduard</first><last>Dragut</last><affiliation>Temple University</affiliation></author>
       <pages>16225-16238</pages>
       <abstract>Individuals engaging in online communication frequently express personal opinions with informal styles (e.g., memes and emojis). While Language Models (LMs) with informal communications have been widely discussed, a unique and emphatic style, the Repetitive Lengthening Form (RLF), has been overlooked for years. In this paper, we explore answers to two research questions: 1) Is RLF important for SA? 2) Can LMs understand RLF? Inspired by previous linguistic research, we curate **Lengthening**, the first multi-domain dataset with 850k samples focused on RLF for sentiment analysis. Moreover, we introduce **Explnstruct**, a two-stage Explainable Instruction Tuning framework aimed at improving both the performance and explainability of LLMs for RLF. We further propose a novel unified approach to quantify LMs’ understanding of informal expressions. We show that RLF sentences are expressive expressions and can serve as signatures of document-level sentiment. Additionally, RLF has potential value for online content analysis. Our comprehensive results show that fine-tuned Pre-trained Language Models (PLMs) can surpass zero-shot GPT-4 in performance but not in explanation for RLF. Finally, we show ExpInstruct can improve the open-sourced LLMs to match zero-shot GPT-4 in performance and explainability for RLF with limited samples. Code and sample data are available at https://github.com/Tom-Owl/OverlookedRLF</abstract>

data/xml/2024.lrec.xml

@@ -5693,7 +5693,7 @@
       <title>Dual Complex Number Knowledge Graph Embeddings</title>
       <author><first>Yao</first><last>Dong</last></author>
       <author><first>Qingchao</first><last>Kong</last></author>
-      <author><first>Lei</first><last>Wang</last></author>
+      <author id="lei-wang"><first>Lei</first><last>Wang</last></author>
       <author><first>Yin</first><last>Luo</last></author>
       <pages>5391–5400</pages>
       <abstract>Knowledge graph embedding, which aims to learn representations of entities and relations in large scale knowledge graphs, plays a crucial part in various downstream applications. The performance of knowledge graph embedding models mainly depends on the ability of modeling relation patterns, such as symmetry/antisymmetry, inversion and composition (commutative composition and non-commutative composition). Most existing methods fail in modeling the non-commutative composition patterns. Several methods support this kind of pattern by modeling in quaternion space or dihedral group. However, extending to such sophisticated spaces leads to a substantial increase in the amount of parameters, which greatly reduces the parameter efficiency. In this paper, we propose a new knowledge graph embedding method called dual complex number knowledge graph embeddings (DCNE), which maps entities to the dual complex number space, and represents relations as rotations in 2D space via dual complex number multiplication. The non-commutativity of the dual complex number multiplication empowers DCNE to model the non-commutative composition patterns. In the meantime, modeling relations as rotations in 2D space can effectively improve the parameter efficiency. Extensive experiments on multiple benchmark knowledge graphs empirically show that DCNE achieves significant performance in link prediction and path query answering.</abstract>