Author page zhiyu chen lehigh (closes #5158) (#5800)

Split conflated author page by adding explicit ID 'zhiyu-chen-lehigh'
for Zhiyu Chen from Lehigh University with ORCID 0000-0002-3096-7912.

Add --first-name flag to script.

Author: mjpost

bin/add_author_id.py

@@ -24,23 +24,21 @@
 
 Usage:
 
-    ./add_author_id.py bill-byrne --last-name Byrne
+    ./add_author_id.py bill-byrne --last-name Byrne --first-name Bill
 """
 
 import argparse
 import os
 
+from pathlib import Path
 from anthology.utils import indent
 from itertools import chain
 
 import lxml.etree as ET
 
 
 def main(args):
-    for xml_file in os.listdir(args.data_dir):
-        if not xml_file.endswith(".xml"):
-            continue
-
+    for xml_file in Path(args.data_dir).glob("**/*.xml"):
         changed_one = False
 
         tree = ET.parse(xml_file)
@@ -53,7 +51,11 @@ def main(args):
                 if "id" in author_xml.attrib:
                     continue
                 last_name = author_xml.find("./last").text
-                if last_name == args.last_name:
+                try:
+                    first_name = author_xml.find("./first").text
+                except AttributeError:
+                    first_name = ""
+                if last_name == args.last_name and first_name == args.first_name:
                     paper_id = (
                         paper_xml.attrib["id"] if paper_xml.text == "paper" else "0"
                     )
@@ -71,6 +73,7 @@ def main(args):
     parser = argparse.ArgumentParser()
     parser.add_argument("id", help="Author ID to add")
     parser.add_argument("--last-name", help="Author's last name")
+    parser.add_argument("--first-name", help="Author's first name")
     parser.add_argument("--confirm", action="store_true", help="Confirm each instance")
     parser.add_argument(
         "--data-dir", default=os.path.join(os.path.dirname(__file__), "..", "data", "xml")

data/xml/2020.acl.xml

@@ -250,7 +250,7 @@
     </paper>
     <paper id="18">
       <title>Few-Shot <fixed-case>NLG</fixed-case> with Pre-Trained Language Model</title>
-      <author><first>Zhiyu</first><last>Chen</last></author>
+      <author id="zhiyu-chen"><first>Zhiyu</first><last>Chen</last></author>
       <author><first>Harini</first><last>Eavani</last></author>
       <author><first>Wenhu</first><last>Chen</last></author>
       <author><first>Yinyin</first><last>Liu</last></author>
@@ -9517,7 +9517,7 @@
       <author><first>Wenhu</first><last>Chen</last></author>
       <author><first>Jianshu</first><last>Chen</last></author>
       <author><first>Yu</first><last>Su</last></author>
-      <author><first>Zhiyu</first><last>Chen</last></author>
+      <author id="zhiyu-chen"><first>Zhiyu</first><last>Chen</last></author>
       <author><first>William Yang</first><last>Wang</last></author>
       <pages>7929–7942</pages>
       <abstract>Neural natural language generation (NLG) models have recently shown remarkable progress in fluency and coherence. However, existing studies on neural NLG are primarily focused on surface-level realizations with limited emphasis on logical inference, an important aspect of human thinking and language. In this paper, we suggest a new NLG task where a model is tasked with generating natural language statements that can be <i>logically entailed</i> by the facts in an open-domain semi-structured table. To facilitate the study of the proposed logical NLG problem, we use the existing TabFact dataset~(CITATION) featured with a wide range of logical/symbolic inferences as our testbed, and propose new automatic metrics to evaluate the fidelity of generation models w.r.t. logical inference. The new task poses challenges to the existing monotonic generation frameworks due to the mismatch between sequence order and logical order. In our experiments, we comprehensively survey different generation architectures (LSTM, Transformer, Pre-Trained LM) trained with different algorithms (RL, Adversarial Training, Coarse-to-Fine) on the dataset and made following observations: 1) Pre-Trained LM can significantly boost both the fluency and logical fidelity metrics, 2) RL and Adversarial Training are trading fluency for fidelity, 3) Coarse-to-Fine generation can help partially alleviate the fidelity issue while maintaining high language fluency. The code and data are available at <url>https://github.com/wenhuchen/LogicNLG</url>.</abstract>

data/xml/2020.findings.xml

@@ -1173,7 +1173,7 @@
       <title><fixed-case>H</fixed-case>ybrid<fixed-case>QA</fixed-case>: A Dataset of Multi-Hop Question Answering over Tabular and Textual Data</title>
       <author><first>Wenhu</first><last>Chen</last></author>
       <author><first>Hanwen</first><last>Zha</last></author>
-      <author><first>Zhiyu</first><last>Chen</last></author>
+      <author id="zhiyu-chen"><first>Zhiyu</first><last>Chen</last></author>
       <author><first>Wenhan</first><last>Xiong</last></author>
       <author><first>Hong</first><last>Wang</last></author>
       <author><first>William Yang</first><last>Wang</last></author>
@@ -2461,7 +2461,7 @@
     </paper>
     <paper id="190">
       <title><fixed-case>L</fixed-case>ogic2<fixed-case>T</fixed-case>ext: High-Fidelity Natural Language Generation from Logical Forms</title>
-      <author><first>Zhiyu</first><last>Chen</last></author>
+      <author id="zhiyu-chen"><first>Zhiyu</first><last>Chen</last></author>
       <author><first>Wenhu</first><last>Chen</last></author>
       <author><first>Hanwen</first><last>Zha</last></author>
       <author><first>Xiyou</first><last>Zhou</last></author>

data/xml/2021.eacl.xml

@@ -1563,8 +1563,8 @@
       <author><first>Anette</first><last>Frank</last></author>
       <pages>1504–1518</pages>
       <abstract>Systems that generate natural language text from abstract meaning representations such as AMR are typically evaluated using automatic surface matching metrics that compare the generated texts to reference texts from which the input meaning representations were constructed. We show that besides well-known issues from which such metrics suffer, an additional problem arises when applying these metrics for AMR-to-text evaluation, since an abstract meaning representation allows for numerous surface realizations. In this work we aim to alleviate these issues by proposing <tex-math>\mathcal{M}\mathcal{F}_\beta</tex-math>, a decomposable metric that builds on two pillars. The first is the <b>principle of meaning preservation <tex-math>\mathcal{M}</tex-math>
-        </b>: it measures to what extent a given AMR can be reconstructed from the generated sentence using SOTA AMR parsers and applying (fine-grained) AMR evaluation metrics to measure the distance between the original and the reconstructed AMR. The second pillar builds on a <b>principle of (grammatical) form <tex-math>\mathcal{F}</tex-math>
-        </b> that measures the linguistic quality of the generated text, which we implement using SOTA language models. In two extensive pilot studies we show that fulfillment of both principles offers benefits for AMR-to-text evaluation, including explainability of scores. Since <tex-math>\mathcal{M}\mathcal{F}_\beta</tex-math> does not necessarily rely on gold AMRs, it may extend to other text generation tasks.</abstract>
+ </b>: it measures to what extent a given AMR can be reconstructed from the generated sentence using SOTA AMR parsers and applying (fine-grained) AMR evaluation metrics to measure the distance between the original and the reconstructed AMR. The second pillar builds on a <b>principle of (grammatical) form <tex-math>\mathcal{F}</tex-math>
+ </b> that measures the linguistic quality of the generated text, which we implement using SOTA language models. In two extensive pilot studies we show that fulfillment of both principles offers benefits for AMR-to-text evaluation, including explainability of scores. Since <tex-math>\mathcal{M}\mathcal{F}_\beta</tex-math> does not necessarily rely on gold AMRs, it may extend to other text generation tasks.</abstract>
       <url hash="7c470968">2021.eacl-main.129</url>
       <bibkey>opitz-frank-2021-towards</bibkey>
       <doi>10.18653/v1/2021.eacl-main.129</doi>
@@ -4448,7 +4448,7 @@
     <paper id="39">
       <title><fixed-case>HULK</fixed-case>: An Energy Efficiency Benchmark Platform for Responsible Natural Language Processing</title>
       <author><first>Xiyou</first><last>Zhou</last></author>
-      <author><first>Zhiyu</first><last>Chen</last></author>
+      <author id="zhiyu-chen"><first>Zhiyu</first><last>Chen</last></author>
       <author><first>Xiaoyong</first><last>Jin</last></author>
       <author><first>William Yang</first><last>Wang</last></author>
       <pages>329–336</pages>

data/xml/2021.emnlp.xml

@@ -4176,7 +4176,7 @@
     </paper>
     <paper id="300">
       <title><fixed-case>F</fixed-case>in<fixed-case>QA</fixed-case>: A Dataset of Numerical Reasoning over Financial Data</title>
-      <author><first>Zhiyu</first><last>Chen</last></author>
+      <author id="zhiyu-chen"><first>Zhiyu</first><last>Chen</last></author>
       <author><first>Wenhu</first><last>Chen</last></author>
       <author><first>Charese</first><last>Smiley</last></author>
       <author><first>Sameena</first><last>Shah</last></author>

data/xml/2021.findings.xml

@@ -10153,7 +10153,7 @@
     </paper>
     <paper id="337">
       <title><fixed-case>NUANCED</fixed-case>: Natural Utterance Annotation for Nuanced Conversation with Estimated Distributions</title>
-      <author><first>Zhiyu</first><last>Chen</last></author>
+      <author id="zhiyu-chen"><first>Zhiyu</first><last>Chen</last></author>
       <author><first>Honglei</first><last>Liu</last></author>
       <author><first>Hu</first><last>Xu</last></author>
       <author><first>Seungwhan</first><last>Moon</last></author>

data/xml/2022.emnlp.xml

@@ -5828,7 +5828,7 @@
     </paper>
     <paper id="421">
       <title><fixed-case>C</fixed-case>onv<fixed-case>F</fixed-case>in<fixed-case>QA</fixed-case>: Exploring the Chain of Numerical Reasoning in Conversational Finance Question Answering</title>
-      <author><first>Zhiyu</first><last>Chen</last><affiliation>Meta</affiliation></author>
+      <author id="zhiyu-chen"><first>Zhiyu</first><last>Chen</last><affiliation>Meta</affiliation></author>
       <author><first>Shiyang</first><last>Li</last><affiliation>UC Santa Barbara</affiliation></author>
       <author><first>Charese</first><last>Smiley</last><affiliation>JPMorgan AI Research</affiliation></author>
       <author><first>Zhiqiang</first><last>Ma</last><affiliation>JPMorgan Chase</affiliation></author>
@@ -12665,7 +12665,7 @@
     </paper>
     <paper id="36">
       <title>Reinforced Question Rewriting for Conversational Question Answering</title>
-      <author><first>Zhiyu</first><last>Chen</last><affiliation>Amazon</affiliation></author>
+      <author id="zhiyu-chen-lehigh"><first>Zhiyu</first><last>Chen</last><affiliation>Amazon</affiliation></author>
       <author><first>Jie</first><last>Zhao</last><affiliation>Amazon</affiliation></author>
       <author><first>Anjie</first><last>Fang</last><affiliation>Amazon</affiliation></author>
       <author><first>Besnik</first><last>Fetahu</last><affiliation>Amazon</affiliation></author>

data/xml/2022.findings.xml

@@ -7260,7 +7260,7 @@
     </paper>
     <paper id="197">
       <title><fixed-case>KETOD</fixed-case>: Knowledge-Enriched Task-Oriented Dialogue</title>
-      <author><first>Zhiyu</first><last>Chen</last></author>
+      <author id="zhiyu-chen"><first>Zhiyu</first><last>Chen</last></author>
       <author><first>Bing</first><last>Liu</last></author>
       <author><first>Seungwhan</first><last>Moon</last></author>
       <author><first>Chinnadhurai</first><last>Sankar</last></author>

data/xml/2022.suki.xml

@@ -5,7 +5,7 @@
       <booktitle>Proceedings of the Workshop on Structured and Unstructured Knowledge Integration (SUKI)</booktitle>
       <editor><first>Wenhu</first><last>Chen</last></editor>
       <editor><first>Xinyun</first><last>Chen</last></editor>
-      <editor><first>Zhiyu</first><last>Chen</last></editor>
+      <editor id="zhiyu-chen"><first>Zhiyu</first><last>Chen</last></editor>
       <editor><first>Ziyu</first><last>Yao</last></editor>
       <editor><first>Michihiro</first><last>Yasunaga</last></editor>
       <editor><first>Tao</first><last>Yu</last></editor>

data/xml/2023.acl.xml

@@ -17644,7 +17644,7 @@
     <paper id="70">
       <title>Answering Unanswered Questions through Semantic Reformulations in Spoken <fixed-case>QA</fixed-case></title>
       <author><first>Pedro</first><last>Faustini</last><affiliation>Macquarie University</affiliation></author>
-      <author><first>Zhiyu</first><last>Chen</last><affiliation>Amazon</affiliation></author>
+      <author id="zhiyu-chen-lehigh"><first>Zhiyu</first><last>Chen</last><affiliation>Amazon</affiliation></author>
       <author><first>Besnik</first><last>Fetahu</last><affiliation>Amazon</affiliation></author>
       <author><first>Oleg</first><last>Rokhlenko</last><affiliation>Amazon Research</affiliation></author>
       <author><first>Shervin</first><last>Malmasi</last><affiliation>Amazon</affiliation></author>
@@ -17684,7 +17684,7 @@
     </paper>
     <paper id="73">
       <title>Generate-then-Retrieve: Intent-Aware <fixed-case>FAQ</fixed-case> Retrieval in Product Search</title>
-      <author><first>Zhiyu</first><last>Chen</last><affiliation>Amazon</affiliation></author>
+      <author id="zhiyu-chen-lehigh"><first>Zhiyu</first><last>Chen</last><affiliation>Amazon</affiliation></author>
       <author><first>Jason</first><last>Choi</last><affiliation>Amazon</affiliation></author>
       <author><first>Besnik</first><last>Fetahu</last><affiliation>Amazon</affiliation></author>
       <author><first>Oleg</first><last>Rokhlenko</last><affiliation>Amazon Research</affiliation></author>