December Corrections (#2921)

* Paper meta data correction for 2023.iwcs-1.29, closes #2867.

* Paper pdf correction for 2023.inlg-main.21, closes #2871.

* Paper meta data correction for 2023.ldk-1.46, closes #2877.

* Paper meta data correction for P17-1159, closes #2881.

* Paper meta data and pdf correction for 2023.insights-1.1, closes #2883.

* Paper meta data correction for 2023.semeval-1.179, closes #2885.

* Paper meta data correction for 2023.sigdial-1.27, closes #2895.

* Paper meta data correction for 2023.conll-1.25, closes #2897.

* Paper meta data correction for 2022.emnlp-main.744, closes #2899.

* Paper meta data correction for 2023.emnlp-demo.13, closes #2900.

* Paper pdf correction for 2023.nllp-1.10, closes #2901.

* Paper pdf correction for 2020.emnlp-main.254, closes #2902.

* Paper meta data correction for 2022.emnlp-main.788, closes #2903.

* Paper meta data correction for 2023.eamt-1.55, closes #2909.

* Paper mata data correction for 2023.banglalp-1.5, closes #2910.

* Paper meta data correction for 2023.banglalp-1.27, closes #2911.

* Paper meta data correction for 2023.banglalp-1.45, closes #2912.

* Paper meta data correction for 2023.emnlp-main.920, closes #2915.

* Paper meta data correction for 2023.findings-emnlp.413, closes #2917.

* Paper meta data correction for 2023.arabicnlp-1.28, closes #2918.

* Paper meta data correction for 2023.findings-emnlp.841, closes #2924.

* Paper meta data correction for  2023.emnlp-main.388, closes #2925.

* Paper meta data correction for 2023.findings-emnlp.805, closes #2926.

* Paper pdf correction for 2023.emnlp-main.167, closes #2932.

* Paper pdf correction for 2023.findings-emnlp.459, closes #2933.

* Meta data correction for 2023.emnlp-main.128, closes #2928.

* Paper meta data correction forr 2022.acl-long.480, closes #2929.

* Paper pdf correction for 2023.emnlp-main.167, closes #2932.

* 2023.emnlp-main.167 fix.

* Paper pdf correction for 2023.jeptalnrecital-short.3, closes #2934.

* Paper title meta data correction for 2023.findings-emnlp.231, closes #2936.

* Paper meta data correction for 2023.emnlp-main.900, closes #2938.

* Author name meta data correction for Author Xuan Long Do, closes #2939.

* Paper pdf correction for 2023.emnlp-main.771, closes #2940.

* Volume meta data update for NLP-OSS 2023, closes #2941.

* Paper pdf correction for 2023.emnlp-main.429, closes #2942.

* Paper pdf correction for 2023.findings-emnlp.949, closes #2949.

* Paper pdf correction for 2023.emnlp-main.167, closes #2952.

* Paper pdf correction for 2023.emnlp-main.633, closes #2953.

* Paper pdf correction for 2023.emnlp-main.814, closes #2957.

* Paper pdf correction for 2023.acl-long.519, closes #2958.

* Paper pdf correction for 2023.findings-emnlp.279, closes #2969.

* Meta data correction for 2023.findings-emnlp.325, closes #2970.

* Author V.S.D.S.Mahesh Akavarapu meta data correction, closes #2967.

* Workshop meta data correction for 2023.bigpicture-1 & 2023.newsum-1, closes #2965.

* Meta data correction for 2023.emnlp-main.166, closes #2964.

* Paper pdf correction for 2023.emnlp-main.73, closes #2961.

* Paper meta data correction for 2023.arabicnlp-1.55, closes #2945.

* Paper meta data correction for 2023.arabicnlp-1.87, closes #2946.

* Author meta data correction for 2023.genbench-1.4, closes #2960.

* Paper meta data correction for 2023.findings-emnlp.993, closes #2959.

* Author meta data correction for 2023.findings-emnlp.469, closes #2956.

* Author meta data correction for 2023.emnlp-main.783, closes #2951.

* Author meta data correction for 2023.emnlp-main.953, closes #2947.

* Author meta data correction for 2023.emnlp-main.470, closes #2972.

* Paper meta data correction for 2023.emnlp-main.237, closes #2978.

* Author name meta data correction for 2023.arabicnlp-1.53, closes #2980.

* Paper pdf correction for 2023.acl-long.513, closes #2982.

* Paper pdf correction for 2023.emnlp-main.434, closes #2985.

* Paper pdf correction for 2023.findings-emnlp.709, closes #2991.

* Author meta data correctioon for 2023.findings-emnlp.382, closes #2984.

* Paper meta data correction for 2023.emnlp-main.608, closes #2987.

* Author meta data correction for 2023.emnlp-main.714, closes #2988.

* Author meta data correction for 2023.emnlp-main.724, closes #2990.

* inserted license comment.

* Author yang-liu-edinburgh update.

---------

Co-authored-by: Dan Gildea <dgildea>
Co-authored-by: Jonathan K. Kummerfeld <[email protected]>
Co-authored-by: m-chaves <[email protected]>

data/xml/2020.emnlp.xml

@@ -3817,11 +3817,13 @@
       <author><first>Ryan</first><last>Cotterell</last></author>
       <pages>3138–3153</pages>
       <abstract>The question of how to probe contextual word representations in a way that is principled and useful has seen significant recent attention. In our contribution to this discussion, we argue, first, for a probe metric that reflects the trade-off between probe complexity and performance: the Pareto hypervolume. To measure complexity, we present a number of parametric and non-parametric metrics. Our experiments with such metrics show that probe’s performance curves often fail to align with widely accepted rankings between language representations (with, e.g., non-contextual representations outperforming contextual ones). These results lead us to argue, second, that common simplistic probe tasks such as POS labeling and dependency arc labeling, are inadequate to evaluate the properties encoded in contextual word representations. We propose full dependency parsing as an example probe task, and demonstrate it with the Pareto hypervolume. In support of our arguments, the results of this illustrative experiment conform closer to accepted rankings among contextual word representations.</abstract>
-      <url hash="a71abac5">2020.emnlp-main.254</url>
+      <url hash="8f9c945f">2020.emnlp-main.254</url>
       <doi>10.18653/v1/2020.emnlp-main.254</doi>
       <video href="https://slideslive.com/38938932"/>
       <bibkey>pimentel-etal-2020-pareto</bibkey>
       <pwccode url="https://github.com/rycolab/pareto-probing" additional="false">rycolab/pareto-probing</pwccode>
+      <revision id="1" href="2020.emnlp-main.254v1" hash="a71abac5"/>
+      <revision id="2" href="2020.emnlp-main.254v2" hash="8f9c945f" date="2023-12-10">Updated appendix.</revision>
     </paper>
     <paper id="255">
       <title>Interpretation of <fixed-case>NLP</fixed-case> models through input marginalization</title>

data/xml/2022.acl.xml

@@ -7691,11 +7691,11 @@ in the Case of Unambiguous Gender</title>
     </paper>
     <paper id="480">
       <title>From Simultaneous to Streaming Machine Translation by Leveraging Streaming History</title>
-      <author><first>Javier</first><last>Iranzo Sanchez</last></author>
+      <author><first>Javier</first><last>Iranzo-Sánchez</last></author>
       <author><first>Jorge</first><last>Civera</last></author>
-      <author><first>Alfons</first><last>Juan-Císcar</last></author>
+      <author><first>Alfons</first><last>Juan</last></author>
       <pages>6972-6985</pages>
-      <abstract>Simultaneous machine translation has recently gained traction thanks to significant quality improvements and the advent of streaming applications. Simultaneous translation systems need to find a trade-off between translation quality and response time, and with this purpose multiple latency measures have been proposed. However, latency evaluations for simultaneous translation are estimated at the sentence level, not taking into account the sequential nature of a streaming scenario. Indeed, these sentence-level latency measures are not well suited for continuous stream translation, resulting in figures that are not coherent with the simultaneous translation policy of the system being assessed. This work proposes a stream-level adaptation of the current latency measures based on a re-segmentation approach applied to the output translation, that is successfully evaluated on streaming conditions for a reference IWSLT task</abstract>
+      <abstract>Simultaneous Machine Translation is the task of incrementally translating an input sentence before it is fully available. Currently, simultaneous translation is carried out by translating each sentence independently of the previously translated text. More generally, Streaming MT can be understood as an extension of Simultaneous MT to the incremental translation of a continuous input text stream. In this work, a state-of-the-art simultaneous sentence-level MT system is extended to the streaming setup by leveraging the streaming history. Extensive empirical results are reported on IWSLT Translation Tasks, showing that leveraging the streaming history leads to significant quality gains. In particular, the proposed system proves to compare favorably to the best performing systems.</abstract>
       <url hash="3089f909">2022.acl-long.480</url>
       <attachment type="software" hash="9007090a">2022.acl-long.480.software.zip</attachment>
       <bibkey>iranzo-sanchez-etal-2022-simultaneous</bibkey>

data/xml/2022.emnlp.xml

@@ -9701,7 +9701,7 @@
     </paper>
     <paper id="744">
       <title><fixed-case>KOLD</fixed-case>: <fixed-case>K</fixed-case>orean Offensive Language Dataset</title>
-      <author><first>Younghoon</first><last>Jeong</last><affiliation>KAIST (Korea Advanced Institute of Science and Technology)</affiliation></author>
+      <author><first>Younghun</first><last>Jeong</last><affiliation>KAIST (Korea Advanced Institute of Science and Technology)</affiliation></author>
       <author><first>Juhyun</first><last>Oh</last><affiliation>Independent Researcher</affiliation></author>
       <author><first>Jongwon</first><last>Lee</last><affiliation>Samsung Research</affiliation></author>
       <author><first>Jaimeen</first><last>Ahn</last><affiliation>Independent Researcher</affiliation></author>
@@ -10260,13 +10260,13 @@
       <doi>10.18653/v1/2022.emnlp-main.787</doi>
     </paper>
     <paper id="788">
-      <title>Attentional Probe: Estimating a Module’s Functional Potential</title>
+      <title>The Architectural Bottleneck Principle</title>
       <author><first>Tiago</first><last>Pimentel</last><affiliation>University of Cambridge</affiliation></author>
       <author><first>Josef</first><last>Valvoda</last><affiliation>University of Cambridge</affiliation></author>
       <author><first>Niklas</first><last>Stoehr</last><affiliation>ETH Zurich</affiliation></author>
       <author><first>Ryan</first><last>Cotterell</last><affiliation>ETH Zürich</affiliation></author>
       <pages>11459-11472</pages>
-      <abstract/>
+      <abstract>In this paper, we seek to measure how much information a component in a neural network could extract from the representations fed into it. Our work stands in contrast to prior probing work, most of which investigates how much information a model's representations contain. This shift in perspective leads us to propose a new principle for probing, the architectural bottleneck principle: In order to estimate how much information a given component could extract, a probe should look exactly like the component. Relying on this principle, we estimate how much syntactic information is available to transformers through our attentional probe, a probe that exactly resembles a transformer's self-attention head. Experimentally, we find that, in three models (BERT, ALBERT, and RoBERTa), a sentence's syntax tree is mostly extractable by our probe, suggesting these models have access to syntactic information while composing their contextual representations. Whether this information is actually used by these models, however, remains an open question..</abstract>
       <url hash="fbec6a23">2022.emnlp-main.788</url>
       <bibkey>pimentel-etal-2022-attentional</bibkey>
       <doi>10.18653/v1/2022.emnlp-main.788</doi>

data/xml/2022.inlg.xml

@@ -578,7 +578,7 @@
       <title><fixed-case>D</fixed-case>ialog<fixed-case>S</fixed-case>um Challenge: Results of the Dialogue Summarization Shared Task</title>
       <author><first>Yulong</first><last>Chen</last></author>
       <author><first>Naihao</first><last>Deng</last></author>
-      <author id="yang-liu"><first>Yang</first><last>Liu</last></author>
+      <author id="yang-liu-edinburgh"><first>Yang</first><last>Liu</last></author>
       <author><first>Yue</first><last>Zhang</last></author>
       <pages>94-103</pages>
       <abstract>We report the results of DialogSum Challenge, the shared task on summarizing real-life sce- nario dialogues at INLG 2022. Four teams participate in this shared task and three submit their system reports, exploring different meth- ods to improve the performance of dialogue summarization. Although there is a great im- provement over the baseline models regarding automatic evaluation metrics, such as ROUGE scores, we find that there is a salient gap be- tween model generated outputs and human an- notated summaries by human evaluation from multiple aspects. These findings demonstrate the difficulty of dialogue summarization and suggest that more fine-grained evaluatuion met- rics are in need.</abstract>

data/xml/2023.acl.xml

@@ -3657,7 +3657,7 @@
       <author><first>Pengcheng</first><last>He</last><affiliation>Microsoft</affiliation></author>
       <author><first>Baolin</first><last>Peng</last><affiliation>Tencent AI Lab</affiliation></author>
       <author><first>Song</first><last>Wang</last><affiliation>Microsoft Azure AI</affiliation></author>
-      <author id="yang-liu"><first>Yang</first><last>Liu</last><affiliation>Microsoft</affiliation></author>
+      <author id="yang-liu-edinburgh"><first>Yang</first><last>Liu</last><affiliation>Microsoft</affiliation></author>
       <author><first>Ruochen</first><last>Xu</last><affiliation>Microsoft</affiliation></author>
       <author><first>Hany</first><last>Hassan</last><affiliation>Microsoft</affiliation></author>
       <author><first>Yu</first><last>Shi</last><affiliation>Microsoft</affiliation></author>
@@ -6827,9 +6827,11 @@
       <author><first>Nanyun</first><last>Peng</last><affiliation>University of California, Los Angeles</affiliation></author>
       <pages>9235-9254</pages>
       <abstract>Automatic melody-to-lyric generation is a task in which song lyrics are generated to go with a given melody. It is of significant practical interest and more challenging than unconstrained lyric generation as the music imposes additional constraints onto the lyrics. The training data is limited as most songs are copyrighted, resulting in models that underfit the complicated cross-modal relationship between melody and lyrics. In this work, we propose a method for generating high-quality lyrics without training on any aligned melody-lyric data. Specifically, we design a hierarchical lyric generation framework that first generates a song outline and second the complete lyrics. The framework enables disentanglement of training (based purely on text) from inference (melody-guided text generation) to circumvent the shortage of parallel data. We leverage the segmentation and rhythm alignment between melody and lyrics to compile the given melody into decoding constraints as guidance during inference. The two-step hierarchical design also enables content control via the lyric outline, a much-desired feature for democratizing collaborative song creation. Experimental results show that our model can generate high-quality lyrics that are more on-topic, singable, intelligible, and coherent than strong baselines, for example SongMASS, a SOTA model trained on a parallel dataset, with a 24% relative overall quality improvement based on human ratings. Our code is available at <url>https://github.com/amazon-science/unsupervised-melody-to-lyrics-generation</url>.</abstract>
-      <url hash="67448d56">2023.acl-long.513</url>
+      <url hash="f8e119cc">2023.acl-long.513</url>
       <bibkey>tian-etal-2023-unsupervised</bibkey>
       <doi>10.18653/v1/2023.acl-long.513</doi>
+      <revision id="1" href="2023.acl-long.513v1" hash="67448d56"/>
+      <revision id="2" href="2023.acl-long.513v2" hash="f8e119cc" date="2023-12-31">Added description of authors contributions.</revision>
     </paper>
     <paper id="514">
       <title>Causality-aware Concept Extraction based on Knowledge-guided Prompting</title>
@@ -6917,9 +6919,11 @@
       <author><first>Yue</first><last>Zhang</last><affiliation>Westlake University</affiliation></author>
       <pages>9332-9351</pages>
       <abstract>Most existing cross-lingual summarization (CLS) work constructs CLS corpora by simply and directly translating pre-annotated summaries from one language to another, which can contain errors from both summarization and translation processes. To address this issue, we propose ConvSumX, a cross-lingual conversation summarization benchmark, through a new annotation schema that explicitly considers source input context. ConvSumX consists of 2 sub-tasks under different real-world scenarios, with each covering 3 language directions. We conduct thorough analysis on ConvSumX and 3 widely-used manually annotated CLS corpora and empirically find that ConvSumX is more faithful towards input text. Additionally, based on the same intuition, we propose a 2-Step method, which takes both conversation and summary as input to simulate human annotation process. Experimental results show that 2-Step method surpasses strong baselines on ConvSumX under both automatic and human evaluation. Analysis shows that both source input text and summary are crucial for modeling cross-lingual summaries.</abstract>
-      <url hash="bfa2f562">2023.acl-long.519</url>
+      <url hash="087cb555">2023.acl-long.519</url>
       <bibkey>chen-etal-2023-revisiting</bibkey>
       <doi>10.18653/v1/2023.acl-long.519</doi>
+      <revision id="1" href="2023.acl-long.519v1" hash="bfa2f562"/>
+      <revision id="2" href="2023.acl-long.519v2" hash="087cb555" date="2023-12-15">Correct acknowledgement.</revision>
     </paper>
     <paper id="520">
       <title>Learning Dynamic Contextualised Word Embeddings via Template-based Temporal Adaptation</title>
@@ -9612,7 +9616,7 @@
     <paper id="718">
       <title><fixed-case>U</fixed-case>ni<fixed-case>S</fixed-case>umm and <fixed-case>S</fixed-case>umm<fixed-case>Z</fixed-case>oo: Unified Model and Diverse Benchmark for Few-Shot Summarization</title>
       <author><first>Yulong</first><last>Chen</last><affiliation>Zhejiang University, Westlake University</affiliation></author>
-      <author id="yang-liu"><first>Yang</first><last>Liu</last><affiliation>Microsoft</affiliation></author>
+      <author id="yang-liu-edinburgh"><first>Yang</first><last>Liu</last><affiliation>Microsoft</affiliation></author>
       <author><first>Ruochen</first><last>Xu</last><affiliation>Microsoft</affiliation></author>
       <author><first>Ziyi</first><last>Yang</last><affiliation>Microsoft Research</affiliation></author>
       <author><first>Chenguang</first><last>Zhu</last><affiliation>Microsoft Cognitive Services Research Group</affiliation></author>

data/xml/2023.arabicnlp.xml

@@ -363,7 +363,7 @@
     <paper id="28">
       <title>In-Context Meta-Learning vs. Semantic Score-Based Similarity: A Comparative Study in <fixed-case>A</fixed-case>rabic Short Answer Grading</title>
       <author><first>Menna</first><last>Fateen</last></author>
-      <author><first>Tsunenori</first><last>Mina</last></author>
+      <author><first>Tsunenori</first><last>Mine</last></author>
       <pages>350-358</pages>
       <abstract>Delegating short answer grading to automated systems enhances efficiency, giving teachers more time for vital human-centered aspects of education. Studies in automatic short answer grading (ASAG) approach the problem from instance-based or reference-based perspectives. Recent studies have favored instance-based methods, but they demand substantial data for training, which is often scarce in classroom settings. This study compares both approaches using an Arabic ASAG dataset. We employ in-context meta-learning for instance-based and semantic score-based similarity for reference-based grading. Results show both methods outperform a baseline and occasionally even surpass human raters when grading unseen answers. Notably, the semantic score-based similarity approach excels in zero-shot settings, outperforming in-context meta-learning. Our work contributes insights to Arabic ASAG and introduces a prompt category classification model, leveraging GPT3.5 to augment Arabic data for improved performance.</abstract>
       <url hash="5e70b1c8">2023.arabicnlp-1.28</url>
@@ -662,7 +662,7 @@
       <title>Itri Amigos at <fixed-case>A</fixed-case>r<fixed-case>AIE</fixed-case>val Shared Task: Transformer vs. Compression-Based Models for Persuasion Techniques and Disinformation Detection</title>
       <author><first>Jehad</first><last>Oumer</last></author>
       <author><first>Nouman</first><last>Ahmed</last></author>
-      <author><first>Natalia</first><last>Manrique</last></author>
+      <author><first>Natalia</first><last>Flechas Manrique</last></author>
       <pages>543-548</pages>
       <abstract>Social media has significantly amplified the dissemination of misinformation. Researchers have employed natural language processing and machine learning techniques to identify and categorize false information on these platforms. While there is a well-established body of research on detecting fake news in English and Latin languages, the study of Arabic fake news detection remains limited. This paper describes the methods used to tackle the challenges of the ArAIEval shared Task 2023. We conducted experiments with both monolingual Arabic and multi-lingual pre-trained Language Models (LM). We found that the monolingual Arabic models outperformed in all four subtasks. Additionally, we explored a novel lossless compression method, which, while not surpassing pretrained LM performance, presents an intriguing avenue for future experimentation to achieve comparable results in a more efficient and rapid manner.</abstract>
       <url hash="5a716bbf">2023.arabicnlp-1.53</url>
@@ -682,7 +682,7 @@
     <paper id="55">
       <title><fixed-case>UL</fixed-case> &amp; <fixed-case>UM</fixed-case>6<fixed-case>P</fixed-case> at <fixed-case>A</fixed-case>r<fixed-case>AIE</fixed-case>val Shared Task: Transformer-based model for Persuasion Techniques and Disinformation detection in <fixed-case>A</fixed-case>rabic</title>
       <author><first>Salima</first><last>Lamsiyah</last></author>
-      <author><first>Abdelkader</first><last>Mahdaouy</last></author>
+      <author><first>Abdelkader</first><last>El Mahdaouy</last></author>
       <author><first>Hamza</first><last>Alami</last></author>
       <author><first>Ismail</first><last>Berrada</last></author>
       <author><first>Christoph</first><last>Schommer</last></author>
@@ -1050,7 +1050,7 @@
     </paper>
     <paper id="87">
       <title><fixed-case>UM</fixed-case>6<fixed-case>P</fixed-case> &amp; <fixed-case>UL</fixed-case> at <fixed-case>W</fixed-case>ojood<fixed-case>NER</fixed-case> shared task: Improving Multi-Task Learning for Flat and Nested <fixed-case>A</fixed-case>rabic Named Entity Recognition</title>
-      <author><first>Abdelkader</first><last>Mahdaouy</last></author>
+      <author><first>Abdelkader</first><last>El Mahdaouy</last></author>
       <author><first>Salima</first><last>Lamsiyah</last></author>
       <author><first>Hamza</first><last>Alami</last></author>
       <author><first>Christoph</first><last>Schommer</last></author>