Thailand

Existing work in multilingual pretraining has demonstrated the potential of cross-lingual transferability by training a unified Transformer encoder for multiple languages. However, much of this work only relies on the shared vocabulary and bilingual contexts to encourage the correlation across languages, which is loose and implicit for aligning the contextual representations between languages. In this paper, we plug a cross-attention module into the Transformer encoder to explicitly build the interdependence between languages. It can effectively avoid the degeneration of predicting masked words only conditioned on the context in its own language. More importantly, when fine-tuning on downstream tasks, the cross-attention module can be plugged in or out on-demand, thus naturally benefiting a wider range of cross-lingual tasks, from language understanding to generation. 

As a result, the proposed cross-lingual model delivers new state-of-the-art results on various cross-lingual understanding tasks of the XTREME benchmark, covering text classification, sequence labeling, question answering, and sentence retrieval. For cross-lingual generation tasks, it also outperforms all existing cross-lingual models and state-of-the-art Transformer variants on WMT14 English-to-German and English-to-French translation datasets, with gains of up to 1~2 BLEU.

IJCNLP-AACL 2021

VECO: Variable and Flexible Cross-lingual Pre-training for Language Understanding and Generation

cross-lingual pre-training

language models

machine translation

**Welcome to ACL-IJCNLP 2021!**

The great event is jointly organized by the Association for Computational Linguistics (ACL) and Asian Federation of Natural Language Processing (AFNLP). 

As in previous years, the program of the conference includes a poster session, tutorials, workshops and demonstrations in addition to the main conference.


We were able to keep the registration fees similar to those charged for the virtual ACL 2020. The one fee allows attendance at the main conference and any/all tutorials and workshops. These fees would be $125 Regular Early and $175 Regular Late; $50 Student Early and $75 Student Late. Early registration closes at midnight July 11, 2021 (Eastern Daylight Time).

**Reminder:** It is ACL’s policy that at least one author of each accepted paper (including ACL Finding papers) must register for the conference.

**Reminder2:** Underline site will open closer to the event. If you already registered you will receive access detail

Registration is now open

The great event is jointly organized by the Association for Computational Linguistics (ACL) and Asian Federation of Natural Language Processing (AFNLP).

technical paper

Atomic clauses are fundamental text units for understanding complex sentences. Identifying the atomic sentences within complex sentences is important for applications such as summarization, argument mining, discourse analysis, discourse parsing, and question answering. Previous work mainly relies on rule-based methods dependent on parsing. We propose a new task to decompose each complex sentence into simple sentences derived from the tensed clauses in the source, and a novel problem formulation as a graph edit task. Our neural model learns to Accept, Break, Copy or Drop elements of a graph that combines word adjacency and grammatical dependencies. The full processing pipeline includes modules for graph construction, graph editing, and sentence generation from the output graph. We introduce DeSSE, a new dataset designed to train and evaluate complex sentence decomposition, and MinWiki, a subset of MinWikiSplit. ABCD achieves comparable performance as two parsing baselines on MinWiki. On DeSSE, which has a more even balance of complex sentence types, our model achieves higher accuracy on the number of atomic sentences than an encoder-decoder baseline. Results include a detailed error analysis.

ABCD: A Graph Framework to Convert Complex Sentences to a Covering Set of Simple Sentences

Many Question-Answering (QA) datasets contain unanswerable questions, but their treatment in QA systems remains primitive. Our analysis of the Natural Questions (Kwiatkowski et al. 2019) dataset reveals that a substantial portion of unanswerable questions (~21%) can be explained based on the presence of unverifiable presuppositions. Through a user preference study, we demonstrate that the oracle behavior of our proposed system---which provides responses based on presupposition failure---is preferred over the oracle behavior of existing QA systems. Then, we present a novel framework for implementing such a system in three steps: presupposition generation, presupposition verification, and explanation generation, reporting progress on each. Finally, we show that a simple modification of adding presuppositions and their verifiability to the input of a competitive end-to-end QA system yields modest gains in QA performance and unanswerability detection, demonstrating the promise of our approach.

Which Linguist Invented the Lightbulb? Presupposition Verification for Question-Answering

Text-level discourse rhetorical structure (DRS) parsing is known to be challenging due to the notorious lack of training data. Although recent top-down DRS parsers can better leverage global document context and have achieved certain success, the performance is still far from perfect. To our knowledge, all previous DRS parsers make local decisions for either bottom-up node composition or top-down split point ranking at each time step, and largely ignore DRS parsing from the global view point. Obviously, it is not sufficient to build an entire DRS tree only through these local decisions. In this work, we present our insight on evaluating the pros and cons of the entire DRS tree for global optimization. Specifically, based on recent well-performing top-down frameworks, we introduce a novel method to transform both gold standard and predicted constituency trees into tree diagrams with two color channels. After that, we learn an adversarial bot between gold and fake tree diagrams to estimate the generated DRS trees from a global perspective. We perform experiments on both RST-DT and CDTB corpora and use the original Parseval for performance evaluation. The experimental results show that our parser can substantially improve the performance when compared with previous state-of-the-art parsers.

Adversarial Learning for Discourse Rhetorical Structure Parsing

Discourse relations among arguments reveal logical structures of a debate conversation. However, no prior work has explicitly studied how the sequence of discourse relations influence a claim's impact. This paper empirically shows that the discourse relations between two arguments along the context path are essential factors for identifying the persuasive power of an argument. We further propose DisCOC to inject and fuse the sentence-level structural discourse information with contextualized features derived from large-scale language models. Experimental results and extensive analysis show that the attention and gate mechanisms that explicitly model contexts and texts can indeed help the argument impact classification task defined by Durmus et al. (2019), and discourse structures among the context path of the claim to be classified can further boost the performance.

Exploring Discourse Structures for Argument Impact Classification

This paper proposes a sophisticated neural architecture to incorporate bilingual dictionaries into Neural Machine Translation (NMT) models. By introducing three novel components: Pointer, Disambiguator, and Copier, our method PDC achieves the following merits inherently compared with previous efforts: (1) Pointer leverages the semantic information from bilingual dictionaries, for the first time, to better locate source words whose translation in dictionaries can potentially be used; (2) Disambiguator synthesizes contextual information from the source view and the target view, both of which contribute to distinguishing the proper translation of a specific source word from multiple candidates in dictionaries; (3) Copier systematically connects Pointer and Disambiguator based on a hierarchical copy mechanism seamlessly integrated with Transformer, thereby building an end-to-end architecture that could avoid error propagation problems in alternative pipe-line methods. The experimental results on Chinese-English and English-Japanese benchmarks demonstrate the PDC's overall superiority and effectiveness of each component.

Point, Disambiguate and Copy: Incorporating Bilingual Dictionaries for Neural Machine Translation

State-of-the-art Variational Auto-Encoders (VAEs) for learning disentangled latent representations give impressive results in discovering features like pitch, pause duration, and accent in speech data, leading to highly controllable text-to-speech (TTS) synthesis. However, these LSTM-based VAEs fail to learn latent clusters of speaker attributes when trained on limited or noisy datasets. Further, different latent variables are found to encode the same features, limiting the control and expressiveness during speech synthesis. To resolve these issues, we propose REMMI (Reordered transformer Encoder with Minimal Mutual Information) where we minimize the mutual information between different latent variables and devise a modified Transformer architecture with layer reordering to learn controllable latent representations in speech data. We show that REMMI reduces the cluster overlap of speaker attributes by at least 30% over LSTM-VAE.

Learning Robust Latent Representations for Controllable Speech Synthesis

Having recognized gender bias as a major issue affecting current translation technologies, researchers have primarily attempted to mitigate it by working on the data front. However, whether algorithmic aspects concur to exacerbate unwanted outputs remains so far under-investigated. In this work, we bring the analysis on gender bias in automatic translation onto a seemingly neutral yet critical component: word segmentation. Can segmenting methods influence the ability to translate gender? Do certain segmentation approaches penalize the representation of feminine linguistic markings? We address these questions by comparing 5 existing segmentation strategies on the target side of speech translation systems. Our results on two language pairs (English-Italian/French) show that state-of-the-art subword splitting (BPE) comes at the cost of higher gender bias. In light of this finding, we propose a combined approach that preserves BPE overall translation quality, while leveraging the higher ability of character-based segmentation to properly translate gender.

How to Split: the Effect of Word Segmentation on Gender Bias in Speech Translation

In this paper, we analyze the interplay between the use of offensive language and mental health. We acquired publicly available datasets created for offensive language identification and depression detection and we train computational models to compare the use of offensive language in social media posts written by groups of individuals with and without self-reported depression diagnosis. We also look at samples written by groups of individuals whose posts show signs of depression according to recent related studies. Our analysis indicates that offensive language is more frequently used in the samples written by individuals with self-reported depression as well as individuals showing signs of depression. The results discussed here open new avenues in research in politeness/offensiveness and mental health.

An Exploratory Analysis of the Relation between Offensive Language and Mental Health

Automated Term Extraction (ATE), even though well-investigated, continues to be a challenging task. Approaches conventionally extract terms on corpus or document level and the benefits of neural models still remain underexplored with very few exceptions. We introduce three transformer-based term extraction models operating on sentence level: a language model for token classification, one for sequence classification, and an innovative use of Neural Machine Translation (NMT), which learns to reduce sentences to terms. All three models are trained and tested on the dataset of the ATE challenge TermEval 2020 in English, French, and Dutch across four specialized domains. The two best performing approaches are also evaluated on the ACL RD-TEC 2.0 dataset. Our models outperform previous baselines, one of which is BERT-based, by a substantial margin, with the token-classifier language model performing best.

Transforming Term Extraction: Transformer-Based Approaches to Multilingual Term Extraction Across Domains

Transformers have been shown to emulate logical deduction over natural language theories (logical rules expressed in natural language), reliably assigning true/false labels to candidate implications. However, their ability to *generate* implications of a theory has not yet been demonstrated, and methods for reconstructing proofs of answers are imperfect. In this work we show that a generative model, called ProofWriter, can reliably generate both implications of a theory and the natural language proofs that support them. In particular, iterating a 1-step implication generator results in proofs that are highly reliable, and represent actual model decisions (rather than post-hoc rationalizations). On the RuleTaker dataset, the accuracy of ProofWriter's proofs exceed previous methods by +9% absolute, and in a way that generalizes to proof depths unseen in training and on out-of-domain problems. We also show that generative techniques can perform a type of abduction with high precision: Given a theory and an unprovable conclusion, identify a missing fact that allows the conclusion to be proved, along with a proof. These results significantly improve the viability of neural methods for systematically reasoning over natural language.

VECO: Variable and Flexible Cross-lingual Pre-training for Language Understanding and Generation

Downloads

Next from IJCNLP-AACL 2021

ABCD: A Graph Framework to Convert Complex Sentences to a Covering Set of Simple Sentences

Similar lecture

Discriminative Reranking for Neural Machine Translation

VECO: Variable and Flexible Cross-lingual Pre-training for Language Understanding and Generation

.css-70qvj9{display:-webkit-box;display:-webkit-flex;display:-ms-flexbox;display:flex;-webkit-align-items:center;-webkit-box-align:center;-ms-flex-align:center;align-items:center;}Downloads

Next from IJCNLP-AACL 2021

ABCD: A Graph Framework to Convert Complex Sentences to a Covering Set of Simple Sentences

Similar lecture

Discriminative Reranking for Neural Machine Translation

Downloads