Thailand

Attention mechanisms have achieved substantial improvements in neural machine translation by dynamically selecting relevant inputs for different predictions. However, recent studies have questioned the attention mechanisms&#39; capability for discovering decisive inputs. In this paper, we propose to calibrate the attention weights by introducing a mask perturbation model that automatically evaluates each input&#39;s contribution to the model outputs. We increase the attention weights assigned to the indispensable tokens, whose removal leads to a dramatic performance decrease. The extensive experiments on the Transformer-based translation have demonstrated the effectiveness of our model. We further find that the calibrated attention weights are more uniform at lower layers to collect multiple information while more concentrated on the specific inputs at higher layers. Detailed analyses also show a great need for calibration in the attention weights with high entropy where the model is unconfident about its decision.

ACL 2021

Attention Calibration for Transformer in Neural Machine Translation

attention mechanisms

neural machine translation

transformer

**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). 

This Joint Conference of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (ACL-IJCNLP 2021) was to be held in Bangkok, Thailand, during August 1-6, 2021 and is now virtual on the Underline platform.

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

Attention mechanisms have achieved substantial improvements in neural machine translation by dynamically selecting relevant inputs for different predictions. However, recent studies have questioned the attention mechanisms' capability for discovering decisive inputs. In this paper, we propose to calibrate the attention weights by introducing a mask perturbation model that automatically evaluates each input's contribution to the model outputs. We increase the attention weights assigned to the indispensable tokens, whose removal leads to a dramatic performance decrease. The extensive experiments on the Transformer-based translation have demonstrated the effectiveness of our model. We further find that the calibrated attention weights are more uniform at lower layers to collect multiple information while more concentrated on the specific inputs at higher layers. Detailed analyses also show a great need for calibration in the attention weights with high entropy where the model is unconfident about its decision.

poster

Optical character recognition (OCR) is crucial for a deeper access to historical collections. OCR needs to account for orthographic variations, typefaces, or language evolution (i.e., new letters, word spellings), as the main source of character, word, or word segmentation transcription errors. For digital corpora of historical prints, the errors are further exacerbated due to low scan quality and lack of language standardization.

For the task of OCR post-hoc correction, we propose a neural approach based on a combination of recurrent (RNN) and deep convolutional network (ConvNet) to correct OCR transcription errors. At character level we flexibly capture errors, and decode the corrected output based on a novel attention mechanism. Accounting for the input and output similarity, we propose a new loss function that rewards the model’s correcting behavior.

Evaluation on a historical book corpus in German language shows that our models
are robust in capturing diverse OCR transcription errors and reduce the word error rate of 32.3% by more than 89%.

Neural OCR Post-Hoc Correction of Historical Corpora

Ideology of legislators is typically estimated by ideal point models from historical records of votes. It represents legislators and legislation as points in a latent space and shows promising results for modeling voting behavior. However, it fails to capture more specific attitudes of legislators toward emerging issues and is unable to model newly-elected legislators without voting histories. In order to mitigate these two problems, we explore to incorporate both voting behavior and public statements on Twitter to jointly model legislators. In addition, we propose a novel task, namely hashtag usage prediction to model the ideology of legislators on Twitter. In practice, we construct a heterogeneous graph for the legislative context and use relational graph neural networks to learn the representation of legislators with the guidance of historical records of their voting and hashtag usage. Experiment results indicate that our model yields significant improvements for the task of roll call vote prediction. Further analysis further demonstrates that legislator representation we learned captures nuances in statements.

Align Voting Behavior with Public Statements for Legislator Representation Learning

Languages are dynamic systems: word usage may change over time, reflecting various societal factors. However, all languages do not evolve identically: the impact of an event, the influence of a trend or thinking, can differ between communities.
In this paper, we propose to track these divergences by comparing the evolution of a word and its translation across two languages. We investigate several methods of building time-varying and bilingual word embeddings, using contextualised and non-contextualised embeddings. We propose a set of scenarios to characterize semantic divergence across two languages, along with a setup to differentiate them in a bilingual corpus. We evaluate the different methods by generating a corpus of synthetic semantic change across two languages, English and French, before applying them to newspaper corpora to detect bilingual semantic divergence and provide qualitative insight for the task. We conclude that BERT embeddings coupled with a clustering step lead to the best performance on synthetic corpora; however, the performance of CBOW embeddings is very competitive and more adapted to an exploratory analysis on a large corpus.

Measure and Evaluation of Semantic Divergence across Two Languages

Multilingual neural machine translation has shown the capability of directly translating between language pairs unseen in training, i.e. zero-shot translation. Despite being conceptually attractive, it often suffers from low output quality. The difficulty of generalizing to new translation directions suggests the model representations are highly specific to those language pairs seen in training. We demonstrate that a main factor causing the language-specific representations is the positional correspondence to input tokens. We show that this can be easily alleviated by removing residual connections in an encoder layer. With this modification, we gain up to 18.5 BLEU points on zero-shot translation while retaining quality on supervised directions. The improvements are particularly prominent between related languages, where our proposed model outperforms pivot-based translation. Moreover, our approach allows easy integration of new languages, which substantially expands translation coverage. By thorough inspections of the hidden layer outputs, we show that our approach indeed leads to more language-independent representations.

Improving Zero-Shot Translation by Disentangling Positional Information

Commonsense reasoning research has so far been limited to English.
We aim to evaluate and improve popular multilingual language models (ML-LMs) to help advance commonsense reasoning (CSR) beyond English.
We collect the Mickey corpus, consisting of 561k sentences in 11 different languages, which
can be used for analyzing and improving ML-LMs.
We propose Mickey Probe, a language-general probing task for fairly evaluating the common sense of popular ML-LMs across different languages.
In addition, we also create two new datasets, X-CSQA and X-CODAH, by translating their English versions to 14 other languages, so that we can evaluate popular ML-LMs for cross-lingual commonsense reasoning.
To improve the performance beyond English, 
we propose a simple yet effective method --- multilingual contrastive pretraining (MCP).
It significantly enhances sentence representations, yielding a large performance gain on both benchmarks (e.g., +2.7% accuracy for X-CSQA over XLM-R_L).

Common Sense Beyond English: Evaluating and Improving Multilingual Language Models for Commonsense Reasoning

It is well-known that typical word embedding methods such as Word2Vec and GloVe have the property that the meaning can be composed by adding up the embeddings (additive compositionality). Several theories have been proposed to explain additive compositionality, but the following questions remain unanswered: (Q1) The assumptions of those theories do not hold for the practical word embedding. (Q2) Ordinary additive compositionality can be seen as an AND operation of word meanings, but it is not well understood how other operations, such as OR and NOT, can be computed by the embeddings. We address these issues by the idea of frequency-weighted centering at its core. This paper proposes a post-processing method for bridging the gap between practical word embedding and the assumption of theory about additive compositionality as an answer to (Q1). It also gives a method for taking OR or NOT of the meaning by linear operation of word embedding as an answer to (Q2). Moreover, we confirm experimentally that the accuracy of AND operation, i.e., the ordinary additive compositionality, can be improved by our post-processing method (3.5x improvement in top-100 accuracy) and that OR and NOT operations can be performed correctly.

Revisiting Additive Compositionality: AND, OR and NOT Operations with Word Embeddings

Various studies show that pretrained language models such as BERT cannot straightforwardly replace encoders in neural machine translation despite their enormous success in other tasks. This is even more astonishing considering the similarities between the architectures. This paper sheds some light on the embedding spaces they create, using average cosine similarity, contextuality metrics and measures for representational similarity for comparison, revealing that BERT and NMT encoder representations look significantly different from one another. 
In order to address this issue, we propose a supervised transformation from one into the other using explicit alignment and fine-tuning. Our results demonstrate the need for such a transformation to improve the applicability of BERT in MT.

On the differences of Bert and MT encoder spaces and how to address them in translation tasks

This paper proposes a novel attention mechanism for Transformer Neural Machine Translation, “Synchronous Syntactic Attention,” inspired by synchronous dependency grammars. The mechanism synchronizes source-side and target-side syntactic self-attentions by minimizing the difference between target-side self-attentions and the source-side self-attentions mapped by the encoder-decoder attention matrix. The experiments show that the proposed method improves the translation performance on WMT14 En-De, WMT16 En-Ro, and ASPEC Ja-En (up to +0.38 points in BLEU).

Synchronous Syntactic Attention for Transformer Neural Machine Translation

This paper investigates bounds on the generative capacity of prosodic processes, by focusing on the complexity of recursive prosody in coordination contexts in English (Wagner, 2010). Although all phonological processes and most prosodic processes are computationally regular string languages, we show that recursive prosody is not. The output string language is instead parallel multiple context-free (Seki et al., 1991). We evaluate the complexity of the pattern over strings, and then move on to a characterization over trees that requires the expressivity of multi bottom-up tree transducers. In doing so, we provide a foundation for future mathematically grounded investigations of the syntax-prosody interface.

Recursive prosody is not finite-state

Raimy (1999; 2000a; 2000b) proposed a graphical formalism for modeling reduplication, originallymostly focused on phonological overapplication in a derivational framework. This framework is now known as Precedence-based phonology or Multiprecedence phonology. Raimy's idea is that the segments at the input to the phonology are not totally ordered by precedence. This paper tackles a challenge that arose with Raimy's work, the development of a deterministic serialization algorithm as part of the derivation of surface forms. The Match-Extend algorithm introduced here requires fewer assumptions and sticks tighter to the attested typology. The algorithm also contains no parameter or constraint specific to individual graphs or topologies, unlike previous proposals. Match-Extend requires nothing except knowing the last added set of links.

Attention Calibration for Transformer in Neural Machine Translation

Downloads

Next from ACL 2021

Neural OCR Post-Hoc Correction of Historical Corpora

Similar lecture

Encouraging Neural Machine Translation to Satisfy Terminology Constraints

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES