Search

While pretrained language models (LMs) have driven impressive gains over morpho-syntactic and semantic tasks, their ability to model discourse and pragmatic phenomena is less clear. As a step towards a better understanding of their discourse modeling capabilities, we propose a sentence intrusion detection task. We examine the performance of a broad range of pretrained LMs on this detection task for English. Lacking a dataset for the task, we introduce INSteD, a novel intruder sentence detection dataset, containing 170,000+ documents constructed from English Wikipedia and CNN news articles. Our experiments show that pretrained LMs perform impressively in in-domain evaluation, but experience a substantial drop in the cross-domain setting, indicating limited generalization capacity. Further results over a novel linguistic probe dataset show that there is substantial room for improvement, especially in the cross- domain setting.

The perceived quality of a document is affected by various factors, including grammat- icality, readability, stylistics, and expertise depth, making the task of document quality assessment a complex one. In this paper, we explore this task in the context of assessing the quality of Wikipedia articles and academic papers. Observing that the visual rendering of a document can capture implicit quality indicators that are not present in the document text — such as images, font choices, and visual layout — we propose a joint model that combines the text content with a visual rendering of the document for document qual- ity assessment. Our joint model achieves state-of-the-art results over five datasets in two domains (Wikipedia and academic papers), which demonstrates the complementarity of textual and visual features, and the general applicability of our model. To examine what kinds of features our model has learned, we further train our model in a multi-task learning setting, where document quality assessment is the primary task and feature learning is an auxiliary task. Experimental results show that visual embeddings are better at learning structural features while textual embeddings are better at learning readability scores, which further verifies the complementarity of visual and textual features.

While automatic summarization evaluation methods developed for English are routinely applied to other languages, this is the first attempt to systematically quantify their panlinguistic efficacy. We take a summarization corpus for eight different languages, and manually annotate generated summaries for focus (precision) and coverage (recall). Based on this, we evaluate 19 summarization evaluation metrics, and find that using multilingual BERT within BERTScore performs well across all languages, at a level above that for English., Findings of ACL 2021

Existing work on probing of pretrained language models (LMs) has predominantly focused on sentence-level syntactic tasks. In this paper, we introduce document-level discourse probing to evaluate the ability of pretrained LMs to capture document-level relations. We experiment with 7 pretrained LMs, 4 languages, and 7 discourse probing tasks, and find BART to be overall the best model at capturing discourse -- but only in its encoder, with BERT performing surprisingly well as the baseline model. Across the different models, there are substantial differences in which layers best capture discourse information, and large disparities between models., Accepted at NAACL 2021

Chemical patents represent a valuable source of information about new chemical compounds, which is critical to the drug discovery process. Automated information extraction over chemical patents is, however, a challenging task due to the large volume of existing patents and the complex linguistic properties of chemical patents. The Cheminformatics Elsevier Melbourne University (ChEMU) evaluation lab 2020, part of the Conference and Labs of the Evaluation Forum 2020 (CLEF2020), was introduced to support the development of advanced text mining techniques for chemical patents. The ChEMU 2020 lab proposed two fundamental information extraction tasks focusing on chemical reaction processes described in chemical patents: (1)chemical named entity recognition, requiring identification of essential chemical entities and their roles in chemical reactions, as well as reaction conditions; and (2)event extraction, which aims at identification of event steps relating the entities involved in chemical reactions. The ChEMU 2020 lab received 37 team registrations and 46 runs. Overall, the performance of submissions for these tasks exceeded our expectations, with the top systems outperforming strong baselines. We further show the methods to be robust to variations in sampling of the test data. We provide a detailed overview of the ChEMU 2020 corpus and its annotation, showing that inter-annotator agreement is very strong. We also present the methods adopted by participants, provide a detailed analysis of their performance, and carefully consider the potential impact of data leakage on interpretation of the results. The ChEMU 2020 Lab has shown the viability of automated methods to support information extraction of key information in chemical patents.

Chemical patents serve as an indispensable source of information about new discoveries of chemical compounds. The ChEMU (Cheminformatics Elsevier Melbourne University) lab addresses information extraction over chemical patents, and aims to advance the state of the art on this topic. ChEMU lab 2021, as part of the 12th Conference and Labs of the Evaluation Forum (CLEF-2021), will be the second ChEMU lab. ChEMU 2021 will provide two distinct tasks related to reference resolution in chemical patents. Task 1—Chemical Reaction Reference Resolution—focuses on paragraph-level references and aims to identify the chemical reactions or general conditions specified in one reaction description referred to by another. Task 2—Anaphora Resolution—focuses on expression-level references and aims to identify the reference relationships between expressions in chemical reaction descriptions. In this paper, we introduce ChEMU 2021, including its motivation, goals, tasks, resources, and evaluation framework.

In this paper, we provide an overview of the Cheminformatics Elsevier Melbourne University (ChEMU) evaluation lab 2021, part of the Conference and Labs of the Evaluation Forum 2021 (CLEF 2021). The ChEMU evaluation lab focuses on information extraction over chemical reactions from patent texts. As the second instance of our ChEMU lab series, we build upon the ChEMU corpus developed for ChEMU 2020, extending it for two distinct tasks related to reference resolution in chemical patents. Task 1—Chemical Reaction Reference Resolution—focuses on paragraph-level references and aims to identify the chemical reactions or general conditions specified in one reaction description referred to by another. Task 2—Anaphora Resolution—focuses on expression-level references and aims to identify the reference relationships between expressions in chemical reaction descriptions. Herein, we describe the resources created for these tasks and the evaluation methodology adopted. We also provide a brief summary of the results obtained in this lab, finding that one submission achieves substantially better results than our baseline models.

Chemical patents contain rich coreference and bridging links, which are the target of this research. Specially, we introduce a novel annotation scheme, based on which we create the ChEMU-Ref dataset from reaction description snippets in English-language chemical patents. We propose a neural approach to anaphora resolution, which we show to achieve strong results, especially when jointly trained over coreference and bridging links.

We introduce a top-down approach to discourse parsing that is conceptually simpler than its predecessors (Kobayashi et al., 2020; Zhang et al., 2020). By framing the task as a sequence labelling problem where the goal is to iteratively segment a document into individual discourse units, we are able to eliminate the decoder and reduce the search space for splitting points. We explore both traditional recurrent models and modern pre-trained transformer models for the task, and additionally introduce a novel dynamic oracle for top-down parsing. Based on the Full metric, our proposed LSTM model sets a new state-of-the-art for RST parsing., Accepted at EACL 2021

Images are core components of multi-modal learning in natural language processing (NLP), and results have varied substantially as to whether images improve NLP tasks or not. One confounding effect has been that previous NLP research has generally focused on sophisticated tasks (in varying settings), generally applied to English only. We focus on text classification, in the context of assigning named entity classes to a given Wikipedia page, where images generally complement the text and the Wikipedia page can be in one of a number of different languages. Our experiments across a range of languages show that images complement NLP models (including BERT) trained without external pre-training, but when combined with BERT models pre-trained on large-scale external data, images contribute nothing.

Adversarial learning can learn fairer and less biased models of language than standard methods. However, current adversarial techniques only partially mitigate model bias, added to which their training procedures are often unstable. In this paper, we propose a novel approach to adversarial learning based on the use of multiple diverse discriminators, whereby discriminators are encouraged to learn orthogonal hidden representations from one another. Experimental results show that our method substantially improves over standard adversarial removal methods, in terms of reducing bias and the stability of training., Comment: EACL 2021 (5 pages + 1 references)

Although the Indonesian language is spoken by almost 200 million people and the 10th most spoken language in the world, it is under-represented in NLP research. Previous work on Indonesian has been hampered by a lack of annotated datasets, a sparsity of language resources, and a lack of resource standardization. In this work, we release the IndoLEM dataset comprising seven tasks for the Indonesian language, spanning morpho-syntax, semantics, and discourse. We additionally release IndoBERT, a new pre-trained language model for Indonesian, and evaluate it over IndoLEM, in addition to benchmarking it against existing resources. Our experiments show that IndoBERT achieves state-of-the-art performance over most of the tasks in IndoLEM., Accepted at COLING 2020 - The 28th International Conference on Computational Linguistics

Existing metonymy resolution approaches rely on features extracted from external resources like dictionaries and hand-crafted lexical resources. In this paper, we propose an end-to-end word-level classification approach based only on BERT, without dependencies on taggers, parsers, curated dictionaries of place names, or other external resources. We show that our approach achieves the state-of-the-art on 5 datasets, surpassing conventional BERT models and benchmarks by a large margin. We also show that our approach generalises well to unseen data., 12 pages

As part of growing NLP capabilities, coupled with an awareness of the ethical dimensions of research, questions have been raised about whether particular datasets and tasks should be deemed off-limits for NLP research. We examine this question with respect to a paper on automatic legal sentencing from EMNLP 2019 which was a source of some debate, in asking whether the paper should have been allowed to be published, who should have been charged with making such a decision, and on what basis. We focus in particular on the role of data statements in ethically assessing research, but also discuss the topic of dual use, and examine the outcomes of similar debates in other scientific disciplines., 6 pages; accepted for ACL2020

We introduce a new evaluation lab named ChEMU (Cheminformatics Elsevier Melbourne University), part of the 11th Conference and Labs of the Evaluation Forum (CLEF-2020). ChEMU involves two key information extraction tasks over chemical reactions from patents. Task 1—Named entity recognition—involves identifying chemical compounds as well as their types in context, i.e., to assign the label of a chemical compound according to the role which the compound plays within a chemical reaction. Task 2—Event extraction over chemical reactions—involves event trigger detection and argument recognition. We briefly present the motivations and goals of the ChEMU tasks, as well as resources and evaluation methodology.

Identifying the reasons for antibiotic administration in veterinary records is a critical component of understanding antimicrobial usage patterns. This informs antimicrobial stewardship programs designed to fight antimicrobial resistance, a major health crisis affecting both humans and animals in which veterinarians have an important role to play. We propose a document classification approach to determine the reason for administration of a given drug, with particular focus on domain adaptation from one drug to another, and instance selection to minimize annotation effort.

In this paper, we provide an overview of the Cheminformatics Elsevier Melbourne University (ChEMU) evaluation lab 2020, part of the Conference and Labs of the Evaluation Forum 2020 (CLEF2020). The ChEMU evaluation lab focuses on information extraction over chemical reactions from patent texts. Using the ChEMU corpus of 1500 “snippets” (text segments) sampled from 170 patent documents and annotated by chemical experts, we defined two key information extraction tasks. Task 1 addresses chemical named entity recognition, the identification of chemical compounds and their specific roles in chemical reactions. Task 2 focuses on event extraction, the identification of reaction steps, relating the chemical compounds involved in a chemical reaction. Herein, we describe the resources created for these tasks and the evaluation methodology adopted. We also provide a brief summary of the participants of this lab and the results obtained across 46 runs from 11 teams, finding that several submissions achieve substantially better results than our baseline methods.

In this paper, we apply pre-trained language models to the Semantic Textual Similarity (STS) task, with a specific focus on the clinical domain. In low-resource setting of clinical STS, these large models tend to be impractical and prone to overfitting. Building on BERT, we study the impact of a number of model design choices, namely different fine-tuning and pooling strategies. We observe that the impact of domain-specific fine-tuning on clinical STS is much less than that in the general domain, likely due to the concept richness of the domain. Based on this, we propose two data augmentation techniques. Experimental results on N2C2-STS 1 demonstrate substantial improvements, validating the utility of the proposed methods.

Domain pretraining followed by task fine-tuning has become the standard paradigm for NLP tasks, but requires in-domain labelled data for task fine-tuning. To overcome this, we propose to utilise domain unlabelled data by assigning pseudo labels from a general model. We evaluate the approach on two clinical STS datasets, and achieve r= 0.80 on N2C2-STS. Further investigation reveals that if the data distribution of unlabelled sentence pairs is closer to the test data, we can obtain better performance. By leveraging a large general-purpose STS dataset and small-scale in-domain training data, we obtain further improvements to r= 0.90, a new SOTA.

Background Currently there is an incomplete understanding of antimicrobial usage patterns in veterinary clinics in Australia, but such knowledge is critical for the successful implementation and monitoring of antimicrobial stewardship programs. Methods VetCompass Australia collects medical records from 181 clinics in Australia (as of May 2018). These records contain detailed information from individual consultations regarding the medications dispensed. One unique aspect of VetCompass Australia is its focus on applying natural language processing (NLP) and machine learning techniques to analyse the records, similar to efforts conducted in other medical studies. Results The free text fields of 4,394,493 veterinary consultation records of dogs and cats between 2013 and 2018 were collated by VetCompass Australia and NLP techniques applied to enable the querying of the antimicrobial usage within these consultations. Conclusion The NLP algorithms developed matched antimicrobial in clinical records with 96.7% accuracy and an F1 Score of 0.85, as evaluated relative to expert annotations. This dataset can be readily queried to demonstrate the antimicrobial usage patterns of companion animal practices throughout Australia.

We present a new model for acquiring comprehensive multiword lexicons from large corpora based on competition among n-gram candidates. In contrast to the standard approach of simple ranking by association measure, in our model n-grams are arranged in a lattice structure based on subsumption and overlap relationships, with nodes inhibiting other nodes in their vicinity when they are selected as a lexical item. We show how the configuration of such a lattice can be optimized tractably, and demonstrate using annotations of sampled n-grams that our method consistently outperforms alternatives by at least 0.05 F-score across several corpora and languages.

We describe SemEval-2017 Task 3 on Community Question Answering. This year, we reran the four subtasks from SemEval-2016:(A) Question-Comment Similarity,(B) Question-Question Similarity,(C) Question-External Comment Similarity, and (D) Rerank the correct answers for a new question in Arabic, providing all the data from 2015 and 2016 for training, and fresh data for testing. Additionally, we added a new subtask E in order to enable experimentation with Multi-domain Question Duplicate Detection in a larger-scale scenario, using StackExchange subforums. A total of 23 teams participated in the task, and submitted a total of 85 runs (36 primary and 49 contrastive) for subtasks A-D. Unfortunately, no teams participated in subtask E. A variety of approaches and features were used by the participating systems to address the different subtasks. The best systems achieved an official score (MAP) of 88.43, 47.22, 15.46, and 61.16 in subtasks A, B, C, and D, respectively. These scores are better than the baselines, especially for subtasks A-C., Comment: community question answering, question-question similarity, question-comment similarity, answer reranking, Multi-domain Question Duplicate Detection, StackExchange, English, Arabic

In this paper, we apply various embedding methods on multiword expressions to study how well they capture the nuances of non-compositional data. Our results from a pool of word-, character-, and document-level embbedings suggest that Word2vec performs the best, followed by FastText and Infersent. Moreover, we find that recently-proposed contextualised embedding models such as Bert and ELMo are not adept at handling non-compositionality in multiword expressions.

Argument component extraction is a challenging and complex high-level semantic extraction task. As such, it is both expensive to annotate (meaning training data is limited and low-resource by nature), and hard for current-generation deep learning methods to model. In this paper, we reevaluate the performance of state-of-the-art approaches in both single- and multi-task learning settings using combinations of character-level, GloVe, ELMo, and BERT encodings using standard BiLSTM-CRF encoders. We use evaluation metrics that are more consistent with evaluation practice in named entity recognition to understand how well current baselines address this challenge and compare their performance to lower-level semantic tasks such as CoNLL named entity recognition. We find that performance utilizing various pre-trained representations and training methodologies often leaves a lot to be desired as it currently stands, and suggest future pathways for improvement.

Accurate, automatic evaluation of machine translation is critical for system tuning, and evaluating progress in the field. We proposed a simple unsupervised metric, and additional supervised metrics which rely on contextual word embeddings to encode the translation and reference sentences. We find that these models rival or surpass all existing metrics in the WMT 2017 sentence-level and system-level tracks, and our trained model has a substantially higher correlation with human judgements than all existing metrics on the WMT 2017 to-English sentence level dataset.

Critical to natural language generation is the production of correctly inflected text. In this paper, we isolate the task of predicting a fully inflected sentence from its partially lemmatized version. Unlike traditional morphological inflection or surface realization, our task input does not provide ``gold'' tags that specify what morphological features to realize on each lemmatized word; rather, such features must be inferred from sentential context. We develop a neural hybrid graphical model that explicitly reconstructs morphological features before predicting the inflected forms, and compare this to a system that directly predicts the inflected forms without relying on any morphological annotation. We experiment on several typologically diverse languages from the Universal Dependencies treebanks, showing the utility of incorporating linguistically-motivated latent variables into NLP models., Comment: NAACL 2019

Supervised models of NLP rely on large collections of text which closely resemble the intended testing setting. Unfortunately matching text is often not available in sufficient quantity, and moreover, within any domain of text, data is often highly heterogenous. In this paper we propose a method to distill the important domain signal as part of a multi-domain learning system, using a latent variable model in which parts of a neural model are stochastically gated based on the inferred domain. We compare the use of discrete versus continuous latent variables, operating in a domain-supervised or a domain semi-supervised setting, where the domain is known only for a subset of training inputs. We show that our model leads to substantial performance improvements over competitive benchmark domain adaptation methods, including methods using adversarial learning., ACL 2019 (9 pages + 2 references + 1 appendices)

In this paper, we present the first attempt to use multitask learning for query segmentation. We use the semantic category of the words as an auxiliary task and show that segmentation improves when the model is also trained to predict the semantic category of the query terms, outperforming benchmark methods over a novel dataset from a popular job search engine. Our further experiments show that the task of modeling the query term semantics performs better as a standalone task, without adding segmentation as an auxiliary task.

While neural networks have been shown to achieve impressive results for sentence-level sentiment analysis, targeted aspect-based sentiment analysis (TABSA) --- extraction of fine-grained opinion polarity w.r.t. a pre-defined set of aspects --- remains a difficult task. Motivated by recent advances in memory-augmented models for machine reading, we propose a novel architecture, utilising external "memory chains" with a delayed memory update mechanism to track entities. On a TABSA task, the proposed model demonstrates substantial improvements over state-of-the-art approaches, including those using external knowledge bases., Accepted to NAACL 2018 (camera-ready)

Social media user geolocation is vital to many applications such as event detection. In this paper, we propose GCN, a multiview geolocation model based on Graph Convolutional Networks, that uses both text and network context. We compare GCN to the state-of-the-art, and to two baselines we propose, and show that our model achieves or is competitive with the state- of-the-art over three benchmark geolocation datasets when sufficient supervision is available. We also evaluate GCN under a minimal supervision scenario, and show it outperforms baselines. We find that highway network gates are essential for controlling the amount of useful neighbourhood expansion in GCN., Comment: ACL2018

Most real world language problems require learning from heterogenous corpora, raising the problem of learning robust models which generalise well to both similar (in domain) and dissimilar (out of domain) instances to those seen in training. This requires learning an underlying task, while not learning irrelevant signals and biases specific to individual domains. We propose a novel method to optimise both in- and out-of-domain accuracy based on joint learning of a structured neural model with domain-specific and domain-general components, coupled with adversarial training for domain. Evaluating on multi-domain language identification and multi-domain sentiment analysis, we show substantial improvements over standard domain adaptation techniques, and domain-adversarial training.

This paper describes a warrant classification system for SemEval 2018 Task 12, that attempts to learn semantic representations of reasons, claims and warrants. The system consists of 3 stacked LSTMs: one for the reason, one for the claim, and one shared Siamese Network for the 2 candidate warrants. Our main contribution is to force the embeddings into a shared feature space using vector operations, semantic similarity classification, Siamese networks, and multi-task learning. In doing so, we learn a form of generative implication, in encoding implication interrelationships between reasons, claims, and the associated correct and incorrect warrants. We augment the limited data in the task further by utilizing WordNet synonym “fuzzing”. When applied to SemEval 2018 Task 12, our system performs well on the development data, and officially ranked 8th among 21 teams.

Story comprehension requires a deep semantic understanding of the narrative, making it a challenging task. Inspired by previous studies on ROC Story Cloze Test, we propose a novel method, tracking various semantic aspects with external neural memory chains while encouraging each to focus on a particular semantic aspect. Evaluated on the task of story ending prediction, our model demonstrates superior performance to a collection of competitive baselines, setting a new state of the art., Comment: Accepted to ACL 2018 (camera-ready)

Crowd-sourced assessments of machine translation quality allow evaluations to be carried out cheaply and on a large scale. It is essential, however, that the crowd's work be filtered to avoid contamination of results through the inclusion of false assessments. One method is to filter via agreement with experts, but even amongst experts agreement levels may not be high. In this paper, we present a new methodology for crowd-sourcing human assessments of translation quality, which allows individual workers to develop their own individual assessment strategy. Agreement with experts is no longer required, and a worker is deemed reliable if they are consistent relative to their own previous work. Individual translations are assessed in isolation from all others in the form of direct estimates of translation quality. This allows more meaningful statistics to be computed for systems and enables significance to be determined on smaller sets of assessments. We demonstrate the methodology's feasibility in large-scale human evaluation through replication of the human evaluation component of Workshop on Statistical Machine Translation shared translation task for two language pairs, Spanish-to-English and English-to-Spanish. Results for measurement based solely on crowd-sourced assessments show system rankings in line with those of the original evaluation. Comparison of results produced by the relative preference approach and the direct estimate method described here demonstrate that the direct estimate method has a substantially increased ability to identify significant differences between translation systems.

The development of precision grammars is an inherently resource-intensive process; their complexity means that changes made to one area of a grammar often introduce unexpected flow-on effects elsewhere in the grammar which may only be discovered after some time has been invested in updating numerous test suite items. In this paper, we present the browser-based gDelta tool, which aims to provide grammar engineers with more immediate feedback on the impact of changes made to a grammar by comparing parser output from two different grammar versions. We describe an attribute weighting algorithm for highlighting components of the grammar that have been strongly impacted by a modification to the grammar, as well as a technique for clustering test suite items whose parsability has changed, in order to locate related groups of effects. These two techniques are used to present the grammar engineer with different views on the grammar to inform them of different aspects of change in a data-driven manner.

Language identification is the task of automatically detecting the language(s) present in a document based on the content of the document. In this work, we address the problem of detecting documents that contain text from more than one language ( multilingual documents). We introduce a method that is able to detect that a document is multilingual, identify the languages present, and estimate their relative proportions. We demonstrate the effectiveness of our method over synthetic data, as well as real-world multilingual documents collected from the web.

A person or other entity is often associated with multiple URL endpoints on the web, motivating the task of determining whether a given pair of webpages is coreferent to a given entity. To strike a balance between unsupervised and supervised methods that require annotated data, we build a positive and unlabelled (PU) learning model, where we obtain positive examples using web search-based distant supervision. We evaluate our proposed approach using the SemEval-2007 WePS and ALTA-2016 shared task datasets.

ive document summarization seeks to automatically generate a summary for a document, based on some abstract “understanding” of the original document. State-of-the-art techniques traditionally use attentive encoder–decoder architectures. However, due to the large number of parameters in these models, they require large training datasets and long training times. In this paper, we propose decoupling the encoder and decoder networks, and training them separately. We encode documents using an unsupervised document encoder, and then feed the document vector to a recurrent neural network decoder. With this decoupled architecture, we decrease the number of parameters in the decoder substantially, and shorten its training time. Experiments show that the decoupled model achieves comparable performance with state-of-the-art models for in-domain documents, but less well for out-of-domain documents.

This paper presents a methodology for identifying and resolving various kinds of inconsistency in the context of merging dependency and multiword expression (MWE) annotations, to generate a dependency treebank with comprehensive MWE annotations. Candidates for correction are identified using a variety of heuristics, including an entirely novel one which identifies violations of MWE constituency in the dependency tree, and resolved by arbitration with minimal human intervention. Using this technique, we identified and corrected several hundred errors across both parse and MWE annotations, representing changes to a significant percentage (well over 10%) of the MWE instances in the joint corpus.

In East Asian languages such as Japanese and Chinese, the semantics of a character are (somewhat) reflected in its sub-character elements. This paper examines the effect of using sub-characters for language modeling in Japanese. This is achieved by decomposing characters according to a range of character decomposition datasets, and training a neural language model over variously decomposed character representations. Our results indicate that language modelling can be improved through the inclusion of sub-characters, though this result depends on a good choice of decomposition dataset and the appropriate granularity of decomposition.

Monolingual evaluation of Machine Translation (MT) aims to simplify human assessment by requiring assessors to compare the meaning of the MT output with a reference translation, opening up the task to a much larger pool of genuinely qualified evaluators. Monolingual evaluation runs the risk, however, of bias in favour of MT systems that happen to produce translations superficially similar to the reference and, consistent with this intuition, previous investigations have concluded monolingual assessment to be strongly biased in this respect. On re-examination of past analyses, we identify a series of potential analytical errors that force some important questions to be raised about the reliability of past conclusions, however. We subsequently carry out further investigation into reference bias via direct human assessment of MT adequacy via quality controlled crowd-sourcing. Contrary to both intuition and past conclusions, results for show no significant evidence of reference bias in monolingual evaluation of MT.

Deep neural networks have achieved remarkable results across many language processing tasks, however they have been shown to be susceptible to overfitting and highly sensitive to noise, including adversarial attacks. In this work, we propose a linguistically-motivated approach for training robust models based on exposing the model to corrupted text examples at training time. We consider several flavours of linguistically plausible corruption, include lexical semantic and syntactic methods. Empirically, we evaluate our method with a convolutional neural model across a range of sentiment analysis datasets. Compared with a baseline and the dropout method, our method achieves better overall performance.

Manual data annotation is a vital component of NLP research. When designing annotation tasks, properties of the annotation interface can unintentionally lead to artefacts in the resulting dataset, biasing the evaluation. In this paper, we explore sequence effects where annotations of an item are affected by the preceding items. Having assigned one label to an instance, the annotator may be less (or more) likely to assign the same label to the next. During rating tasks, seeing a low quality item may affect the score given to the next item either positively or negatively. We see clear evidence of both types of effects using auto-correlation studies over three different crowdsourced datasets. We then recommend a simple way to minimise sequence effects.

Derivational morphology is a fundamental and complex characteristic of language. In this paper we propose the new task of predicting the derivational form of a given base-form lemma that is appropriate for a given context. We present an encoder--decoder style neural network to produce a derived form character-by-character, based on its corresponding character-level representation of the base form and the context. We demonstrate that our model is able to generate valid context-sensitive derivations from known base forms, but is less accurate under a lexicon agnostic setting.

© 2017 Association for Computational Linguistics. Meaningful conclusions about the relative performance of NLP systems are only possible if the gold standard employed in a given evaluation is both valid and reliable. In this paper, we explore the validity of human annotations currently employed in the evaluation of document-level quality estimation for machine translation (MT).We demonstrate the degree to which MT system rankings are dependent on weights employed in the construction of the gold standard, before proposing direct human assessment as a valid alternative. Experiments show direct assessment (DA) scores for documents to be highly reliable, achieving a correlation of above 0.9 in a self-replication experiment, in addition to a substantial estimated cost reduction through quality controlled crowdsourcing. The original gold standard based on post-edits incurs a 10-20 times greater cost than DA.

We propose a simple yet effective text- based user geolocation model based on a neural network with one hidden layer, which achieves state of the art performance over three Twitter benchmark geolocation datasets, in addition to producing word and phrase embeddings in the hidden layer that we show to be useful for detecting dialectal terms. As part of our analysis of dialectal terms, we release DAREDS, a dataset for evaluating dialect term detection methods.

Language models are typically applied at the sentence level, without access to the broader document context. We present a neural language model that incorporates document context in the form of a topic model-like architecture, thus providing a succinct representation of the broader document context outside of the current sentence. Experiments over a range of datasets demonstrate that our model outperforms a pure sentence-based model in terms of language model perplexity, and leads to topics that are potentially more coherent than those produced by a standard LDA topic model. Our model also has the ability to generate related sentences for a topic, providing another way to interpret topics., Comment: 11 pages, Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (ACL 2017) (to appear)

Geographical location is vital to geospatial applications like local search and event detection. In this paper, we investigate and improve on the task of text-based geolocation prediction of Twitter users. Previous studies on this topic have typically assumed that geographical references (e.g., gazetteer terms, dialectal words) in a text are indicative of its author’s location. However, these references are often buried in informal, ungrammatical, and multilingual data, and are therefore non-trivial to identify and exploit. We present an integrated geolocation prediction framework and investigate what factors impact on prediction accuracy. First, we evaluate a range of feature selection methods to obtain “location indicative words”. We then evaluate the impact of non-geotagged tweets, language, and user-declared metadata on geolocation prediction. In addition, we evaluate the impact of temporal variance on model generalisation, and discuss how users differ in terms of their geolocatability. We achieve state-of-the-art results for the text-based Twitter user geolocation task, and also provide the most extensive exploration of the task to date. Our findings provide valuable insights into the design of robust, practical text-based geolocation prediction systems.

Corpora of documents whose metadata includes GPS coordinates have recently become widely available through online social media such as Twitter. This has created opportunities for statistical corpus methods that describe the geographical spread of words, but such techniques do not appear to be widely used in corpus linguistics and lexicography. This paper presents several methods for describing the spread of a set of points, corresponding to documents containing a given word and applies the methods to a corpus of GPS-tagged tweets from Twitter. In experiments on known regionalisms, we show that these methods could be used to help identify such expressions. We analyze the words in the corpus identified as having the most geographically restricted usage and identify some expressions that appear to be previously undocumented regionalisms with highly localized usage.

In this article, we investigate word sense distributions in noun compounds (NCs). Our primary goal is to disambiguate the word sense of component words in NCs, based on investigation of “semantic collocation” between them. We use sense collocation and lexical substitution to build supervised and unsupervised word sense disambiguation (WSD) classifiers, and show our unsupervised learner to be superior to a benchmark WSD system. Further, we develop a word sense-based approach to interpreting the semantic relations in NCs.