Spain

In this paper, we described our systems for the first and second subtasks of Social Media Mining for Health Applications (SMM4H) shared task in 2020. The two subtasks are automatic classi-fication of medication mentions and adverse effect in tweets. Our systems for both subtasks are based on Robustly optimized BERT approach (RoBERTa) and our previous work at SMM4H’19. The best F1-scores achieved by our systems for subtask 1 and 2 were 0.7974 and 0.64 respec-tively, which outperformed the average F1-scores among all teams’ best runs by at least 0.13.

COLING 2020

ISLab System for SMM4H Shared Task 2020

stages 

COLING, the International Conference on Computational Linguistics, is one of the premier conferences for natural language processing and computational linguistics. Often grouped within the field of artificial intelligence, but actually pre-dating the development of artificial intelligence, advances in computational linguistics and natural language processing are now some of the major drivers behind the use of artificial intelligence for commercial and social applications – for example, on-line search, machine translation and with voice-assisted conversational devices.

First established in 1965, the biennial COLING conference is held in diverse parts of the globe and attracts participants from both top-ranked research centers and emerging countries. Today, the most important developments in our field are taking place not only in universities and academic research institutes, but also in industrial research departments and in technological startups. COLING conferences provide opportunities for all these communities to showcase their exciting developments.

COLING, the International Conference on Computational Linguistics, is one of the premier conferences for natural language processing and computational linguistics. Often grouped within the field of artificial intelligence, but actually pre-dating the development of artificial intelligence, advances in computational linguistics and natural language processing are now some of the major drivers behind the use of artificial intelligence for commercial and social applications – for example, on-line search, machine translation and with voice-assisted conversational devices.

technical paper

This presentation describes neural models developed for the Social Media Mining for Health (SMM4H) 2020 shared tasks. Specifically, we participated in two tasks. We investigate the use of a language representation model BERT pre-trained on a large-scale corpus of 5 million health-related user reviews in English and Russian. The ensemble of neural networks for extraction and normalization of adverse drug reactions ranked first among 7 teams at the SMM4H 2020 Task 3 and obtained a relaxed F1 of 46%. The BERT-based multilingual model for classification of English and Russian tweets that report adverse reactions ranked second among 16 and 7 teams at two first subtasks of the SMM4H 2019 Task 2 and obtained a relaxed F1 of 58% on English tweets and 51% on Russian tweets.

KFU NLP Team at SMM4H 2020 Tasks: Cross-lingual Transfer Learning with Pretrained Language Models for Drug Reactions

Our team participated in the second subtask for Russian language creating
a system to detect adverse drug reaction (ADR) presence in a text. For our submission, we exploited an ensemble model architecture, combining BERT’s extension for Russian language, Logistic Regression and domain-specific preprocessing pipeline. Our system was ranked first among others, achieving F-score of 0.51.

BERT implementation for detecting adverse drug effects mentions in Russian

The vast amount of data on social media presents significant opportunities and challenges for utilizing it as a resource for health informatics. The fifth iteration of the Social Media Mining for Health Applications (#SMM4H) shared tasks sought to advance the use of Twitter data (tweets) for pharmacovigilance, toxicovigilance, and epidemiology of birth defects. In addition to re-runs of three tasks, #SMM4H 2020 included new tasks for detecting adverse effects of medications in French and Russian tweets, characterizing chatter related to prescription medication abuse, and detecting self reports of birth defect pregnancy outcomes. The five tasks required methods for binary classification, multi-class classification, and named entity recognition (NER). With 29 teams and a total of 130 system submissions, participation in the #SMM4H shared tasks continues to grow.

Overview of the #SMM4H 2020 Shared Tasks

Prior work in Argument Mining frequently alludes to its potential applications in automatic de-bating systems. Despite this focus, almost no datasets or models exist which apply natural language processing techniques to problems found within competitive formal debate. To remedy this, we present the DebateSum dataset1. DebateSum consists of 187,386 unique pieces of evidence with corresponding argument and extractive summaries. DebateSum was made using data compiled by competitors within the National Speech and Debate Association over a 7-year period. We train several transformer summarization models to benchmark summarization performance on DebateSum. We also introduce a set of fasttext word-vectors trained on DebateSum called debate2vec. Finally, we present a search engine for this dataset which is utilized extensively by members of the National Speech and Debate Association today. The DebateSum search engine is available to the public here: http://www.debate.cards

DebateSum: A large-scale argument mining and summarization dataset

Habernal and Gurevych (2016) in their feasibility study of Arfgument Mining on User Generated Content (UGC) indicated the major challenge facing the field: there seems to be a big gap between the substantial argumentation research and the limited use NLP applications seem to make of these. This also leads to a myriad of annotation approaches, which either start from scratch (with a proprietory scheme which is rarely motivated in terms of the theories underlying it) or apply an existing scheme (Stephen Toulmin being a popular choice for many different text genres, ranging from more formal texts (legal documents) to more informal data like forum posts). The pilot study into the annotation of Dutch Facebook comments (for topics, aspects, stance, argumentativeness and claim) we present was motivated by our hypothesis that the divide between the unpredictable nature of the data and the traditionally restrictive definitions of concepts like argumentativeness may be bridged by gaining more knowledge on the great variety of linguistic expressions of argumentation in user comments. We motivate some of the theoretical principles underlying our annotation guidelines as well as illustrate some of the annotation difficulties which have prompted the adjustments of our initial guidelines.

Annotating topics, stance, argumentativeness and claims in Dutch social media comments: A pilot study

Debate portals and similar web platforms constitute one of the main text sources in computational argumentation research and its applications. While the corpora built upon these sources are rich of argumentatively relevant content and structure, they also include text that is irrelevant, or even detrimental, to their purpose. In this paper, we present a precision-oriented approach to detecting such irrelevant text in a semi-supervised way. Given a few seed examples, the approach automatically learns basic lexical patterns of relevance and irrelevance and then incrementally bootstraps new patterns from sentences matching the patterns. In the existing args.me corpus with 400k argumentative texts, our approach detects almost 87k irrelevant sentences, at a precision of 0.97 according to manual evaluation. With low effort, the approach can be adapted to other web argument corpora, providing a generic way to improve corpus quality.

Semi-Supervised Cleansing of Web Argument Corpora

Sentiment and stance are two important concepts for the analysis of arguments. We propose to add another perspective to the analysis, namely moral sentiment. We argue that moral values are crucial for ideological debates and can thus add useful information for argument mining. In the paper, we present different models for automatically predicting moral sentiment in debates and evaluate them on a manually annotated test set. We then apply our models to investigate how moral values in arguments relate to argument quality, stance, and audience reactions.

Exploring Morality in Argumentation

Computational Argumentation in general and Argument Mining in particular are important research fields. In previous works, many of the challenges to automatically extract and to some degree reason over natural language arguments were addressed. The tools to extract argument units are increasingly available and further open problems can be addressed. In this work, we are presenting the task of Aspect-Based Argument Mining (ABAM), with the essential subtasks of Aspect Term Extraction (ATE) and Nested Segmentation (NS). At the first instance, we create and release an annotated corpus with aspect information on the token-level. We consider aspects as the main point(s) argument units are addressing. This information is important for further downstream tasks such as argument ranking, argument summarization and generation, as well as the search for counter-arguments on the aspect-level. We present several experiments using state-of-the-art supervised architectures and demonstrate their performance for both of the subtasks. The annotated benchmark is available at https://github.com/trtm/ABAM.

Aspect-Based Argument Mining

Notwithstanding the increasing role Twitter plays in modern political and social discourse, resources built for conducting argument mining on tweets remain limited. In this paper, we present a new corpus of German tweets annotated for argument components. To the best of our knowledge, this is the first corpus containing not only annotated full tweets but also argumentative spans within tweets. We further report first promising results using supervised classification (F1: 0.82) and sequence labeling (F1: 0.72) approaches.

Annotation and Detection of Arguments in Tweets

In this paper, we publicly release an annotated corpus of 42 decisions of the European Court of Human Rights (ECHR). The corpus is annotated in terms of three types of clauses useful in argument mining: premise, conclusion, and non-argument parts of the text. Furthermore, relationships among the premises and conclusions are mapped. We present baselines for three tasks that lead from unstructured texts to structured arguments. The tasks are argument clause recognition, clause relation prediction, and premise/conclusion recognition. Despite a straightforward application of the bidirectional encoders from Transformers (BERT), we obtained very promising results F1 0.765 on argument recognition, 0.511 on relation prediction, and 0.859/0.628 on premise/conclusion recognition). The results suggest the usefulness of pre-trained language models based on deep neural network architectures in argument mining. Because of the simplicity of the baselines, there is ample space for improvement in future work based on the released corpus.

Downloads

Next from COLING 2020

KFU NLP Team at SMM4H 2020 Tasks: Cross-lingual Transfer Learning with Pretrained Language Models for Drug Reactions

Similar lecture

KFU NLP Team at SMM4H 2020 Tasks: Cross-lingual Transfer Learning with Pretrained Language Models for Drug Reactions

.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 COLING 2020

KFU NLP Team at SMM4H 2020 Tasks: Cross-lingual Transfer Learning with Pretrained Language Models for Drug Reactions

Similar lecture

KFU NLP Team at SMM4H 2020 Tasks: Cross-lingual Transfer Learning with Pretrained Language Models for Drug Reactions

Downloads