Yes, we can! Mining Arguments in 50 Years of US Presidential Campaign Debates

Haddadan, Shohreh; Cabrio, Elena; Villata, Serena

doi:10.18653/v1/p19-1463

Cited by 29 publications

(26 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Here, it is important to make emphasis that the way we considered to represent argumentative relations (i.e., IAT labelling) make this task harder than most of the previous work (i.e., attack/support) in this area. We obtained a 0.70 macro F1-score with RoBERTa-large, outperforming the most recent work in argument relation identification, even with a harder instance of the task (macro F1-score of 0.67 on Claim/Premise classification in [5]). In order to have a more strong reference to compare with previous results, we carried out an experiment using the same parameters but restricting the classes to be learnt to only attack and support relations.…”

Section: Resultsmentioning

confidence: 64%

“…With the advent of Neural Networks (NNs), a performance gap between previous works and this new approach could be observed. In [3] and [5] the empirical results obtained by Recurrent Neural Network (RNN) models for AM were significantly better. However, there is an interesting observation to make emphasis on, which makes it hard to compare AM works.…”

Section: Related Workmentioning

confidence: 99%

“…However, there is an interesting observation to make emphasis on, which makes it hard to compare AM works. As it can be pointed out after looking at the results depicted in works like [12] or [5], the corpus used in each work has a lot of influence in the results. This is due to many different factors such as class distributions, variable language complexity (e.g., use of irony, enthymemes, etc.)…”

Section: Related Workmentioning

confidence: 99%

See 2 more Smart Citations

Transformer-Based Models for Automatic Identification of Argument Relations: A Cross-Domain Evaluation

Ruiz-Dolz,

Heras,

Alemany

et al. 2020

Preprint

View full text Add to dashboard Cite

Argument Mining is defined as the task of automatically identifying and extracting argumentative components (e.g., premises, claims, etc.) and detecting the existing relations among them (i.e., support, attack, rephrase, no relation). One of the main issues when approaching this problem is the lack of data, and the size of the publicly available corpora. In this work, we use the recently annotated US2016 debate corpus. US2016 is the largest existing argument annotated corpus, which allows exploring the benefits of the most recent advances in Natural Language Processing in a complex domain like Argument (relation) Mining. We present an exhaustive analysis of the behavior of transformer-based models (i.e., BERT, XLNET, RoBERTa, DistilBERT and ALBERT) when predicting argument relations. Finally, we evaluate the models in five different domains, with the objective of finding the less domain dependent model. We obtain a macro F1-score of 0.70 with the US2016 evaluation corpus, and a macro F1-score of 0.61 with the Moral Maze cross-domain corpus.

show abstract

Section: Resultsmentioning

confidence: 64%

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Transformer-Based Models for Automatic Identification of Argument Relations: A Cross-Domain Evaluation

Ruiz-Dolz,

Heras,

Alemany

et al. 2020

Preprint

View full text Add to dashboard Cite

show abstract

“…Different methods have been employed to address these tasks, from standard Support Vector Machines (SVMs) to Neural Networks (NNs). AM methods have been applied to heterogeneous types of textual documents, e.g., persuasive essays [22], scientific articles [23], Wikipedia articles [24], political speeches and debates [25,26], and peer reviews [27]. However, only few approaches [28,15,16,17,18] focused on automatically detecting argumentative structures from textual documents in the medical domain, such as clinical trials, clinical guidelines, and Electronic Health Records.…”

Section: Related Workmentioning

confidence: 99%

Enhancing evidence-based medicine with natural language argumentative analysis of clinical trials

Mayer

Marro

Cabrio

et al. 2021

Artificial Intelligence in Medicine

Self Cite

View full text Add to dashboard Cite

In the latest years, the healthcare domain has seen an increasing interest in the definition of intelligent systems to support clinicians in their everyday tasks and activities. Among others, also the field of Evidence-Based Medicine is impacted by this twist, with the aim to combine the reasoning frameworks proposed thus far in the field with mining algorithms to extract structured information from clinical trials, clinical guidelines, and Electronic Health Records. In this paper, we go beyond the state of the art by proposing a new end-to-end pipeline to address argumentative outcome analysis on clinical trials. More precisely, our pipeline is composed of (i) an Argument Mining module to extract and classify argumentative components (i.e., evidence and claims of the trial) and their relations (i.e., support, attack), and (ii) an outcome analysis module to identify and classify the effects (i.e., improved, increased, decreased, no difference, no occurrence) of an intervention on the outcome of the trial, based on PICO elements. We annotated a dataset composed of more than 500 abstracts of Randomized Controlled Trials (RCT) from the MEDLINE database, leading to a labeled dataset with 4198 argument components, 2601 argument relations, and 3351 outcomes on five different diseases (i.e., neoplasm, glaucoma, hepatitis, diabetes, hypertension). We experiment with deep bidirectional transformers in combination with different neural architectures (i.e., LSTM, GRU and CRF) and obtain a macro F1-score of .87 for component detection and .68 for relation prediction, outperforming current state-of-the-art end-to-end Argument Mining systems, and a macro F1-score of .80 for outcome classification.

show abstract

“…Distinctive feature removal methods namely n-grams, POS tagging etc are utilised to extract significant features. Shohreh Haddadan et al [9] presented an approach to mine sentiments from political debates. He considered US presidential campaign debates for sentiment classification.…”

Section: Literature Reviewmentioning

confidence: 99%

Twitter Sentiment Analysis: A Political View

Pinto¹

2020

IJATCSE

View full text Add to dashboard Cite

Twitter is a famous micro-blogging service for tracking public mood with respect to an object or entity. It involves an application of sentiment analysis on naturally written language to extract sentiments conveyed by the users. Twitter sentiment analysis is demanding. Because of the raw nature of text, limited size, use of slang words, abbreviations, tweet processing task became quite complicated. But, one cannot ignore the twitter data, as it hold sentiments towards variety of topics. This research aims to analyse political orientation of twitter users as positive or negative. Different machine learning algorithms are adopted to identify the user point of view on Ayodhya issue. Then the efficiencies of these built models are contrasted with one another to discover the best machine learning algorithm on text categorisation. The result is analysed based on the measurement metrics namely accuracy, precision, recall, F1score measures for each sentiment class.

show abstract

Yes, we can! Mining Arguments in 50 Years of US Presidential Campaign Debates

Cited by 29 publications

References 19 publications

Transformer-Based Models for Automatic Identification of Argument Relations: A Cross-Domain Evaluation

Transformer-Based Models for Automatic Identification of Argument Relations: A Cross-Domain Evaluation

Enhancing evidence-based medicine with natural language argumentative analysis of clinical trials

Twitter Sentiment Analysis: A Political View

Contact Info

Product

Resources

About