With the rapid growth of social network services, misinformation has spread uncontrollably. Most recent approaches to fake news detection use neural network models to predict whether the input text is fake or real. Some of them even provide explanations, in addition to veracity, generated by Large Language Models (LLMs). However, they do not utilize factual evidence, nor do they allude to it or provide evidence/justification, thereby making their predictions less credible. This paper proposes a new fake news detection method that predicts the truth or false-hood of a claim based on relevant factual evidence (if exists) or LLM’s inference mechanisms (such as common-sense reasoning) otherwise. Our method produces the final synthesized prediction, along with well-founded facts or reasoning. Experimental results on several large COVID-19 fake news datasets show that our method achieves state-of-the-art (SOTA) detection and evidence explanation performance. Our source codes are available online.

Aspect-category-based sentiment analysis (ACSA), which aims to identify aspect categories and predict their sentiments has been intensively studied due to its wide range of NLP applications. Most approaches mainly utilize intrasentential features. However, a review often includes multiple different aspect categories, and some of them do not explicitly appear in the review. Even in a sentence, there is more than one aspect category with its sentiments, and they are entangled intra-sentence, which makes the model fail to discriminately preserve all sentiment characteristics. In this paper, we propose an enhanced coherence-aware network with hierarchical disentanglement (ECAN) for ACSA tasks. Specifically, we explore coherence modeling to capture the contexts across the whole review and to help the implicit aspect and sentiment identification. To address the issue of multiple aspect categories and sentiment entanglement, we propose a hierarchical disentanglement module to extract distinct categories and sentiment features. Extensive experimental and visualization results show that our ECAN effectively decouples multiple categories and sentiments entangled in the coherence representations and achieves state-of-the-art (SOTA) performance. Our codes and data are available online: https://github.com/cuijin-23/ECAN.

Many current language models such as BERT utilize attention mechanisms to transform sequence representations. We ask whether we can influence BERT’s attention with human reading patterns by using eye-tracking and brain imaging data. We fine-tune BERT for relation extraction with auxiliary attention supervision in which BERT’s attention weights are supervised by cognitive data. Through a variety of metrics we find that this attention supervision can be used to increase similarity between model attention distributions over sequences and the cognitive data without significantly affecting classification performance while making unique errors from the baseline. In particular, models with cognitive attention supervision more often correctly classified samples misclassified by the baseline.

Manual text annotation is a resource-consuming endeavor necessary for NLP systems when they target new tasks or domains for which there are no existing annotated corpora. Distributing the annotation work across multiple contributors is a natural solution to reduce and manage the effort required. Although there are a few publicly available tools which support distributed collaborative text annotation, most of them have complex user interfaces and require a significant amount of involvement from the annotators/contributors as well as the project developers and administrators. We present a light-weight web application for highly distributed annotation projects - Djangology. The application takes advantage of the recent advances in web framework architecture that allow rapid development and deployment of web applications thus minimizing development time for customization. The application's web-based interface gives project administrators the ability to easily upload data, define project schemas, assign annotators, monitor progress, and review inter-annotator agreement statistics. The intuitive web-based user interface encourages annotator participation as contributors are not burdened by tool manuals, local installation, or configuration. The system has achieved a user response rate of 70% in two annotation projects involving more than 250 medical experts from various geographic locations.

We describe various syntactic and semantic conditions for finding abstractnouns which refer to concepts of adjectives from a text, in an attempt to explore the creation of a thesaurus from text. Depending on usages, six kinds of syntactic patterns are shown. In the syntactic and semantic conditions an omission of an abstract noun is mainly used, but in addition, various linguistic clues are needed. We then compare our results with synsets of Japanese WordNet. From a viewpoint of Japanese WordNet, the degree of agreement of ?Attribute? between our data and Japanese WordNet was 22%. On the other hand, the total number of differences of obtained abstract nouns was 267. From a viewpoint of our data,the degree of agreement of abstract nouns between our data and Japanese WordNet was 54%.