Knowledge graph informed fake news classification via heterogeneous representation ensembles

Koloski, Boshko; Perdih, Timen Stepišnik; Robnik–Šikonja, Marko; Pollak, Senja; Škrlj, Blaž

doi:10.1016/j.neucom.2022.01.096

Cited by 50 publications

(20 citation statements)

References 45 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Knowledge-backed representation of documents has proven to be useful in text classification tasks (Koloski et al, 2022). We explore how these representations perform in the problem of the detection of depression.…”

Section: Knowledge Graphsmentioning

confidence: 99%

E8-IJS@LT-EDI-ACL2022 - BERT, AutoML and Knowledge-graph backed Detection of Depression

Ilija¹,

Koloski²,

Škrlj³

et al. 2022

Proceedings of the Second Workshop on Language Technology for Equality, Diversity and Inclusion

Self Cite

View full text Add to dashboard Cite

Depression is a mental illness that negatively affects a person's well-being and can, if left untreated, lead to serious consequences such as suicide. Therefore, it is important to recognize the signs of depression early. In the last decade, social media has become one of the most common places to express one's feelings. Hence, there is a possibility of text processing and applying machine learning techniques to detect possible signs of depression. In this paper, we present our approaches to solving the shared task titled Detecting Signs of Depression from Social Media Text. We explore three different approaches to solve the challenge: finetuning BERT model, leveraging AutoML for the construction of features and classifier selection and finally, we explore latent spaces derived from the combination of textual and knowledge-based representations. We ranked 9th out of 31 teams in the competition. Our best solution, based on knowledge graph and textual representations, was 4.9% behind the best model in terms of Macro F1, and only 1.9% behind in terms of Recall.

show abstract

Section: Knowledge Graphsmentioning

confidence: 99%

E8-IJS@LT-EDI-ACL2022 - BERT, AutoML and Knowledge-graph backed Detection of Depression

Ilija¹,

Koloski²,

Škrlj³

et al. 2022

Proceedings of the Second Workshop on Language Technology for Equality, Diversity and Inclusion

Self Cite

View full text Add to dashboard Cite

show abstract

“…We use the Wikidata5m (Wang et al, 2021) knowledge graph (KG) to retrieve knowledge-based features as used by Koloski et al (2022). Similarly, we exploit six different knowledge graph embeddings: transE (Bordes et al, 2013), rotatE (Sun et al, 2019), complEx (Trouillon et al, 2016), distmult (Yang et al, 2015), simplE (Kazemi and Poole, 2018), and quate (Zhang et al, 2019).…”

Section: Knowledge Graph Similarity and Regressionmentioning

confidence: 99%

EMBEDDIA at SemEval-2022 Task 8: Investigating Sentence, Image, and Knowledge Graph Representations for Multilingual News Article Similarity

Zosa¹,

Boroş²,

Koloski³

et al. 2022

Proceedings of the 16th International Workshop on Semantic Evaluation (SemEval-2022)

Self Cite

View full text Add to dashboard Cite

In this paper, we present the participation of the EMBEDDIA team in the SemEval-2022 Task 8 (Multilingual News Article Similarity). We cover several techniques and propose different methods for finding the multilingual news article similarity by exploring the dataset in its entirety. We take advantage of the textual content of the articles, the provided metadata (e.g., titles, keywords, topics), the translated articles, the images (those that were available), and knowledge graph-based representations for entities and relations present in the articles. We, then, compute the semantic similarity between the different features and predict through regression the similarity scores. Our findings show that, while our proposed methods obtained promising results, exploiting the semantic textual similarity with sentence representations is unbeatable. Finally, in the official SemEval-2022 Task 8, we ranked fifth in the overall team ranking cross-lingual results, and second in the English-only results.

show abstract

“…The method used in the literature adopted a symbolic reasoning system along with a neural network to improve the efficiency and accuracy of enterprise risk management and false news identification. Besides, there is a literature on the application of type 1 neuro-symbolic AI method in the fields of journalism and communication [22] and education [23]. Researchers mainly rely on deep neural networks to carry out pattern recognition, but the input calculation data is replaced by symbolic knowledge or knowledge map.…”

Section: What Can Be Done With Neuro-symbolic Ai?mentioning

confidence: 99%

Introducing Neuro-Symbolic Artificial Intelligence to Humanities and Social Sciences: Why Is It Possible and What Can Be Done?

Huang¹

2022

TEM Journal

View full text Add to dashboard Cite

With the support of artificial intelligence (AI), the smart applications in all walks of life have brought great changes to human society. Not only being concerned, analysed and criticized by scholars from Humanities and social sciences, AI also plays an important role in empirical research methods, thus facilitating the transformation of research paradigms in these fields. At present, neuro-symbolic AI, as a new product of the integration of two major factions in the field of artificial intelligence - connectionism and symbolism, has high application value in studying and solving the humanistic and social problems involving massive data due to its learning capability of perceiving the environment as well as reasoning capability of manipulating symbols. The introduction of neuro-symbolic AI is also of great significance for the development of emerging interdisciplinary fields such as digital humanities and computational social sciences. This paper aims to clarify the connections between neuro-symbolic AI and Humanities and social sciences, summarize the latest developing trends and representative applications, and explore a feasible path for the expansion of pluralistic methodologies in Humanities and social sciences to adapt to the age of big data.

show abstract

Knowledge graph informed fake news classification via heterogeneous representation ensembles

Cited by 50 publications

References 45 publications

E8-IJS@LT-EDI-ACL2022 - BERT, AutoML and Knowledge-graph backed Detection of Depression

E8-IJS@LT-EDI-ACL2022 - BERT, AutoML and Knowledge-graph backed Detection of Depression

EMBEDDIA at SemEval-2022 Task 8: Investigating Sentence, Image, and Knowledge Graph Representations for Multilingual News Article Similarity

Introducing Neuro-Symbolic Artificial Intelligence to Humanities and Social Sciences: Why Is It Possible and What Can Be Done?

Contact Info

Product

Resources

About