Browsing by Author "Reforgiato Recupero, Diego"
Now showing 1 - 5 of 5
Results Per Page
Sort Options
- ItemAdvances in Semantics and Explainability for NLP: Joint Proceedings of the 2nd International Workshop on Deep Learning meets Ontologies and Natural Language Processing (DeepOntoNLP 2021) & 6th International Workshop on Explainable Sentiment Mining and Emotion Detection (X-SENTIMENT 2021), co-located with the 18th Extended Semantic Web Conference (ESWC 2021)(Aachen : RWTH Aachen, 2021) Ben Abbès, Sarra; Hantach, Rim; Calvez, Philippe; Buscaldi, Davide; Dessì, Danilo; Dragoni, Mauro; Reforgiato Recupero, Diego; Sack, Harald[no abstract available]
- ItemChallenges of Applying Knowledge Graph and their Embeddings to a Real-world Use-case(Aachen, Germany : RWTH Aachen, 2021) Petzold, Rick; Gesese, Genet Asefa; Bogdanova, Viktoria; Zylowski, Thorsten; Sack, Harald; Alam, Mehwish; Alam, Mehwish; Buscaldi, Davide; Cochez, Michael; Osborne, Francesco; Reforgiato Recupero, Diego; Sack, HaraldDifferent Knowledge Graph Embedding (KGE) models have been proposed so far which are trained on some specific KG completion tasks such as link prediction and evaluated on datasets which are mainly created for such purpose. Mostly, the embeddings learnt on link prediction tasks are not applied for downstream tasks in real-world use-cases such as data available in different companies/organizations. In this paper, the challenges with enriching a KG which is generated from a real-world relational database (RDB) about companies, with information from external sources such as Wikidata and learning representations for the KG are presented. Moreover, a comparative analysis is presented between the KGEs and various text embeddings on some downstream clustering tasks. The results of experiments indicate that in use-cases like the one used in this paper, where the KG is highly skewed, it is beneficial to use text embeddings or language models instead of KGEs.
- ItemImproving Language Model Predictions via Prompts Enriched with Knowledge Graphs(Aachen, Germany : RWTH Aachen, 2023) Brate, Ryan; Minh-Dang, Hoang; Hoppe, Fabian; He, Yuan; Meroño-Peñuela, Albert; Sadashivaiah, Vijay; Alam, Mehwish; Buscaldi, Davide; Cochez, Michael; Osborne, Francesco; Reforgiato Recupero, DiegoDespite advances in deep learning and knowledge graphs (KGs), using language models for natural language understanding and question answering remains a challenging task. Pre-trained language models (PLMs) have shown to be able to leverage contextual information, to complete cloze prompts, next sentence completion and question answering tasks in various domains. Unlike structured data querying in e.g. KGs, mapping an input question to data that may or may not be stored by the language model is not a simple task. Recent studies have highlighted the improvements that can be made to the quality of information retrieved from PLMs by adding auxiliary data to otherwise naive prompts. In this paper, we explore the effects of enriching prompts with additional contextual information leveraged from the Wikidata KG on language model performance. Specifically, we compare the performance of naive vs. KG-engineered cloze prompts for entity genre classification in the movie domain. Selecting a broad range of commonly available Wikidata properties, we show that enrichment of cloze-style prompts with Wikidata information can result in a significantly higher recall for the investigated BERT and RoBERTa large PLMs. However, it is also apparent that the optimum level of data enrichment differs between models.
- ItemSemantic role labeling for knowledge graph extraction from text(Berlin ; Heidelberg : Springer, 2021) Alam, Mehwish; Gangemi, Aldo; Pressutti, Valentina; Reforgiato Recupero, DiegoThis paper introduces TakeFive, a new semantic role labeling method that transforms a text into a frame-oriented knowledge graph. It performs dependency parsing, identifies the words that evoke lexical frames, locates the roles and fillers for each frame, runs coercion techniques, and formalizes the results as a knowledge graph. This formal representation complies with the frame semantics used in Framester, a factual-linguistic linked data resource. We tested our method on the WSJ section of the Peen Treebank annotated with VerbNet and PropBank labels and on the Brown corpus. The evaluation has been performed according to the CoNLL Shared Task on Joint Parsing of Syntactic and Semantic Dependencies. The obtained precision, recall, and F1 values indicate that TakeFive is competitive with other existing methods such as SEMAFOR, Pikes, PathLSTM, and FRED. We finally discuss how to combine TakeFive and FRED, obtaining higher values of precision, recall, and F1 measure.
- ItemUnderstanding Class Representations: An Intrinsic Evaluation of Zero-Shot Text Classification(Aachen, Germany : RWTH Aachen, 2021) Hoppe, Fabian; Dessì, Danilo; Sack, Harald; Alam, Mehwish; Buscaldi, Davide; Cochez, Michael; Osborne, Francesco; Reforgiato Recupero, Diego; Sack, HaraldFrequently, Text Classification is limited by insufficient training data. This problem is addressed by Zero-Shot Classification through the inclusion of external class definitions and then exploiting the relations between classes seen during training and unseen classes (Zero-shot). However, it requires a class embedding space capable of accurately representing the semantic relatedness between classes. This work defines an intrinsic evaluation based on greater-than constraints to provide a better understanding of this relatedness. The results imply that textual embeddings are able to capture more semantics than Knowledge Graph embeddings, but combining both modalities yields the best performance.