Papers by Itziar Gonzalez-Dios
Cross-checking WordNet and SUMO Using Meronymy (L18-1)
Copied to clipboard
| Challenge: | Existing methods to validate knowledge encoded in WordNet, SUMO and their mapping have been mainly manual . |
| Approach: | They propose to use WordNet and SUMO to validate knowledge by using automated theorem provers to evaluate the competency of SU MO-based ontologies. |
| Outcome: | The proposed method enables validation of some pieces of information and also the detection of missing information or inconsistencies among knowledge resources. |
Textual Entailment for Event Argument Extraction: Zero- and Few-Shot with Multi-Source Learning (2022.findings-naacl)
Copied to clipboard
| Challenge: | Recent work shows that Relation Extraction tasks can be recasted as Textual Entailment tasks using verbalizations. |
| Approach: | They propose to recasted RE tasks as Textual Entailment tasks using verbalizations . they show that entailment reduces the need for manual annotation to 50% and 20% . |
| Outcome: | The proposed method reduces the need for manual annotation to 50% and 20% in event argument extraction tasks while achieving the same performance as with full training. |
This is not a Dataset: A Large Negation Benchmark to Challenge Large Language Models (2023.emnlp-main)
Copied to clipboard
| Challenge: | Large language models (LLMs) have grammatical knowledge but fail to interpret negation . a recent study shows that LLMs struggle with negative sentences . |
| Approach: | They propose to use a dataset to grasp LLMs' generalization and inference capability . they also fine-tuned models to assess whether the understanding of negation can be trained . |
| Outcome: | The proposed model is able to generalize and infer negation in 400,000 sentences . but it is suboptimal when it comes to negation, a key step in natural language processing . |
A Multi-layered Approach to Physical Commonsense Understanding: Creation and Evaluation of an Italian Dataset (2024.lrec-main)
Copied to clipboard
| Challenge: | Using a multilingual model, we examine the ability of large language models to perform reasoning tasks. |
| Approach: | They propose to use a multilingual model to analyze commonsense reasoning in large language models for Italian and to provide a semi-automated system to complete the annotation. |
| Outcome: | The proposed model performs at high-level classification tasks but its easoning is inconsistent and unverifiable, since it does not capture intermediate evidence. |
A Preliminary Study of ChatGPT for Spanish E2R Text Adaptation (2024.lrec-main)
Copied to clipboard
| Challenge: | 16% of the world's population lacks literacy skills, according to the United Nations . |
| Approach: | They propose to use ChatGPT-4 to generate E2R text prompts and a checklist-based manual evaluation to evaluate 10 texts adapted by ChatGPt-4. |
| Outcome: | The proposed model is able to generate written language and adapt to LF in 10 texts. |