Although large language models (LLMs) have apparently acquired a certain level of grammatical knowledge and the ability to make generalizations, they fail to interpret negation, a crucial step in Natural Language Processing. We try to clarify the reasons for the sub-optimal performance of LLMs understanding negation. We introduce a large semi-automatically generated dataset of circa 400,000 descriptive sentences about commonsense knowledge that can be true or false in which negation is present in about 2/3 of the corpus in different forms. We have used our dataset with the largest available open LLMs in a zero-shot approach to grasp their generalization and inference capability and we have also fine-tuned some of the models to assess whether the understanding of negation can be trained. Our findings show that, while LLMs are proficient at classifying affirmative sentences, they struggle with negative sentences and lack a deep understanding of negation, often relying on superficial cues. Although fine-tuning the models on negative sentences improves their performance, the lack of generalization in handling negation is persistent, highlighting the ongoing challenges of LLMs regarding negation understanding and generalization. The dataset and code are publicly available.

In this paper, we analyse and compare several correction methods of knowledge resources with the purpose of improving the abilities of systems that require commonsense reasoning with the least possible human-effort. To this end, we cross-check the WordNet meronymy relation member against the knowledge encoded in a SUMO-based first-order logic ontology on the basis of the mapping between WordNet and SUMO. In particular, we focus on the knowledge in WordNet regarding the taxonomy of animals and plants. Despite being created manually, these knowledge resources — WordNet, SUMO and their mapping — are not free of errors and discrepancies. Thus, we propose three correction methods by semi-automatically improving the alignment between WordNet and SUMO, by performing some few corrections in SUMO and by combining the above two strategies. The evaluation of each method includes the required human-effort and the achieved improvement on unseen data from the WebChild project, that is tested using first-order logic automated theorem provers.

Previous studies have shown that the knowledge about attributes and properties in the SUMO ontology and its mapping to WordNet adjectives lacks of an accurate and complete characterization. A proper characterization of this type of knowledge is required to perform formal commonsense reasoning based on the SUMO properties, for instance to distinguish one concept from another based on their properties. In this context, we propose a new semi-automatic approach to model the knowledge about properties and attributes in SUMO by exploiting the information encoded in WordNet adjectives and its mapping to SUMO. To that end, we considered clusters of semantically related groups of WordNet adjectival and nominal synsets. Based on these clusters, we propose a new semi-automatic model for SUMO attributes and their mapping to WordNet, which also includes polarity information. In this paper, as an exploratory approach, we focus on qualities.

We describe a detailed analysis of a sample of large benchmark of commonsense reasoning problems that has been automatically obtained from WordNet, SUMO and their mapping. The objective is to provide a better assessment of the quality of both the benchmark and the involved knowledge resources for advanced commonsense reasoning tasks. By means of this analysis, we are able to detect some knowledge misalignments, mapping errors and lack of knowledge and resources. Our final objective is the extraction of some guidelines towards a better exploitation of this commonsense knowledge framework by the improvement of the included resources.

We describe the practical application of a black-box testing methodology for the validation of the knowledge encoded in WordNet, SUMO and their mapping by using automated theorem provers. In this paper,weconcentrateonthepart-whole information provided by WordNet and create a large set of tests on the basis of few question patterns. From our preliminary evaluation results, we report on some of the detected inconsistencies.

This paper presents the complete and consistent ontological annotation of the nominal part of WordNet. The annotation has been carried out using the semantic features defined in the EuroWordNet Top Concept Ontology and made available to the NLP community. Up to now only an initial core set of 1,024 synsets, the so-called Base Concepts, was ontologized in such a way. The work has been achieved by following a methodology based on an iterative and incremental expansion of the initial labeling through the hierarchy while setting inheritance blockage points. Since this labeling has been set on the EuroWordNet’s Interlingual Index (ILI), it can be also used to populate any other wordnet linked to it through a simple porting process. This feature-annotated WordNet is intended to be useful for a large number of semantic NLP tasks and for testing for the first time componential analysis on real environments. Moreover, the quantitative analysis of the work shows that more than 40% of the nominal part of WordNet is involved in structure errors or inadequacies.