Hybrid Artificial Intelligence to extract patterns and rules from argumentative and legal texts

Doctoral thesis (2022)

This Thesis is composed of a selection of studies realized between 2019 and 2022, whose aim is to find working methodologies of Artificial Intelligence (AI) and Machine Learning for the detection and ... [more ▼]

This Thesis is composed of a selection of studies realized between 2019 and 2022, whose aim is to find working methodologies of Artificial Intelligence (AI) and Machine Learning for the detection and classification of patterns and rules in argumentative and legal texts. We define our approach as “hybrid”, since different methods have been employed combining symbolic AI (which involves “top-dow” structured knowledge) and sub-symbolic AI (which involves “bottom-up” data-driven knowledge). The first group of these works was dedicated to the classification of argumentative patterns. Following the Waltonian model of argument (according to which arguments are composed by a set of premises and a conclusion), and the theory of Argumentation Schemes, this group of studies was focused on the detection of argumentative evidences of support and opposition. More precisely, the aim of these first works was to show that argumentative patterns of opposition and support could be classified at fine-grained levels and without resorting to highly engineered features. To show this, we firstly employed methodologies based on Tree Kernel classifiers and TFIDF. In these experiments, we explored different combinations of Tree Kernel calculation and different data structures (i.e., different tree structures). Also, some of these combinations employs a hybrid approach where the calculation of similarity among trees is influenced not only by the tree structures but also by a semantic layer (e.g. those using “smoothed” trees and “compositional” trees). After the encouraging results of this first phase, we explored the use of a new methodology which was deeply changing the NLP landscape exactly in that year, fostered and promoted by actors like Google, i.e. Transfer Learning and the use of language models. These newcomer methodologies markedly improved our previous results and provided us with stronger NLP tools. Using Transfer Learning, we were also able to perform a Sequence Labelling task for the recognition of the exact span of argumentative components (i.e. claims and premises), which is crucial to connect the sphere of natural language to the sphere of logic. The last part of this work was finally dedicated to show how to use Transfer Learning for the detection of rules and deontic modalities. In this case, we tried to explore a hybrid approach which combines structured knowledge coming from two LegalXML formats (i.e., Akoma Ntoso and LegalRuleML) with sub-symbolic knowledge coming from pre-trained (and then fine-tuned) neural architectures. [less ▲]

Hybrid AI Framework for Legal Analysis of the EU Legislation Corrigenda

in Legal Knowledge and Information Systems (2021)

This paper presents an AI use-case developed in the project “Study on legislation in the era of artificial intelligence and digitization” promoted by the EU Commission Directorate-General for Informatics ... [more ▼]

This paper presents an AI use-case developed in the project “Study on legislation in the era of artificial intelligence and digitization” promoted by the EU Commission Directorate-General for Informatics. We propose a hybrid technical framework where AI techniques, Data Analytics, Semantic Web approaches and LegalXML modelisation produce benefits in legal drafting activity. This paper aims to classify the corrigenda of the EU legislation with the goal to detect some criteria that could prevent errors during the drafting or during the publication process. We use a pipeline of different techniques combining AI, NLP, Data Analytics, Semantic annotation and LegalXML instruments for enriching the non-symbolic AI tools with legal knowledge interpretation to offer to the legal experts. [less ▲]

Argumentation Schemes as Templates? Combining Bottom-up and Top-down Knowledge Representation

in Proceedings of the 20th Workshop on Computational Models of Natural Argument (2020)

This paper describes a long-term research goal which aims at creating a middleware interface between Argumentation Schemes and natural language. This idea comes from the need to face some challenges ... [more ▼]

This paper describes a long-term research goal which aims at creating a middleware interface between Argumentation Schemes and natural language. This idea comes from the need to face some challenges related to the automatic extraction of Argumentation Schemes from Nat- ural Language: for example the ability to extract Argumentation Schemes at different level of granularity. In the paper we describe how this process can be designed and how the structures of Argumentation Schemes can be modeled to this aim. [less ▲]

Transfer Learning with Sentence Embeddings for Argumentative Evidence Classification

in Proceedings of the 20th Workshop on Computational Models of Natural Argument (2020)

Comparing Tree Kernels performances in argumentative evidence classification

in CLADAG 2019 - Book of Short Papers (2019)

The purpose of this study is to deploy a novel methodology for classifying argumentative support (or evidence) in arguments. The methodology shows that Tree Kernel can discriminate between different types ... [more ▼]

The purpose of this study is to deploy a novel methodology for classifying argumentative support (or evidence) in arguments. The methodology shows that Tree Kernel can discriminate between different types of argumentative evidence with high scores, while keeping a good generalization. Moreover, the results of two different Tree Kernels are evaluated. [less ▲]

Detecting “slippery slope” and other argumentative stances of opposition using tree kernels in monologic discourse

in International Joint Conference on Rules and Reasoning (2019)

The aim of this study is to propose an innovative methodology to classify argumentative stances in a monologic argumentative context. Particularly, the proposed approach shows that Tree Kernels can be ... [more ▼]

The aim of this study is to propose an innovative methodology to classify argumentative stances in a monologic argumentative context. Particularly, the proposed approach shows that Tree Kernels can be used in combination with traditional textual vectorization to discriminate between different stances of opposition without the need of extracting highly engineered features. This can be useful in many Argument Mining sub-tasks. In particular, this work explores the possibility of classifying opposition stances by training multiple classifiers to reach different degrees of granularity. Noticeably, discriminating support and opposition stances can be particularly useful when trying to detect Argument Schemes, one of the most challenging sub-task in the Argument Mining pipeline. In this sense, the approach can be also considered as an attempt to classify stances of opposition that are related to specific Argument Schemes. [less ▲]