Automatic linguistic analysis can provide cost-effective, valuable clues to the diagnosis of cognitive difficulties and to therapeutic practice, and hence impact positively on wellbeing. In this work, we analyzed transcribed conversations between elderly individuals living with dementia and healthcare professionals. The material came from the Anchise 2022 Corpus, a large collection of transcripts of conversations in Italian recorded in naturalistic conditions. The aim of the work was to test the effectiveness of a number of automatic analyzes in finding correlations with the progression of dementia in individuals with cognitive decline as measured by the Mini-Mental State Examination (MMSE) score, which is the only psychometric-clinical information available on the participants in the conversations. Healthy controls (HC) were not considered in this study, nor does the corpus itself include HCs. The main innovation and strength of the work consists in the high ecological validity of the language analyzed (most of the literature to date concerns controlled language experiments); in the use of Italian (there is little corpora for Italian); in the size of the analyzed data (more than 200 conversations were considered); in the adoption of a wide range of NLP methods, that span from traditional morphosyntactic investigation to deep linguistic models for conducting analyzes such as through perplexity, sentiment (polarity) and emotions. Analyzing real-world interactions not designed with computational analysis in mind, such as is the case of the Anchise Corpus, is particularly challenging. To achieve the research goals, a wide variety of tools were employed. These included traditional morphosyntactic analysis based on digital linguistic biomarkers (DLBs), transformer-based language models, sentiment and emotion analysis, and perplexity metrics. Analyzes were conducted both on the continuous range of MMSE values and on the severe/moderate/mild categorization suggested by AIFA (Italian Medicines Agency) guidelines, based on MMSE threshold values. Correlations between MMSE and individual DLBs were weak, up to 0.19 for positive, and -0.21 for negative correlation values. Nevertheless, some correlations were statistically significant and consistent with the literature, suggesting that people with a greater degree of impairment tend to show a reduced vocabulary, to have anomia, to adopt a more informal linguist register, and to display a simplified use of verbs, with a decrease in the use of participles, gerunds, subjunctive moods, modal verbs, as well as a flattening in the use of the tenses towards the present to the detriment of the past. The -0.26 inverse correlation between perplexity and MMSE suggests that perplexity captures slightly more specific linguistic information, which can complement the MMSE scores. In the categorization tasks, the classifier based on DLBs achieved an F1 score of 0.79 for binary classification between SEVERE and MILD, and 0.61 for multi-label categorization. Sentiment and emotion analyzes showed inverse trends for joy while MMSE scores suggested that less impaired individuals were less joyful, or more “negative”, than others. Considering the real-world context, this is consistent with the hypothesis of a gradual reduction in awareness in individuals affected by dementia. Finally, integrating various profiles of analysis has been proved to be effective in offering a wider picture of linguistic and communication deficits, as well as more precise data regarding the progression of dementia.

Marco Antonio Stranisci, Rossana Damiano, Enrico Mensa, Viviana Patti, Daniele Radicioni, Tommaso Caselli. Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2023.

We present LESSLEX, a novel multilingual lexical resource. Different from the vast majority of existing approaches, we ground our embeddings on a sense inventory made available from the BabelNet semantic network. In this setting, multilingual access is governed by the mapping of terms onto their underlying sense descriptions, such that all vectors co-exist in the same semantic space. As a result, for each term we have thus the “blended” terminological vector along with those describing all senses associated to that term. LESSLEX has been tested on three tasks relevant to lexical semantics: conceptual similarity, contextual similarity, and semantic text similarity. We experimented over the principal data sets for such tasks in their multilingual and crosslingual variants, improving on or closely approaching state-of-the-art results. We conclude by arguing that LESSLEX vectors may be relevant for practical applications and for research on conceptual and lexical access and competence.

In this article we present an advanced version of Dual-PECCS, a cognitively-inspired knowledge representation and reasoning system aimed at extending the capabilities of artificial systems in conceptual categorization tasks. It combines different sorts of common-sense categorization (prototypical and exemplars-based categorization) with standard monotonic categorization procedures. These different types of inferential procedures are reconciled according to the tenets coming from the dual process theory of reasoning. On the other hand, from a representational perspective, the system relies on the hypothesis of conceptual structures represented as heterogeneous proxytypes. Dual-PECCS has been experimentally assessed in a task of conceptual categorization where a target concept illustrated by a simple common-sense linguistic description had to be identified by resorting to a mix of categorization strategies, and its output has been compared to human responses. The obtained results suggest that our approach can be beneficial to improve the representational and reasoning conceptual capabilities of standard cognitive artificial systems, and – in addition – that it may be plausibly applied to different general computational models of cognition. The current version of the system, in fact, extends our previous work, in that Dual- PECCS is now integrated and tested into two cognitive architectures, ACT-R and CLARION, implementing different assumptions on the underlying invariant structures governing human cognition. Such integration allowed us to extend our previous evaluation.

In this work we present a knowledge-based system equipped with a hybrid, cognitively inspired architecture for the representation of conceptual information. The proposed system aims at extending the classical representational and reasoning capabilities of the ontology-based frameworks towards the realm of the prototype theory. It is based on a hybrid knowledge base, composed of a classical symbolic component (grounded on a formal ontology) with a typicality based one (grounded on the conceptual spaces framework). The resulting system attempts to reconcile the heterogeneous approach to the concepts in Cognitive Science with the dual process theories of reasoning and rationality. The system has been experimentally assessed in a conceptual categorisation task where common sense linguistic descriptions were given in input, and the corresponding target concepts had to be identified. The results show that the proposed solution substantially extends the representational and reasoning ‘conceptual’ capabilities of standard ontology-based systems.

In this work we present an ontology-based system equipped with a hybrid architecture for the representation of conceptual information. The proposed system aims at extending the representational and reasoning capabilities of classical ontology-based systems towards more realistic and cognitively grounded scenarios, such as those envisioned by the prototype theory. The resulting system attempts to reconcile the heterogeneous approach to the concepts in Cognitive Science and the dual process theories of reasoning and rationality. The system has been experimentally assessed in a conceptual categorization task where common sense linguistic descriptions were given in input, and the corresponding target concepts had to be identified. The results show that the proposed solution substantially improves on the representational and reasoning “conceptual” capabilities of standard ontology-based systems

In the last few years University of Turin and CIRSFID University of Bologna collaborated to pair NLP techniques and legal knowledge to detect modificatory provisions in normative texts. Annotating these modifications is a relevant and interesting problem, in that modifications affect the whole normative system; and legal language, though more regular than unrestricted language, is sometimes particularly convoluted, and poses specific linguistic issues. This paper focuses on two major aspects. First, we explore a combination between parsing and regular expressions; to the best of our knowledge, such hybrid strategy has never been proposed before to tackle the problem at hand. Secondly, we significantly extend past works coverage (basically focussed on substitution, integration and repeal modifications) in order to account for further twelve modification kinds. For the sake of conciseness, we fully illustrate and discuss only few modification types that are more relevant and interesting: suspension, prorogation of efficacy, postponement of efficacy and exception/derogation. These sorts of modifications appear particularly challenging, in that modifications in these categories make use of similar linguistic speech acts and verbs, and exhibit strong similarities in the linguistic syntactical patterns, to such an extent that to discern them is difficult for the legal expert, too. We describe the implemented system and report about an extensive experimentation on the new modificatory provisions. Results are discussed in order to improve both system's accuracy and annotation practice.

In this paper we investigate how to categorize text excerpts from Italian normative texts. Although text categorization is a problem of broader interest, we single out a specific issue. Namely, we are concerned with categorizing the set of subjects in which Italian Regions are allowed to produce norms: this is the so-called residual legislative power problem. It basically consists in making explicit a set of subjects that was originally defined only in a residual and negative fashion. The categorization of legal text fragments is acknowledged to be a difficult problem, featured by abstract concepts along with a variety of locutions used to denote them, by convoluted sentence structure, and by several other facets. In addition, in the present case subjects are often partially overlapped, and a training set of sufficient size (for the problem under consideration) does not exist: all these aspects make our task challenging. In this setting, classical feature-based approaches provide poor quality results, so we explored algorithms based on compression techniques. We tested three such techniques: we illustrate their main features and report the results of an experimentation where our implementation of such algorithms is compared with the output of standard machine learning algorithms. Far from having found a silver bullet, we show that compression-based techniques provide the best results for the problem at hand, and argue that these approaches can be effectively coupled with more informative and semantically grounded ones.

In this paper we present a prototype for automatically identifying and classifying types of modifications in Italian legal text. The prototype is part of the Eunomos system, a legal knowledge management service that integrates and makes available legislation from various sources, while finding definitions and explanations of legal concepts in a given context. The design of the prototype is grounded on the error analysis of a previous prototype. The latter made use of dependency relations provided by the TUP parser, a multi-purpose parser for Italian. Since those syntactic relations were responsible of the majority of errors, we decided in the present tool to ignore them, and to rewrite an ad-hoc shallow parsing, based on the morphological analysis of the legal text (still provided by the TUP parser). We obtained performances much greater than those of the initial prototype. In particular, the level of precision of the classification in output is now close to 100%.

One open problem in the AI & Law community is how to provide computers with a basic understanding of legal concepts, and their relationship with legal texts and with the legal lexicon. We propose to add a layer to connect the linguistic description of the provisions to syntactic patterns using FramNet that can be exploited thought NLP tools. A deep-parsing and shallow-semantics approach has been devised to interpret and retrieve the characterizing components of legal modificatory provisions. In this paper we single out the case of efficacy suspension and show how FrameNet approach can provide profit especially to isolate temporal parameters and their interpretation.

In this paper we illustrare a research based on NLP techniques aimed at automatically annotate modificatory provisions. We propose an approach which pairs deep syntactic parsing with rule-based shallow semantic analysis relying on a fine-grained taxonomy of modificatory provisions. The implemented system is evaluated on a large dataset hand-crafted by legal experts; the results are discussed and future directions of the research outlined.

We are concerned with the automatic semantic interpretation of legal modificatory provisions. We propose a novel approach which pairs deep syntactic parsing and a fine-grained taxonomy of legal modifications. Although still in a developmental stage, the implemented system can be used to annotate with meta-information modificatory provisions of NormaInRete documents.

In this paper we present a novel algorithm, CarpeDiem. It significantly improves on the time complexity of Viterbi algorithm, preserving the optimality of the result. This fact has consequences on Machine Learning systems that use Viterbi algorithm during learning or classification. We show how the algorithm applies to the Supervised Sequential Learning task and, in particular, to the HMPerceptron algorithm. We illustrate CarpeDiem in full details, and provide experimental results that support the proposed approach.