Search

In this paper, we present an entity-centric infrastructure to manage legal documents, especially court judgments, based on the organization of a textual document repository and on the annotation of these documents to serve a variety of downstream tasks. Documents are pre-processed and then iteratively annotated using a set of NLP services that combine complementary approaches based on machine learning and syntactic rules. We present a framework that has been designed to be developed and maintained in a sustainable way, allowing for multiple services and uses of the annotated document repository and considering the scarcity of annotated data as an intrinsic challenge for its development. The design activity is the result of a cooperative project where a scientific team, institutional bodies, and companies appointed to implement the final system are involved in co-design activities. We describe experiments to demonstrate the feasibility of the solution and discuss the main challenges to scaling the system at a national level. In particular, we report the results we obtained in annotating data with different low-resource methods and with solutions designed to combine these approaches in a meaningful way. An essential aspect of the proposed solution is a human-in-the-loop approach to control the output of the annotation algorithms in agreement with the organizational processes in place in Italian courts. Based on these results we advocate for the feasibility of the proposed approach and discuss the challenges that must be addressed to ensure the scalability and robustness of the proposed solution.

The extraction of named entities from court judgments is useful in several downstream applications, such as document anonymization and semantic search engines. In this paper, we discuss the application of named entity recognition and linking (NEEL) to extract entities from Italian civil court judgments. To develop and evaluate our work, we use a corpus of 146 manually annotated court judgments. We use a pipeline that combines a transformer-based Named Entity Recognition (NER) component, a transformer-based Named Entity Linking (NEL) component, and a NIL prediction component. While the NEL and NIL prediction components are not fine-tuned on domain-specific data, the NER component is fine-tuned on the annotated corpus. In addition, we compare different masked language modeling (MLM) adaptation strategies to optimize the result and investigate their impact. Results obtained on a 30-document test set reveal satisfactory performance, especially on the NER task, and emphasize challenges to improve NEEL on similar documents. Our code is available on GitHub.(https://github.com/rpo19/pozzi_aixia_2023. We are not allowed to publish sensitive data and the NER models trained on sensitive data.)

In this paper, we present MammoTab, a dataset composed of 1M Wikipedia tables extracted from over 20M Wikipedia pages and annotated through Wikidata. The lack of this kind of datasets in the state- of-the-art makes MammoTab a good resource for testing and training Semantic Table Interpretation approaches. The dataset has been designed to cover several key challenges, such as disambiguation, homonymy, and NIL-mentions. The dataset has been evaluated using MTab, one of the best approaches of the SemTab challenge.

Named entity extraction is a crucial task to support the population of Knowledge Bases (KBs) from documents written in natural language. However, in many application domains, these documents must be collected and processed incrementally to update the KB as more data are ingested. In some cases, quality concerns may even require human validation mechanisms along the process. While very recent work in the NLP community has discussed the importance of evaluating and benchmarking continuous entity extraction, it has proposed methods and datasets that avoid Named Entity Linking (NEL) as a component of the extraction process. In this paper, we advocate for batch-based incremental entity extraction methods that can exploit NEL with a background KB, detect mentions of entities that are not present in the KB yet (NIL mentions), and update the KB with the novel entities. Based on this assumption, we present a methodology to evaluate NEL-based incremental entity extraction, which can be applied to a “static” dataset for evaluating NEL into a dataset for evaluating incremental entity extraction. We apply this methodology to an existing benchmark for evaluating NEL algorithms, and evaluate an incremental extraction pipeline that orchestrates different strong state-of-the-art and baseline algorithms for the tasks involved in the extraction process, namely, NEL, NIL prediction, and NIL clustering. In presenting our experiments, we demonstrate the increased difficulty of the information extraction task in incremental settings and discuss the strengths of the available solutions as well as open challenges.

In this paper, we present MammoTab, a dataset composed of 1M Wikipedia tables extracted from over 20M Wikipedia pages and annotated through Wikidata. The lack of this kind of datasets in the state- of-the-art makes MammoTab a good resource for testing and training Semantic Table Interpretation approaches. The dataset has been designed to cover several key challenges, such as disambiguation, homonymy, and NIL-mentions. The dataset has been evaluated using MTab, one of the best approaches of the SemTab challenge.

Chronic immunosuppression is associated with increased and more severe viral infections. However, little is known about the association between immunosuppression and severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection. Our aim was to describe the clinical course of patients with immunosuppressed autoimmune hepatitis (AIH) during coronavirus disease 2019 (COVID-19) infection in Italy. Our study is a case series of patients with AIH treated with immunosuppression, who tested positive for SARS-CoV-2 in March 2020 during the outbreak of COVID-19. Ten patients from seven different hospitals in Italy were diagnosed with COVID-19 during the outbreak of SARS-CoV-2 in March 2020. Seven subjects were female (70%), and age ranged from 27 to 73 years. Before the onset of SARS-CoV-2 infection, all patients were taking immunosuppressive therapy for AIH, and eight of them were on biochemical remission. Two other patients had recent acute onset of their AIH, and consequently started high-dose steroids, as per induction protocol. All patients had a respiratory syndrome and a positive nasal swab for SARS-CoV-2. Five patients developed a computed tomography-confirmed COVID-19 pneumonia. Six subjects received a combination of antiretroviral and antimalarial drugs. In seven patients, the dosage of immunosuppressive medication was changed. Liver enzymes were repeated during SARS-CoV-2 infection in all hospitalized cases; they remained within the normal range in all cases, and improved in the two acute cases treated with high-dose steroids. The clinical outcome was comparable to the reported cases occurring in non-immunosuppressed subjects.Conclusion:Patients under immunosuppressive therapy for AIH developing COVID-19 show a disease course presumptively similar to that reported in the non-immunosuppressed population. These data might aid in medical decisions when dealing with SARS-CoV-2 infection in immunocompromised patients.

Background - There are limited data on aspirin (ASA) desensitization for patients with coronary artery disease. The aim of the present study was to assess the safety and efficacy of a standard rapid desensitization protocol in patients with ASA sensitivity undergoing coronary angiography. Methods and Results - This is a prospective, multicenter, observational study including 7 Italian centers including patients with a history of ASA sensitivity undergoing coronary angiography with intent to undergo percutaneous coronary intervention. A total of 330 patients with history of ASA sensitivity with known/suspected stable coronary artery disease or presenting with an acute coronary syndrome, including ST-segment-elevation myocardial infarction were enrolled. Adverse effects to aspirin included urticaria (n=177, 53.6%), angioedema (n=69, 20.9%), asthma (n=65, 19.7%), and anaphylactic reaction (n=19, 5.8%). Among patients with urticaria/angioedema, 13 patients (3.9%) had a history of idiopathic chronic urticaria. All patients underwent a rapid ASA (5.5 hours) desensitization procedure. The desensitization procedure was performed before cardiac catheterization in all patients, except for those (n=78, 23.6%) presenting with ST-segment-elevation myocardial infarction who underwent the desensitization after primary percutaneous coronary intervention. Percutaneous coronary intervention was performed in 235 patients (71%) of the overall study population. The desensitization procedure was successful in 315 patients (95.4%) and in all patients with a history of anaphylactic reaction. Among the 15 patients (4.6%) who did not successfully respond to the desensitization protocol, adverse reactions were minor and responded to treatment with corticosteroids and antihistamines. Among patients with successful in-hospital ASA desensitization, 253 patients (80.3%) continued ASA for at least 12 months. Discontinuation of ASA in the 62 patients (19.7%) who had responded to the desensitization

Background: Many studies have investigated the relationships between cardiovascular diseases and patients' depression; nevertheless, few is still known as regard the impact of illness severity on depression and whether psychosocial variables mediate this association. Purpose: The aim of this study is to investigate the putative mediating role of illness representations, self-efficacy beliefs, and perceived social support on the relationship between illness severity and depression. Methods: A total of 75 consecutive patients with cardiovascular disease (80 % men; mean age=65.44, SD=10.20) were enrolled in an Italian hospital. Illness severity was measured in terms of left ventricular ejection fraction, whereas psychological factors were assessed using self-report questionnaires. Results: The relationship between left ventricular ejection fraction and depression was mediated by identity illness perception, self-efficacy beliefs in managing cardiac risk factors, and perceived social support. Conclusion: The treatment of depression in cardiovascular disease patients may therefore benefit from a psychological intervention focused on patients' illness representations, self-efficacy beliefs, and their perceived social support. © 2013 International Society of Behavioral Medicine.