142 research outputs found
XAI.it 2024 : Preface to the Fifth Italian Workshop on eXplainable Artificial Intelligence
No full textAs Artificial Intelligence (AI) systems become integral to daily life, ensuring transparency and interpretability in their decision-making processes is critical. The General Data Protection Regulation (GDPR) has underscored users’ right to understand how AI-driven systems make decisions that affect them. However, the pursuit of model performance often compromises explainability, creating a tension between achieving high accuracy and maintaining transparency. Core research questions focus on reconciling the high performance of LLMs and other AI models with interpretability requirements. Emerging research focuses on designing transparent systems, understanding the effects of opaque models on users, developing explanation strategies, and enhancing user control over AI behaviors. The workshop on eXplainable AI (XAI.it) provides a platform for addressing these challenges, fostering collaboration within the XAI community to explore novel solutions and share insights across this evolving multifaceted field
CHILab at HaSpeeDe3: Overview of the Taks A Textual
No full textThis technical report illustrates the system developed by the CHILab team for the competition HaSpeeDe3 as part of the EVALITA 2023 campaign. The key idea for HaSpeeDe3 task A - Political Hate Speech Detection - Textual, was to develop different systems arranged as suitable combinations of the Pre-Trained Language Model (PTLM) used for embedding extraction, neural architectures for further elaborations over the embeddings and a classifier. In particular, dense layers, LSTM, BiLSTM and Transformers were used. The best performing system across the ones investigated in this report was made by embeddings extracted via XLM-RoBERTa coupled with BiLSTM that reaches a macro-F1 score of 0.876
Novel method for the simultaneous characterization of bulk impurities and surface states by photocurrent measurements
No full textCHILab at HODI: A minimalist approach
No full textThis technical report illustrates the system developed by the CHILab team for the competition HODI at EVALITA 2023. The key idea of the method we proposed for the HODI Subtask A - Homotransphobia detection, was to develop different systems arranged as suitable combinations of Pre-Trained Language Model (PTLM) for embedding extraction, neural architectures for further elaborations over the embeddings and a classifier. In particular dense layers, LSTM, BiLSTM and Transformers were used as neural architectures. The best performing system across the ones investigated in this report was made by embeddings extracted via AlBERTo coupled with a Transformer that reaches a macro-F1 score of 0.753
Ibuprofen slow-release foam dressing reduces wound pain in painful exuding wounds: preliminary findings from an international real-life study
No full textBackground: Wound pain is a serious problem for people with chronic wounds. The aim of this real-life study was to
compare the effect of a foam dressing that releases ibuprofen (Biatain IbuH) with local best practice on the treatment of
painful exuding wounds. Methods: A total of 185 patients with painful exuding wounds were randomized to either ibuprofen
foam treatment (n598) or local best practice (n587). The primary endpoint was pain relief over 7 days of treatment,
assessed daily using a 5-point verbal rating scale (no relief, slight relief, moderate relief, lots of relief, and complete relief).
Secondary endpoints included a total reduction in pain intensity for the whole study period (using an 11-point Numeric Box
Scale: 05no pain to 105worst possible pain) and incidence of adverse events (AEs). Results: More patients in the ibuprofen
foam treatment group reported wound pain relief and lower wound pain intensity values after 7 days (pv0.0001 for both
variables). Within the four most common ulcer aetiolgies, patients reported significantly more effective pain relief with
ibuprofen foam treatment (venous: p50.009, mixed arterial venous: pv0.0001, arterial: p50.0009, and vasculitis:
p50.009). In all groups, patients from the ibuprofen foam group reported lower pain intensities. The results were significant
for patients with venous (pv0.002) and arterial (pv0.0001) leg ulcers. Two AEs were reported. Conclusions: The ibuprofen
foam represents an effective and safe alternative to local best practice in the management of painful exuding wounds.
Key words: Ibuprofen, leg ulcer, occlusive dressings, pain management, wound healing
Introduction
Pain represents a major problem for patients with
chronic wounds (1–5). Poor management of pain
has been shown to affect patients’ quality of life (6,7)
and potentially influences healing (8). Chronic pain
in leg ulcers is often poorly managed (9), with the
prevalence of pain reported as being as high as 64%
(2). A participant in one study of painful leg ulcers
described his/her situation as being ‘locked in a shell
of pain’ (7). Almost 50% of patients in one study
investigating the impact of dressings and cleansing
agents could not use compression dressings because
of the associated pain (8). People with ulcers of
venous, arterial, mixed arterial venous, and vasculitis
origin are reporting serious problems with
wound pain (1,2). However, there is limited
literature on the effect of active treatments in these
patients.
Foam dressings provide moisture balance to wounds
and have been shown to have excellent absorbency
properties with fewer dressing changes (10). Orally
administered non-steroidal anti-inflammatory drugs,
such as ibuprofen, are excellent for pain management,
but their use can be precluded in some patients owing
to weak blood circulation reducing the local effect, and
side effects such as gatrointestinal bleeding and
decreased renal function (11).
Biatain IbuH (Coloplast A/S, Denmark) is a nonadhesive
foam dressing that continually releases
ibuprofen to the applied wound (7) relative to the
exudate level of the wound. Previous studies in
patients with leg ulcers have shown that use of
ibuprofen foam (Ibu Foam) is associated with
significantly reduced pain associated with leg ulcers
(7,12,13). In addition, use of the Ibu Foam
improved quality of life (12), reduced peri-wound
erythema, and increased healthy granulation tissu
Conditioning Chat-GPT for Information Retrieval: The Unipa-GPT Case Study
No full textThis paper illustrates the architecture and training of Unipa-GPT, a Large Language Model based chatbot developed for assisting students in choosing a bachelor/master degree course at the University of Palermo. Unipa-GPT relies on gpt-3.5-turbo, it was presented in the context of the European Researchers' Night SHARPER event. In our experiments we adopted both the Retrieval Augmented Generation (RAG) approach and fine-tuning to develop the system. The whole architecture of Unipa-GPT is presented, both the RAG and the fine-tuned systems are compared, and a brief discussion on their performance is reported
Contextualized BERT Sentence Embeddings for Author Profiling: The Cost of Performances
No full textThe necessity to know information about the real identity of an online subject is a highly relevant issue in User Profiling, especially for analysis from digital sources such as social media. The digital identity of a user does not always present explicit data about her offline life such as age, gender, work, and more. This problem makes the task of user profiling complex and incomplete. For many years this issue has received a considerable amount of attention from the whole community, which has developed several solutions, also based on machine learning, to estimate user characteristics. The increasing diffusion of deep learning approaches has allowed, on the one hand, to obtain a considerable increase in predictive performance, but on the other hand, to have available models that cannot be interpreted and that require very high computational power. Considering the validity of new pre-trained language models on extensive data for resolving many natural language processing and classification tasks, we decided to propose a BERT-based approach (BERT-DNN) also for the author profiling task. In a first analysis, we compared the results obtained by our model with them of more classical approaches. As a follow, a critical analysis was carried out. We analyze the advantages and disadvantages of these approaches also in terms of resources needed to run them. The results obtained by our model are encouraging in terms of reliability but very disappointing if we consider the computational power required for running it
ATE_ABSITA@EVALITA2020 Task
No full text***************************** TASK DESCRIPTION *********************************
In our challenge, we would like to propose three different annotation tasks regarding Aspect Term Extraction (ATE), Aspect Based Sentiment Analysis (ABSA), and sentence Sentiment Analysis (SA). Aspect Term Extraction (ATE) is the task of identifying an "aspect" in a text without knowing a priori the list that contains it. According to the literature definition, a term/phrase is considered as an aspect when it co-occurs with “opinion words” that indicate a sentiment polarity on it.
More details and examples are available at: http://www.di.uniba.it/~swap/ate_absita/examples.html
Aspect-based Sentiment Analysis (ABSA) is an evolution of Sentiment Analysis that aims at capturing the aspect-level opinions expressed in natural language texts. In the Aspect-based Sentiment Analysis (ABSA) task, the polarity of each expressed aspect is recognized. Sentiment Analysis (or Opinion Mining) is the task of identifying what the user thinks about a particular piece of text. In particular, it often takes the form of an annotation task with the purpose of annotating a portion of text with a positive, negative, or neutral label. In our Sentiment Analysis (SA) task, the polarity of the review is provided. In particular, we decided to use the score left by the user at the item as value of polarity. It is defined as an integer number into the range 1:5.
*************************** ATE ABSITA - EVALITA 2020 **************************
Contacts:
website: http://www.di.uniba.it/~swap/ate_absita/index.html
email: [email protected]
email: [email protected]
PLEASE CITE:
@InProceedings{ateabsita2020,
author = {Lorenzo de Mattei and Graziella de Martino and Andrea Iovine and Alessio Miaschi and Marco Polignano and Giulia Rambelli},
title = {{ATE\_ABSITA@EVALITA2020: Overview of the Aspect Term Extraction and Aspect-based Sentiment Analysis Task}},
booktitle = {{Proceedings of the 7th evaluation campaign of Natural Language Processing and Speech tools for Italian (EVALITA 2020)}},
editor = {Basile, Valerio and Croce, Danilo and Di Maro, Maria and Passaro, Lucia C.},
year = {2020},
publisher = {CEUR.org},
address = {Online}
Measurement by bioluminescence technique of erythrocyte membrane Na+,K+-ATPase activity in hypertensive patients
No full textErythrocyte membrane Na+,K+-ATPase activity was measured using a bioluminescence technique in 28 hypertensive patients (24 with essential hypertension, 2 with renovascular hypertension and 2 with hypertension secondary to primary hyperaldosteronism) and in 28 normotensive control subjects matched for age and sex. Erythrocyte Na+,K+-ATPase activity was significantly reduced in the patients with essential hypertension (130.9 +/- 11.4 vs. 186.6 +/- 19.5 nmol ATP/mg prot per h; mean values +/- SEM; p less than 0.05) and in the patients with secondary hypertension. A significant negative correlation was found between erythrocyte Na+,K+-ATPase and systolic blood pressure (r = -0.603; p less than 0.01), but not between Na+,K+-ATPase and plasma renin activity or plasma aldosterone levels. These data confirm the findings of a number of previous studies reporting reduced activity of erythrocyte Na+,K+-ATPase possibly related to the presence of a circulatory inhibitor of sodium pump. The method, based on ATP assay by bioluminescence, presents a high degree of specificity as well as simple, rapid execution
Preface to the EVALITA 2023 Proceedings
No full textEVALITA 2023 is an initiative of AILC (Associazione
Italiana di Linguistica Computazionale)1
and it is
endorsed by the Italian Association for Artificial
Intelligence (AIxIA)2
and the Italian Association for
Speech Sciences (AISV)3
. As in the previous editions
(http://www.evalita.it/), EVALITA 2023 is organized
along a set of selected tasks, which provide participants
with opportunities to discuss and explore both emerging
and traditional areas of Natural Language Processing
and Speech for Italian. The participation is encouraged
for teams working both in academic institutions and
industrial organizations
- …
