Nowadays, there is growing attention on enhancing the quality of teaching, learning and assessment processes. As a recent EU Report underlines, the assessment and feedback area remains a problematic issue regarding educational professionals training and adopting new practices. In fact, traditional summative assessment practices are predominantly used in European countries, against the recommendations of the Bologna Process guidelines that promote the implementation of alternative assessment practices that seem crucial in order to engage and provide lifelong learning skills for students, also with the use of technology. Looking at the literature, a series of sustainability problems arise when these requests meet real-world teaching, particularly when academic instructors face the assessment of extensive classes. With the fast advancement in Large Language Models (LLMs) and their increasing availability, affordability and capability, part of the solution to these problems might be at hand. In fact, LLMs can process large amounts of text, summarise and give feedback about it following predetermined criteria. The insights of that analysis can be used both for giving feedback to the student and helping the instructor assess the text. With the proper pedagogical and technological framework, LLMs can disengage instructors from some of the time-related sustainability issues and so from the only choice of the multiple-choice test and similar. For this reason, as a first step, we are designing and validating a theoretical framework and a teaching model for fostering the use of LLMs in assessment practice, with the approaches that can be most beneficial.
{"title":"Large Language Models for Sustainable Assessment and Feedback in Higher Education: Towards a Pedagogical and Technological Framework","authors":"Daniele Agostini, Federica Picasso","doi":"10.3233/ia-240033","DOIUrl":"https://doi.org/10.3233/ia-240033","url":null,"abstract":" Nowadays, there is growing attention on enhancing the quality of teaching, learning and assessment processes. As a recent EU Report underlines, the assessment and feedback area remains a problematic issue regarding educational professionals training and adopting new practices. In fact, traditional summative assessment practices are predominantly used in European countries, against the recommendations of the Bologna Process guidelines that promote the implementation of alternative assessment practices that seem crucial in order to engage and provide lifelong learning skills for students, also with the use of technology. Looking at the literature, a series of sustainability problems arise when these requests meet real-world teaching, particularly when academic instructors face the assessment of extensive classes. With the fast advancement in Large Language Models (LLMs) and their increasing availability, affordability and capability, part of the solution to these problems might be at hand. In fact, LLMs can process large amounts of text, summarise and give feedback about it following predetermined criteria. The insights of that analysis can be used both for giving feedback to the student and helping the instructor assess the text. With the proper pedagogical and technological framework, LLMs can disengage instructors from some of the time-related sustainability issues and so from the only choice of the multiple-choice test and similar. For this reason, as a first step, we are designing and validating a theoretical framework and a teaching model for fostering the use of LLMs in assessment practice, with the approaches that can be most beneficial.","PeriodicalId":504988,"journal":{"name":"Intelligenza Artificiale","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-07-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141831272","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Matteo Gabardi, Aurora Saibene, Francesca Gasparini, Daniele Rizzo, F. Stella
The artifacts affecting electroencephalographic (EEG) signals may undermine the correct interpretation of neural data that are used in a variety of applications spanning from diagnosis support systems to recreational brain-computer interfaces. Therefore, removing or - at least - reducing the noise content in respect to the actual brain activity data becomes of fundamental importance. However, manual removal of artifacts is not always applicable and appropriate, and sometimes the standard denoising techniques may encounter problems when dealing with noise frequency components overlapping with neural responses. In recent years, deep learning (DL) based denoising strategies have been developed to overcome these challenges and learn noise-related patterns to better discriminate actual EEG signals from artifact-related data. This study presents a novel DL-based EEG denoising model that leverages the prior knowledge on noise spectral features to adaptively compute optimal convolutional filters for multi-artifact noise removal. The proposed strategy is evaluated on a state-of-the-art benchmark dataset, namely EEGdenoiseNet, and achieves comparable to better performances in respect to other literature works considering both temporal and spectral metrics, providing a unique solution to remove muscle or ocular artifacts without needing a specific training on a particular artifact type.
{"title":"DL-based multi-artifact EEG denoising exploiting spectral information","authors":"Matteo Gabardi, Aurora Saibene, Francesca Gasparini, Daniele Rizzo, F. Stella","doi":"10.3233/ia-240025","DOIUrl":"https://doi.org/10.3233/ia-240025","url":null,"abstract":"The artifacts affecting electroencephalographic (EEG) signals may undermine the correct interpretation of neural data that are used in a variety of applications spanning from diagnosis support systems to recreational brain-computer interfaces. Therefore, removing or - at least - reducing the noise content in respect to the actual brain activity data becomes of fundamental importance. However, manual removal of artifacts is not always applicable and appropriate, and sometimes the standard denoising techniques may encounter problems when dealing with noise frequency components overlapping with neural responses. In recent years, deep learning (DL) based denoising strategies have been developed to overcome these challenges and learn noise-related patterns to better discriminate actual EEG signals from artifact-related data. This study presents a novel DL-based EEG denoising model that leverages the prior knowledge on noise spectral features to adaptively compute optimal convolutional filters for multi-artifact noise removal. The proposed strategy is evaluated on a state-of-the-art benchmark dataset, namely EEGdenoiseNet, and achieves comparable to better performances in respect to other literature works considering both temporal and spectral metrics, providing a unique solution to remove muscle or ocular artifacts without needing a specific training on a particular artifact type.","PeriodicalId":504988,"journal":{"name":"Intelligenza Artificiale","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-07-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141683890","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Informal or unpaid caregivers, commonly known as family caregivers, are responsible for providing the 80% of long-term care in Europe, which constitutes a significant portion of health and social care services offered to elderly or disabled individuals. However, the demand for informal care among the elderly is expected to outnumber available supply by 2060. The increasing decline in the caregiver-to-patient ratio is expected to lead to a substantial expansion in the integration of intelligent assistance within general care. The aim of this systematic review was to thoroughly investigate the most recent advancements in AI-enabled technologies, as well as those encompassed within the broader category of assistive technology (AT), which are designed with the primary or secondary goal to assist informal carers. The review sought to identify the specific needs that these technologies fulfill in the caregiver’s activities related to the care of older individuals, the identification of caregivers’ needs domains that are currently neglected by the existing AI-supporting technologies and ATs, as well as shedding light on the informal caregiver groups that are primarily targeted by those currently available. Three databases (Scopus, IEEE Xplore, ACM Digital Libraries) were searched. The search yielded 1002 articles, with 24 articles that met the inclusion and exclusion criteria. Our results showed that AI-powered technologies significantly facilitate ambient assisted living (AAL) applications, wherein the integration of home sensors serves to improve remote monitoring for informal caregivers. Additionally, AI solutions contribute to improve care coordination between formal and informal caregivers, that could lead to advanced telehealth assistance. However, limited research on assistive technologies like robots and mHealth apps suggests further exploration. Future AI-based solutions and assistive technologies (ATs) may benefit from a more targeted approach to appeasing specific user groups based on their informal care type. Potential areas for future research also include the integration of novel methodological approaches to improve the screening process of conventional systematic reviews through the automation of tasks using AI-powered technologies based on active learning approach.
{"title":"Fostering Artificial Intelligence-based supports for informal caregivers: a systematic review of the literature","authors":"Frida Milella, Stefania Bandini","doi":"10.3233/ia-240028","DOIUrl":"https://doi.org/10.3233/ia-240028","url":null,"abstract":"Informal or unpaid caregivers, commonly known as family caregivers, are responsible for providing the 80% of long-term care in Europe, which constitutes a significant portion of health and social care services offered to elderly or disabled individuals. However, the demand for informal care among the elderly is expected to outnumber available supply by 2060. The increasing decline in the caregiver-to-patient ratio is expected to lead to a substantial expansion in the integration of intelligent assistance within general care. The aim of this systematic review was to thoroughly investigate the most recent advancements in AI-enabled technologies, as well as those encompassed within the broader category of assistive technology (AT), which are designed with the primary or secondary goal to assist informal carers. The review sought to identify the specific needs that these technologies fulfill in the caregiver’s activities related to the care of older individuals, the identification of caregivers’ needs domains that are currently neglected by the existing AI-supporting technologies and ATs, as well as shedding light on the informal caregiver groups that are primarily targeted by those currently available. Three databases (Scopus, IEEE Xplore, ACM Digital Libraries) were searched. The search yielded 1002 articles, with 24 articles that met the inclusion and exclusion criteria. Our results showed that AI-powered technologies significantly facilitate ambient assisted living (AAL) applications, wherein the integration of home sensors serves to improve remote monitoring for informal caregivers. Additionally, AI solutions contribute to improve care coordination between formal and informal caregivers, that could lead to advanced telehealth assistance. However, limited research on assistive technologies like robots and mHealth apps suggests further exploration. Future AI-based solutions and assistive technologies (ATs) may benefit from a more targeted approach to appeasing specific user groups based on their informal care type. Potential areas for future research also include the integration of novel methodological approaches to improve the screening process of conventional systematic reviews through the automation of tasks using AI-powered technologies based on active learning approach.","PeriodicalId":504988,"journal":{"name":"Intelligenza Artificiale","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141686232","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Andrea Tinterri, Federica Pelizzari, Marilena di Padova, Francesco Palladino, Giordano Vignoli, Anna Dipace
Game-Based Learning (GBL) and its subset, Board Game-Based Learning (bGBL), are dynamic pedagogical approaches leveraging the immersive power of games to enrich the learning experience. bGBL is distinguished by its tactile and social dimensions, fostering interactive exploration, collaboration, and strategic thinking; however, its adoption is limited due to lack of preparation by teachers and educators and of pedagogical and instructional frameworks in scientific literature. Artificial intelligence (AI) tools have the potential to automate or assist instructional design, but carry significant open questions, including bias, lack of context sensitivity, privacy issues, and limited evidence. This study investigates ChatGPT as a tool for selecting board games for educational purposes, testing its reliability, accuracy, and context-sensitivity through comparison with human experts evaluation. Results show high internal consistency, whereas correlation analyses reveal moderate to high agreement with expert ratings. Contextual factors are shown to influence rankings, emphasizing the need to better understand both bGBL expert decision-making processes and AI limitations. This research provides a novel approach to bGBL, provides empirical evidence of the benefits of integrating AI into instructional design, and highlights current challenges and limitations in both AI and bGBL theory, paving the way for more effective and personalized educational experiences.
{"title":"Automating board-game based learning. A comprehensive study to assess reliability and accuracy of AI in game evaluation","authors":"Andrea Tinterri, Federica Pelizzari, Marilena di Padova, Francesco Palladino, Giordano Vignoli, Anna Dipace","doi":"10.3233/ia-240030","DOIUrl":"https://doi.org/10.3233/ia-240030","url":null,"abstract":"Game-Based Learning (GBL) and its subset, Board Game-Based Learning (bGBL), are dynamic pedagogical approaches leveraging the immersive power of games to enrich the learning experience. bGBL is distinguished by its tactile and social dimensions, fostering interactive exploration, collaboration, and strategic thinking; however, its adoption is limited due to lack of preparation by teachers and educators and of pedagogical and instructional frameworks in scientific literature. Artificial intelligence (AI) tools have the potential to automate or assist instructional design, but carry significant open questions, including bias, lack of context sensitivity, privacy issues, and limited evidence. This study investigates ChatGPT as a tool for selecting board games for educational purposes, testing its reliability, accuracy, and context-sensitivity through comparison with human experts evaluation. Results show high internal consistency, whereas correlation analyses reveal moderate to high agreement with expert ratings. Contextual factors are shown to influence rankings, emphasizing the need to better understand both bGBL expert decision-making processes and AI limitations. This research provides a novel approach to bGBL, provides empirical evidence of the benefits of integrating AI into instructional design, and highlights current challenges and limitations in both AI and bGBL theory, paving the way for more effective and personalized educational experiences.","PeriodicalId":504988,"journal":{"name":"Intelligenza Artificiale","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141687966","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Lorenzo Olearo, Fabio D'Adda, Enza Messina, Marco Cremaschi, Stefania Bandini, Francesca Gasparini
Despite the rapid development in very recent years of Artificial Intelligence models to predict poverty risk, this problem still remains an unsolved open challenge, especially from a multidimensional perspective. One of the main challenges is related to the scarcity of labelled and high-quality data for training models coupled with the lack of a general reference model to build good predictors. This results in the proposal of a variety of approaches tailored to specific contexts. This paper presents our proposal to address multidimensional poverty prediction, starting from an unlabelled dataset. We focus on the case of a fragile population, the older adults; our approach is highly flexible and can be easily adapted to various scenarios. Firstly, starting from expert knowledge, we apply a stochastic method for estimating the probability of an individual being poor, and we use this probability to identify three levels of risk. Then, we train an XGBoost classification model and exploit its tree structure to define a ranking of feature relevance. This information is used to create a new set of aggregated features representative of different poverty dimensions. An explainable novel Naive Bayes model is then trained for predicting individuals’ deprivation level in our particular domain. The capacity to identify which variables are predominantly associated with poverty among older adults offers valuable insights for policymakers and decision-makers to address poverty effectively.
{"title":"Facing multidimensional poverty in older adults: An artificial intelligence approach that reveals the variable relevance","authors":"Lorenzo Olearo, Fabio D'Adda, Enza Messina, Marco Cremaschi, Stefania Bandini, Francesca Gasparini","doi":"10.3233/ia-240027","DOIUrl":"https://doi.org/10.3233/ia-240027","url":null,"abstract":"Despite the rapid development in very recent years of Artificial Intelligence models to predict poverty risk, this problem still remains an unsolved open challenge, especially from a multidimensional perspective. One of the main challenges is related to the scarcity of labelled and high-quality data for training models coupled with the lack of a general reference model to build good predictors. This results in the proposal of a variety of approaches tailored to specific contexts. This paper presents our proposal to address multidimensional poverty prediction, starting from an unlabelled dataset. We focus on the case of a fragile population, the older adults; our approach is highly flexible and can be easily adapted to various scenarios. Firstly, starting from expert knowledge, we apply a stochastic method for estimating the probability of an individual being poor, and we use this probability to identify three levels of risk. Then, we train an XGBoost classification model and exploit its tree structure to define a ranking of feature relevance. This information is used to create a new set of aggregated features representative of different poverty dimensions. An explainable novel Naive Bayes model is then trained for predicting individuals’ deprivation level in our particular domain. The capacity to identify which variables are predominantly associated with poverty among older adults offers valuable insights for policymakers and decision-makers to address poverty effectively.","PeriodicalId":504988,"journal":{"name":"Intelligenza Artificiale","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141704853","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Mario Alviano, Danilo Amendola, Luis Angel Rodriguez Reiners
Marketplaces bring together products from multiple providers and automatically manage orders that involve several suppliers. We document the use of Answer Set Programming to automatically choose products from various warehouses within a marketplace network to fulfill a specified order. The proposed solution seamlessly adapts to various objective functions utilized at different stages of order management, leading to cost savings for customers and simplifying logistics for both the marketplace and its suppliers.
{"title":"Addressing marketplace logistic tasks in answer set programming","authors":"Mario Alviano, Danilo Amendola, Luis Angel Rodriguez Reiners","doi":"10.3233/ia-240024","DOIUrl":"https://doi.org/10.3233/ia-240024","url":null,"abstract":"Marketplaces bring together products from multiple providers and automatically manage orders that involve several suppliers. We document the use of Answer Set Programming to automatically choose products from various warehouses within a marketplace network to fulfill a specified order. The proposed solution seamlessly adapts to various objective functions utilized at different stages of order management, leading to cost savings for customers and simplifying logistics for both the marketplace and its suppliers.","PeriodicalId":504988,"journal":{"name":"Intelligenza Artificiale","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-05-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141125257","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Accurate tooth segmentation in panoramic radiographs is a useful tool for dentists to diagnose and treat dental diseases. Segmenting and labeling individual teeth in panoramic radiographs helps dentists monitor the formation of caries, detect bone loss due to periodontal disease, and determine the location and orientation of damaged teeth. It can also aid in both the planning and placement of dental implants, as well as in forensic dentistry for the identification of individuals in criminal cases or human remains. With the advancement of artificial intelligence, many deep learning-based methods are being developed and improved. Although convolutional neural networks have been extensively used in medical image segmentation, the UNet and its advanced architectures stand out for their superior segmentation capacities. This study presents four semantic segmentation UNets (Classic UNet, Attention UNet, UNet3+, and Transformer UNet) for accurate tooth segmentation in panoramic radiographs using the new Tufts Dental dataset. Each model was performed using transfer learning from ImageNet-trained VGG19 and ResNet50 models. The models achieved the best results compared to the other literature models with dice coefficients (DC) and intersection over union (IoU) of 94.64% to 96.98% and 84.27% to 94.19%, respectively. This result suggests that Unet and its variants are more suitable for segmenting panoramic radiographs and could be useful for potential dental clinical applications.
{"title":"Unleashing the potential of applied UNet architectures and transfer learning in teeth segmentation on panoramic radiographs","authors":"Rime Bouali, Oussama Mahboub, Mohamed Lazaar","doi":"10.3233/ia-230067","DOIUrl":"https://doi.org/10.3233/ia-230067","url":null,"abstract":"Accurate tooth segmentation in panoramic radiographs is a useful tool for dentists to diagnose and treat dental diseases. Segmenting and labeling individual teeth in panoramic radiographs helps dentists monitor the formation of caries, detect bone loss due to periodontal disease, and determine the location and orientation of damaged teeth. It can also aid in both the planning and placement of dental implants, as well as in forensic dentistry for the identification of individuals in criminal cases or human remains. With the advancement of artificial intelligence, many deep learning-based methods are being developed and improved. Although convolutional neural networks have been extensively used in medical image segmentation, the UNet and its advanced architectures stand out for their superior segmentation capacities. This study presents four semantic segmentation UNets (Classic UNet, Attention UNet, UNet3+, and Transformer UNet) for accurate tooth segmentation in panoramic radiographs using the new Tufts Dental dataset. Each model was performed using transfer learning from ImageNet-trained VGG19 and ResNet50 models. The models achieved the best results compared to the other literature models with dice coefficients (DC) and intersection over union (IoU) of 94.64% to 96.98% and 84.27% to 94.19%, respectively. This result suggests that Unet and its variants are more suitable for segmenting panoramic radiographs and could be useful for potential dental clinical applications.","PeriodicalId":504988,"journal":{"name":"Intelligenza Artificiale","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-04-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140698109","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：