<div dir="ltr"><div class="gmail_default" style="font-family:monospace"><div class="gmail_default">[APOLOGIZE FOR MULTIPLE POSTINGS]</div><div class="gmail_default"><br></div><div class="gmail_default">=</div><div class="gmail_default"><br></div><div class="gmail_default">It's a pleasure to announce the OPEN special issue of the <b>Journal of Biomedical Informatics</b>, published by <b>ELSEVIER </b>(<i>ISSN: 1532-0480</i>) entitled: "<b>learning from multiple data sources for decision making in health care</b>". </div><div class="gmail_default"><br></div><div class="gmail_default">Deadline is <b>15 October, 2024</b>.</div><div class="gmail_default"><br></div><div class="gmail_default">For your convenience, some details are reported below; you can find the full call here, along with directions, here: </div><div class="gmail_default">    <a href="https://www.sciencedirect.com/science/article/pii/S1532046424000637" target="_blank">https://www.sciencedirect.com/science/article/pii/S1532046424000637</a></div><div class="gmail_default"><br></div><div class="gmail_default">Please forward this to your colleagues and collaborators, and anyone potentially interested; this issue is going to build a road towards one of the main goals of the HC@AIxIA working group, namely fostering an effective application of AI to medicine and the healthcare domain. </div><div class="gmail_default"><br></div><div class="gmail_default">Feel free to contact us at </div><div class="gmail_default">    <a href="mailto:hc-aixia@googlegroups.com" target="_blank">hc-aixia@googlegroups.com</a></div><div class="gmail_default">and visit </div><div class="gmail_default">    <a href="https://aixia.it/en/gruppi/hc/" target="_blank">https://aixia.it/en/gruppi/hc/</a></div><div class="gmail_default">for further information about the group activities and initiatives.</div><div class="gmail_default"><br></div><div class="gmail_default">Sincerely,</div><div class="gmail_default">Francesco Calimeri, Mauro Dragoni, Fabio Stella</div><div class="gmail_default">Coordinators of the Working Group on Artificial Intelligence for Healthcare</div><div class="gmail_default"><br></div><div class="gmail_default"><br></div><div class="gmail_default"><br></div><div class="gmail_default">=== === === === === === === === ===</div><div class="gmail_default"><br></div><div class="gmail_default"><br></div><div class="gmail_default" style="font-family:Arial,Helvetica,sans-serif"><h2 id="m_-4712258747280641771gmail-publication-title" style="box-sizing:border-box;margin:0px;padding:0px;color:rgb(31,31,31);font-family:ElsevierSans,Arial,Helvetica,Roboto,"Lucida Sans Unicode","Microsoft Sans Serif","Segoe UI Symbol",STIXGeneral,"Cambria Math","Arial Unicode MS",sans-serif;text-align:center;font-weight:400;font-size:1.2rem"><a href="https://www.sciencedirect.com/journal/journal-of-biomedical-informatics" title="Go to Journal of Biomedical Informatics on ScienceDirect" target="_blank" style="color:rgb(31,31,31);box-sizing:border-box;margin:0px;padding:0px;background-color:transparent;word-break:break-word;text-decoration-line:none;display:inline-block"><span style="box-sizing:border-box;margin:0px;padding:0px;border-bottom:2px solid transparent">Journal of Biomedical Informatics</span></a></h2><h1 id="m_-4712258747280641771gmail-screen-reader-main-title" style="font-family:ElsevierGulliver,Georgia,"Times New Roman",Times,STIXGeneral,"Cambria Math","Lucida Sans Unicode","Microsoft Sans Serif","Segoe UI Symbol","Arial Unicode MS",serif;box-sizing:border-box;margin:16px 0px;padding:0px;word-break:break-word;color:rgb(31,31,31);font-weight:400"><span style="box-sizing:border-box;margin:0px;padding:0px"><font size="4">Special issue on learning from multiple data sources for decision making in health care</font></span></h1><div id="m_-4712258747280641771gmail-banner" style="box-sizing:border-box;margin:0px 0px 8px;padding:0px;color:rgb(31,31,31)"><div style="box-sizing:border-box;margin:0px;padding:0px"><div style="box-sizing:border-box;margin:0px;padding:0px"><div id="m_-4712258747280641771gmail-author-group" style="box-sizing:border-box;margin:0px;padding:0px"><span style="font-family:ElsevierGulliver,Georgia,"Times New Roman",Times,STIXGeneral,"Cambria Math","Lucida Sans Unicode","Microsoft Sans Serif","Segoe UI Symbol","Arial Unicode MS",serif;font-size:0.8rem">The increasing availability of digital data, along with recent developments in Artificial Intelligence, especially in the Machine Learning and Deep Learning fields, led the scientific community to debate whether data alone is sufficient for decision making and scientific exploration. We focus the attention on the healthcare domain, where peculiar issues affect data: indeed, data are usually collected under heterogeneous conditions (i.e., different populations, regimes, and sampling methods), suffer missingness – very often not at random – and their use is strongly constrained by privacy issues. In such a complex setting, this special issue challenges computer scientists to contribute to the above debate by designing and developing innovative methodological approaches, for solving complex decision-making problems in health care, leveraging on observational data.</span><br></div></div></div></div><div id="m_-4712258747280641771gmail-body" style="font-family:ElsevierGulliver,Georgia,"Times New Roman",Times,STIXGeneral,"Cambria Math","Lucida Sans Unicode","Microsoft Sans Serif","Segoe UI Symbol","Arial Unicode MS",serif;box-sizing:border-box;margin:0px;padding:0px;font-size:0.8rem;color:rgb(31,31,31)"><div style="box-sizing:border-box;margin:0px;padding:0px"><p id="m_-4712258747280641771gmail-p0010" style="box-sizing:border-box;margin:0px 0px 16px;padding:0px">Topics of interest include, but are not limited to, the following with an emphasis on novel generalizable methods applied to the healthcare domain:</p><ul style="box-sizing:border-box;margin:0px 0px 24px;padding:0px;list-style:none"><li style="margin:0px;box-sizing:border-box;padding:0px;display:flex"><span style="box-sizing:border-box;margin:0px;padding:0px;width:24px">•</span><span style="box-sizing:border-box;margin:0px;padding:0px"><p id="m_-4712258747280641771gmail-p0015" style="box-sizing:border-box;margin:0px 0px 16px;padding:0px">Causal discovery from multiple data sets.</p></span></li><li style="margin:0px;box-sizing:border-box;padding:0px;display:flex"><span style="box-sizing:border-box;margin:0px;padding:0px;width:24px">•</span><span style="box-sizing:border-box;margin:0px;padding:0px"><p id="m_-4712258747280641771gmail-p0020" style="box-sizing:border-box;margin:0px 0px 16px;padding:0px">Federated causal discovery.</p></span></li><li style="margin:0px;box-sizing:border-box;padding:0px;display:flex"><span style="box-sizing:border-box;margin:0px;padding:0px;width:24px">•</span><span style="box-sizing:border-box;margin:0px;padding:0px"><p id="m_-4712258747280641771gmail-p0025" style="box-sizing:border-box;margin:0px 0px 16px;padding:0px">Causal discovery from heterogeneous data sets.</p></span></li><li style="margin:0px;box-sizing:border-box;padding:0px;display:flex"><span style="box-sizing:border-box;margin:0px;padding:0px;width:24px">•</span><span style="box-sizing:border-box;margin:0px;padding:0px"><p id="m_-4712258747280641771gmail-p0030" style="box-sizing:border-box;margin:0px 0px 16px;padding:0px">Transportability of causal models and inference.</p></span></li><li style="margin:0px;box-sizing:border-box;padding:0px;display:flex"><span style="box-sizing:border-box;margin:0px;padding:0px;width:24px">•</span><span style="box-sizing:border-box;margin:0px;padding:0px"><p id="m_-4712258747280641771gmail-p0035" style="box-sizing:border-box;margin:0px 0px 16px;padding:0px">Neuro-symbolic approaches to learn from heterogeneous data sources.</p></span></li><li style="margin:0px;box-sizing:border-box;padding:0px;display:flex"><span style="box-sizing:border-box;margin:0px;padding:0px;width:24px">•</span><span style="box-sizing:border-box;margin:0px;padding:0px"><p id="m_-4712258747280641771gmail-p0040" style="box-sizing:border-box;margin:0px 0px 16px;padding:0px">Continual learning on streams from multiple data sources.</p></span></li><li style="margin:0px;box-sizing:border-box;padding:0px;display:flex"><span style="box-sizing:border-box;margin:0px;padding:0px;width:24px">•</span><span style="box-sizing:border-box;margin:0px;padding:0px"><p id="m_-4712258747280641771gmail-p0045" style="box-sizing:border-box;margin:0px 0px 16px;padding:0px">Computational intelligent strategies to support causal inference.</p></span></li><li style="margin:0px;box-sizing:border-box;padding:0px;display:flex"><span style="box-sizing:border-box;margin:0px;padding:0px;width:24px">•</span><span style="box-sizing:border-box;margin:0px;padding:0px"><p id="m_-4712258747280641771gmail-p0050" style="box-sizing:border-box;margin:0px 0px 16px;padding:0px">Edge computing for decision making in healthcare.</p></span></li><li style="margin:0px;box-sizing:border-box;padding:0px;display:flex"><span style="box-sizing:border-box;margin:0px;padding:0px;width:24px">•</span><span style="box-sizing:border-box;margin:0px;padding:0px"><p id="m_-4712258747280641771gmail-p0055" style="box-sizing:border-box;margin:0px 0px 16px;padding:0px">Integrative AI methodologies.</p></span></li><li style="margin:0px;box-sizing:border-box;padding:0px;display:flex"><span style="box-sizing:border-box;margin:0px;padding:0px;width:24px">•</span><span style="box-sizing:border-box;margin:0px;padding:0px"><p id="m_-4712258747280641771gmail-p0060" style="box-sizing:border-box;margin:0px 0px 16px;padding:0px">Distributed inference methods.</p></span></li><li style="margin:0px;box-sizing:border-box;padding:0px;display:flex"><span style="box-sizing:border-box;margin:0px;padding:0px;width:24px">•</span><span style="box-sizing:border-box;margin:0px;padding:0px"><p id="m_-4712258747280641771gmail-p0065" style="box-sizing:border-box;margin:0px 0px 16px;padding:0px">Continual Learning.</p></span></li><li style="margin:0px;box-sizing:border-box;padding:0px;display:flex"><span style="box-sizing:border-box;margin:0px;padding:0px;width:24px">•</span><span style="box-sizing:border-box;margin:0px;padding:0px"><p id="m_-4712258747280641771gmail-p0070" style="box-sizing:border-box;margin:0px 0px 16px;padding:0px">Knowledge Discovery and Integration.</p></span></li><li style="margin:0px;box-sizing:border-box;padding:0px;display:flex"><span style="box-sizing:border-box;margin:0px;padding:0px;width:24px">•</span><span style="box-sizing:border-box;margin:0px;padding:0px"><p id="m_-4712258747280641771gmail-p0075" style="box-sizing:border-box;margin:0px 0px 16px;padding:0px">Combination of deductive approaches with ML models.</p></span></li><li style="margin:0px;box-sizing:border-box;padding:0px;display:flex"><span style="box-sizing:border-box;margin:0px;padding:0px;width:24px">•</span><span style="box-sizing:border-box;margin:0px;padding:0px"><p id="m_-4712258747280641771gmail-p0080" style="box-sizing:border-box;margin:0px;padding:0px">Combination of ontologies and/or knowledge-bases with ML to support decision making.</p></span></li></ul><p></p><p id="m_-4712258747280641771gmail-p0085" style="box-sizing:border-box;margin:0px 0px 16px;padding:0px"><span style="box-sizing:border-box;margin:0px;padding:0px;font-weight:bolder">Peer Review Process:</span></p><p id="m_-4712258747280641771gmail-p0090" style="box-sizing:border-box;margin:0px 0px 16px;padding:0px">All submitted papers will undergo a rigorous peer-review process featuring at least two reviewers. All submissions should follow the guidelines for authors available at the Journal of Biomedical Informatics website (<a href="http://www.elsevier.com/locate/yjbin" rel="noreferrer noopener" target="_blank" style="color:rgb(2,114,177);box-sizing:border-box;margin:0px;padding:0px;background-color:transparent;word-break:break-word;text-decoration-line:none;display:inline"><span style="box-sizing:border-box;margin:0px;padding:0px;border-bottom:2px solid transparent">http://www.elsevier.com/locate/yjbin</span></a>). JBI’s editorial policy outlined on that page will be strictly enforced by special issue reviewers.</p><p id="m_-4712258747280641771gmail-p0095" style="box-sizing:border-box;margin:0px 0px 16px;padding:0px">Note that JBI emphasizes the publication of papers that introduce innovative and generalizable methods of interest to the informatics community. Specific applications can be described to motivate the methodology being introduced, but papers that focus solely on a specific application are not suitable. A few examples of papers focused on methods previously published in JBI include: Kyrimi, et al. <a href="https://www.sciencedirect.com/science/article/pii/S1532046424000637#b0005" name="m_-4712258747280641771_bb0005" target="_blank" style="color:rgb(2,114,177);box-sizing:border-box;margin:0px;padding:0px;background-color:transparent;word-break:break-word;text-decoration-line:none;display:inline"><span style="box-sizing:border-box;margin:0px;padding:0px;border-bottom:2px solid transparent">[1]</span></a>, Huang, et al. <a href="https://www.sciencedirect.com/science/article/pii/S1532046424000637#b0010" name="m_-4712258747280641771_bb0010" target="_blank" style="color:rgb(2,114,177);box-sizing:border-box;margin:0px;padding:0px;background-color:transparent;word-break:break-word;text-decoration-line:none;display:inline"><span style="box-sizing:border-box;margin:0px;padding:0px;border-bottom:2px solid transparent">[2]</span></a>, Kocbek et al. <a href="https://www.sciencedirect.com/science/article/pii/S1532046424000637#b0015" name="m_-4712258747280641771_bb0015" target="_blank" style="color:rgb(2,114,177);box-sizing:border-box;margin:0px;padding:0px;background-color:transparent;word-break:break-word;text-decoration-line:none;display:inline"><span style="box-sizing:border-box;margin:0px;padding:0px;border-bottom:2px solid transparent">[3]</span></a>, Houston et al. <a href="https://www.sciencedirect.com/science/article/pii/S1532046424000637#b0020" name="m_-4712258747280641771_bb0020" target="_blank" style="color:rgb(2,114,177);box-sizing:border-box;margin:0px;padding:0px;background-color:transparent;word-break:break-word;text-decoration-line:none;display:inline"><span style="box-sizing:border-box;margin:0px;padding:0px;border-bottom:2px solid transparent">[4]</span></a>, García Del Valle et al. <a href="https://www.sciencedirect.com/science/article/pii/S1532046424000637#b0025" name="m_-4712258747280641771_bb0025" target="_blank" style="color:rgb(2,114,177);box-sizing:border-box;margin:0px;padding:0px;background-color:transparent;word-break:break-word;text-decoration-line:none;display:inline"><span style="box-sizing:border-box;margin:0px;padding:0px;border-bottom:2px solid transparent">[5]</span></a>, Graudenzi et al. <a href="https://www.sciencedirect.com/science/article/pii/S1532046424000637#b0030" name="m_-4712258747280641771_bb0030" target="_blank" style="color:rgb(2,114,177);box-sizing:border-box;margin:0px;padding:0px;background-color:transparent;word-break:break-word;text-decoration-line:none;display:inline"><span style="box-sizing:border-box;margin:0px;padding:0px;border-bottom:2px solid transparent">[6]</span></a> and Sims et al. <a href="https://www.sciencedirect.com/science/article/pii/S1532046424000637#b0035" name="m_-4712258747280641771_bb0035" target="_blank" style="color:rgb(2,114,177);box-sizing:border-box;margin:0px;padding:0px;background-color:transparent;word-break:break-word;text-decoration-line:none;display:inline"><span style="box-sizing:border-box;margin:0px;padding:0px;border-bottom:2px solid transparent">[7]</span></a>.</p><p id="m_-4712258747280641771gmail-p0100" style="box-sizing:border-box;margin:0px 0px 16px;padding:0px">In particular, the authors of <a href="https://www.sciencedirect.com/science/article/pii/S1532046424000637#b0005" name="m_-4712258747280641771_bb0005" target="_blank" style="color:rgb(2,114,177);box-sizing:border-box;margin:0px;padding:0px;background-color:transparent;word-break:break-word;text-decoration-line:none;display:inline"><span style="box-sizing:border-box;margin:0px;padding:0px;border-bottom:2px solid transparent">[1]</span></a> showed the relevance of causal models and expert knowledge to develop credible models, i.e., capable of achieving good predictive performances when transported from the study cohort to the target population. Furthermore, <a href="https://www.sciencedirect.com/science/article/pii/S1532046424000637#b0010" name="m_-4712258747280641771_bb0010" target="_blank" style="color:rgb(2,114,177);box-sizing:border-box;margin:0px;padding:0px;background-color:transparent;word-break:break-word;text-decoration-line:none;display:inline"><span style="box-sizing:border-box;margin:0px;padding:0px;border-bottom:2px solid transparent">[2]</span></a> tackles the relevant issue of partially overlapping variables when data are collected from multiple data sources. This problem is extremely relevant both in theoretical and practical terms for decision making in the healthcare sector.</p><p id="m_-4712258747280641771gmail-p0105" style="box-sizing:border-box;margin:0px 0px 16px;padding:0px">The contribution provided in <a href="https://www.sciencedirect.com/science/article/pii/S1532046424000637#b0015" name="m_-4712258747280641771_bb0015" target="_blank" style="color:rgb(2,114,177);box-sizing:border-box;margin:0px;padding:0px;background-color:transparent;word-break:break-word;text-decoration-line:none;display:inline"><span style="box-sizing:border-box;margin:0px;padding:0px;border-bottom:2px solid transparent">[3]</span></a> stressed the importance of working in a multi-source context by demonstrating how the linking of different repositories can improve the overall understanding of patients' conditions. Similarly, in <a href="https://www.sciencedirect.com/science/article/pii/S1532046424000637#b0020" name="m_-4712258747280641771_bb0020" target="_blank" style="color:rgb(2,114,177);box-sizing:border-box;margin:0px;padding:0px;background-color:transparent;word-break:break-word;text-decoration-line:none;display:inline"><span style="box-sizing:border-box;margin:0px;padding:0px;border-bottom:2px solid transparent">[4]</span></a> the authors extended this concept by introducing a methodology to evaluate to audit the data quality of the sources exploited by healthcare information systems. Then, in <a href="https://www.sciencedirect.com/science/article/pii/S1532046424000637#b0025" name="m_-4712258747280641771_bb0025" target="_blank" style="color:rgb(2,114,177);box-sizing:border-box;margin:0px;padding:0px;background-color:transparent;word-break:break-word;text-decoration-line:none;display:inline"><span style="box-sizing:border-box;margin:0px;padding:0px;border-bottom:2px solid transparent">[5]</span></a> the multi-source concept is transferred within the multi-modal environment and the authors surveyed the importance of considering different modalities to obtain a better disease understanding.</p><p id="m_-4712258747280641771gmail-p0110" style="box-sizing:border-box;margin:0px 0px 16px;padding:0px">The works in <a href="https://www.sciencedirect.com/science/article/pii/S1532046424000637#b0030" name="m_-4712258747280641771_bb0030" target="_blank" style="color:rgb(2,114,177);box-sizing:border-box;margin:0px;padding:0px;background-color:transparent;word-break:break-word;text-decoration-line:none;display:inline"><span style="box-sizing:border-box;margin:0px;padding:0px;border-bottom:2px solid transparent">[6]</span></a> and <a href="https://www.sciencedirect.com/science/article/pii/S1532046424000637#b0035" name="m_-4712258747280641771_bb0035" target="_blank" style="color:rgb(2,114,177);box-sizing:border-box;margin:0px;padding:0px;background-color:transparent;word-break:break-word;text-decoration-line:none;display:inline"><span style="box-sizing:border-box;margin:0px;padding:0px;border-bottom:2px solid transparent">[7]</span></a> focuses on the importance of data. In <a href="https://www.sciencedirect.com/science/article/pii/S1532046424000637#b0030" name="m_-4712258747280641771_bb0030" target="_blank" style="color:rgb(2,114,177);box-sizing:border-box;margin:0px;padding:0px;background-color:transparent;word-break:break-word;text-decoration-line:none;display:inline"><span style="box-sizing:border-box;margin:0px;padding:0px;border-bottom:2px solid transparent">[6]</span></a> a data integration framework is defined for characterizing the metabolic deregulations that distinguish cancer phenotypes, by projecting RNA-seq data onto metabolic networks without the need for metabolic measurements; in <a href="https://www.sciencedirect.com/science/article/pii/S1532046424000637#b0035" name="m_-4712258747280641771_bb0035" target="_blank" style="color:rgb(2,114,177);box-sizing:border-box;margin:0px;padding:0px;background-color:transparent;word-break:break-word;text-decoration-line:none;display:inline"><span style="box-sizing:border-box;margin:0px;padding:0px;border-bottom:2px solid transparent">[7]</span></a> a biomedical informatics method is introduced that uses multiple public health data sources to perform surveillance of methadone-related adverse drug events. Interestingly, even if patient data are not linked between different data sources, results show that the integration of multiple public data sources can capture more cases and provide more clinical details than individual data sources alone.</p><p id="m_-4712258747280641771gmail-p0115" style="box-sizing:border-box;margin:0px 0px 16px;padding:0px">Key requirements for JBI ML papers in addition to presenting novel methods (not simply application of existing methods to a new healthcare domain) are as follows: 1) projects must have clinicians involved in research question/problem formulation, defining input data, and assessing the results. 2) An explanation (with clinicians) of how the proposed method would fit into the clinical workflow is expected. It must be translational to practice. 3) Data sets should preferably be collected from hospitals after the research question was formulated, thus avoiding the use of available data (MIMIC) to define a very wide research problem that could potentially be answered with available open datasets (as an example: detecting if someone has COVID from Chest X-Rays would not be acceptable, as the gold standard test is the laboratory test). 4) As for explainability, SHAP values and related diagrams would not be enough: the paper should clearly describe and explain how clinicians use the visualization to make decisions. For further details please refer to <a href="https://www.sciencedirect.com/journal/journal-of-biomedical-informatics/publish/guide-for-authors" rel="noreferrer noopener" target="_blank" style="color:rgb(2,114,177);box-sizing:border-box;margin:0px;padding:0px;background-color:transparent;word-break:break-word;text-decoration-line:none;display:inline"><span style="box-sizing:border-box;margin:0px;padding:0px;border-bottom:2px solid transparent">https://www.sciencedirect.com/journal/journal-of-biomedical-informatics/publish/guide-for-authors</span></a>.</p><p id="m_-4712258747280641771gmail-p0120" style="box-sizing:border-box;margin:0px 0px 16px;padding:0px"><span style="box-sizing:border-box;margin:0px;padding:0px;font-weight:bolder">Submission process:</span></p><p id="m_-4712258747280641771gmail-p0125" style="box-sizing:border-box;margin:0px 0px 16px;padding:0px">Authors must submit their paper via the online Elsevier Editorial System (EES) at <a href="http://ees.elsevier.com/jbi" target="_blank">http://ees.elsevier.com/jbi</a> by October 15th, 2024. Authors can register and upload their text, tables, and figures as well as subsequent revisions through this website. Potential authors may contact the Publishing Services Coordinator in the journal’s editorial office (<a href="http://jbi%40elsevier.com/" rel="noreferrer noopener" target="_blank" style="color:rgb(2,114,177);box-sizing:border-box;margin:0px;padding:0px;background-color:transparent;word-break:break-word;text-decoration-line:none;display:inline"><span style="box-sizing:border-box;margin:0px;padding:0px;border-bottom:2px solid transparent">jbi@elsevier.com</span></a>) for questions regarding this process. When asked for the category of their submission, they should indicate that it is for the special issue on Learning from multiple data sources for decision making in health care.</p></div></div></div></div></div>