{"id":85,"date":"2025-11-01T19:49:19","date_gmt":"2025-11-01T18:49:19","guid":{"rendered":"https:\/\/cornerstone.web.roma2.infn.it\/?p=85"},"modified":"2025-11-01T19:50:10","modified_gmt":"2025-11-01T18:50:10","slug":"cornerstone-at-esww-2025-dr-chierichini-presents-attention-based-flare-forecasting","status":"publish","type":"post","link":"https:\/\/cornerstone.web.roma2.infn.it\/?p=85","title":{"rendered":"CORNERSTONE at ESWW 2025: Dr. Chierichini Presents Attention-Based Flare Forecasting"},"content":{"rendered":"\n

Bringing CORNERSTONE Research to the International Stage<\/strong><\/p>\n\n\n\n

We’re proud to share that Dr. Simone Chierichini<\/strong> presented CORNERSTONE’s cutting-edge research at the European Space Weather Week (ESWW) 2025<\/strong> in Ume\u00e5, Sweden (October 27-31, 2025). His contribution showcased the project’s innovative approach to solar flare forecasting using state-of-the-art deep learning techniques.<\/p>\n\n\n\n

The Presentation: Attention-Based Deep Learning for Operational Forecasting<\/h2>\n\n\n\n

Dr. Chierichini’s presentation focused on employing attention-based deep learning<\/strong> to predict solar flare occurrence from multivariate time series of SHARP magnetogram features\u2014the same carefully curated data prepared by Dr. Stefano Scardigli as part of the CORNERSTONE project.<\/p>\n\n\n\n

Key Research Highlights<\/h3>\n\n\n\n

Temporal Investigation<\/strong>: One of the most important questions in flare forecasting is: how much history do we need to see?<\/em> Dr. Chierichini’s work systematically investigated how the temporal extent of input data affects forecasting skill by evaluating the model on sequences ranging from short snapshots to full 24-hour histories. This comparative analysis helps determine whether longer observation windows allow the model to recognize precursor patterns that may not be evident over shorter intervals.<\/p>\n\n\n\n

Advanced Architecture<\/strong>: The model leverages self-attention mechanisms and learnable positional embeddings to capture the temporal evolution of active regions. This enables it to model dependencies across time\u2014identifying which past observations are most relevant for predicting future flare activity\u2014even in the presence of missing data.<\/p>\n\n\n\n

Operational Readiness<\/strong>: A critical innovation is the model’s resilience to real-world data challenges. Space weather monitoring often faces incomplete or irregular time series due to data gaps, satellite issues, or processing delays. Dr. Chierichini’s approach incorporates a masking mechanism that guides the attention layers to focus only on valid observations, maintaining predictive capability even when data is imperfect. This makes the system suitable for operational settings where data quality can be variable.<\/p>\n\n\n\n

Multi-Class Prediction<\/strong>: The model outputs probabilistic predictions for C+-, M+-, and X+-class flare events, providing forecasters with not just binary yes\/no predictions, but graded likelihood assessments for different flare intensities.<\/p>\n\n\n\n

Why ESWW Matters<\/h2>\n\n\n\n

The European Space Weather Week is the premier annual event in the European space weather calendar, bringing together scientists, engineers, satellite operators, forecasters, and end-users from around the world. Presenting at ESWW 2025 placed CORNERSTONE’s research before the international community actively working to protect critical infrastructure from space weather impacts.<\/p>\n\n\n\n

The conference’s 2025 theme\u2014”Technological expansion of the Arctic: the new frontiers of space weather”\u2014emphasized the growing importance of space weather monitoring and forecasting as technology increasingly extends into polar regions and space.<\/p>\n\n\n\n

The CORNERSTONE Advantage: From Data to Deployment<\/h2>\n\n\n\n

Dr. Chierichini’s presentation exemplifies the strength of the CORNERSTONE approach: combining rigorous data preparation (Dr. Scardigli’s curated SHARP datasets) with sophisticated machine learning techniques (Dr. Chierichini’s attention-based architectures) to create robust, operationally-relevant forecasting tools.<\/p>\n\n\n\n

The work demonstrates that CORNERSTONE is not just advancing academic research\u2014it’s building systems designed for real-world deployment. The focus on handling missing data, investigating optimal time windows, and providing probabilistic forecasts addresses the practical needs of operational space weather forecasting centers.<\/p>\n\n\n\n

Moving Toward Real-Time Forecasting<\/h2>\n\n\n\n

As Dr. Chierichini concluded in his presentation, this approach represents “a step toward developing real-time forecasting tools that are resilient to data quality variability.” This is precisely CORNERSTONE’s mission: bridging the gap between fundamental solar physics research and operational space weather services that protect society’s technological infrastructure.<\/p>\n\n\n\n

The positive reception at ESWW 2025 and the ongoing collaboration between team members demonstrates that CORNERSTONE is making significant progress toward this goal. By combining expertise in data science, machine learning, and solar physics, the project is advancing the state-of-the-art in solar flare forecasting.<\/p>\n\n\n\n

We congratulate Dr. Chierichini on his excellent presentation and look forward to seeing this research continue to develop and eventually transition to operational use!<\/p>\n\n\n\n

\n\n\n\n

CORNERSTONE is funded under MUR – PRIN 2022 PNRR (P2022RKXH9 – CUP: E53D23021410001)<\/em><\/p>\n\n\n\n

Dr. Chierichini’s work was conducted in collaboration with colleagues from University of Rome Tor Vergata, University of Coimbra, University of Sheffield, and University of Trento, demonstrating the international collaborative nature of the CORNERSTONE project.<\/em><\/p>\n","protected":false},"excerpt":{"rendered":"

Bringing CORNERSTONE Research to the International Stage We’re proud to share that Dr. Simone Chierichini presented CORNERSTONE’s cutting-edge research at the European Space Weather Week (ESWW) 2025 in Ume\u00e5, Sweden (October 27-31, 2025). His contribution showcased the project’s innovative approach to solar flare forecasting using state-of-the-art deep learning techniques. The Presentation: Attention-Based Deep Learning for […]<\/p>\n","protected":false},"author":1,"featured_media":86,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[11,4],"tags":[12,5,8],"class_list":["post-85","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-data","category-project-action","tag-data","tag-flare","tag-project"],"_links":{"self":[{"href":"https:\/\/cornerstone.web.roma2.infn.it\/index.php?rest_route=\/wp\/v2\/posts\/85","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/cornerstone.web.roma2.infn.it\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/cornerstone.web.roma2.infn.it\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/cornerstone.web.roma2.infn.it\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/cornerstone.web.roma2.infn.it\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=85"}],"version-history":[{"count":1,"href":"https:\/\/cornerstone.web.roma2.infn.it\/index.php?rest_route=\/wp\/v2\/posts\/85\/revisions"}],"predecessor-version":[{"id":87,"href":"https:\/\/cornerstone.web.roma2.infn.it\/index.php?rest_route=\/wp\/v2\/posts\/85\/revisions\/87"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/cornerstone.web.roma2.infn.it\/index.php?rest_route=\/wp\/v2\/media\/86"}],"wp:attachment":[{"href":"https:\/\/cornerstone.web.roma2.infn.it\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=85"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/cornerstone.web.roma2.infn.it\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=85"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/cornerstone.web.roma2.infn.it\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=85"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}