# 1st International Workshop on Logic, Statistical and Neural AI (LoStaN 2025)
Updates here: <https://lostan-workshop.github.io/2025/> ## Workshop at [JELIA 2025](https://viam.science.tsu.ge/jelia2025/) Organized by: - [Elena Umili](mailto:um...@diag.uniroma1.it<mailto:um...@diag.uniroma1.it>), Sapienza University of Rome, Italy - [Federico Sabbatini](mailto:f.sabbati...@campus.uniurb.it<mailto:f.sabbati...@campus.uniurb.it>), University of Urbino, Italy - [Giovanni Ciatto](mailto:giovanni.cia...@unibo.it<mailto:giovanni.cia...@unibo.it>), University of Bologna, Italy - [Matteo Magnini](mailto:matteo.magn...@unibo.it<mailto:matteo.magn...@unibo.it>), University of Bologna, Italy - [Yves Lesperance](lespe...@yorku.ca<mailto:lespe...@yorku.ca>), York University, Canada ## Aim The LoStaN workshop aims to bring together researchers and practitioners interested in bridging the gap between two traditionally distinct paradigms in Artificial Intelligence: **symbolic reasoning** based on **logical knowledge**, and **data-driven** approaches grounded in **statistics** and (deep) **machine learning**. During its history, AI has seen tremendous progress from both ends of this spectrum. Symbolic methods offer interpretability, generalizability, and the ability to incorporate structured domain knowledge, while neural and statistical techniques excel in learning from vast amounts of data, dealing with uncertainty, and scaling to complex tasks. However, unifying these paradigms remains a grand challenge. The workshop aims to foster a research community that sees logical reasoning and statistical learning not as competing paradigms, but as complementary facets of intelligent behavior, and their integration not only as a technical challenge but rather as a transformative opportunity. LoStaN provides a forum for discussing foundational questions, novel methodologies, and practical applications that seek to integrate logical and statistical AI, with the aim to energize this interdisciplinary space, promote collaboration between research communities, and inspire innovative solutions to AI. ## Important Dates - Paper submission deadline: July 11th, 2025, AoE - Author Notification: July 25th, 2025, AoE - Workshop Date: __TBD__, in any case between September 1st and 5th, 2025, __GET__ * main conference dates: <https://viam.science.tsu.ge/jelia2025/dates> - Camera-ready Submission: __TBD__, in any case _after the workshop_, and only for authors who are willing to publish their paper in post-preceedings. ## Submission site Authors are invited to submit original papers in PDF format to ___[link to OpenReview to be defined]___. Further information will be available on the [CFP page](https://lostan-workshop.github.io/2025/cfp) soon. ## Submission Types We invite the following types of contribution: 1. __Full__ or __short__ papers about approaches involving logic, statistical and neural models, and their combination/integration. Full papers should be limited to __14 pages__, while short papers should be limited to __7 pages__ _(excluding references)_. 2. __Position papers__ concerning the application of logic, statistical and neural models in specific fields. Position papers should be limited to __7 pages__ _(excluding references)_. 3. __Survey papers__, concerning any of the above. Survey papers should be limited to __14 pages__ _(excluding references)_. 4. __Extended abstracts__, providing an overview of recently published papers regarding the workshop topics. Extended abstract should be limited to __3 pages__ _(excluding references)_. ## Relevant topics We welcome contributions on theoretical foundations, practical applications, and interdisciplinary approaches combining symbolic, statistical, and neural methods in AI. Topics of interest include, but are not limited to: - Integration of logic and statistics into machine learning models - Explainable AI (XAI) and interpretable machine learning via symbolic methods - Symbolic knoweldge extraction and injection from/into machine learning models - Neuro-symbolig approaches to AI - Statistical relational learning and probabilistic logic - Learning symbolic representations from data - Inductive Logic Programming (ILP) and relational learning - Causal reasoning and logic-based causal discovery - Neuro-symbolic integration and hybrid AI systems - Fairness, accountability, and transparency in hybrid AI systems - Logic and learning for multi-agent systems - Human-centric AI: reasoning, learning, and interaction - Deep learning and knowledge extraction - Formal methods for the verification of neural networks - Benchmarking and evaluation of hybrid AI systems - Applications in robotics, natural language understanding, bioinformatics, etc. - Toolkits and platforms for integrating logic, statistics, and neural networks ## Submission Guidelines - Submissions should be in [__CEUR Workshop Proceedings__ style](https://it.overleaf.com/latex/templates/template-for-submissions-to-ceur-workshop-proceedings-ceur-ws-dot-org/wqyfdgftmcfw); - Submissions must be exported in __PDF format__; - _Camera-ready_ versions shall include the LaTeX __source files__; - Submissions are *not* anonymous __(single-blind)__. ## Reviewing Format Each submission will undergo a traditional single-blind review process. This phase applies to all papers, regardless of type, and will determine the final acceptance or rejection decision. Accepted contributions are presented at the workshop, and authors may collect further comments and insights. After that, authors who are willing to publish their contribution to the workshop proceedings will be requested to produce a final version of the paper, addressing the reviewers’ suggestions and possibly integrating comments and insights from the workshop. All papers will undergo the standard assessment in which reviewers will be asked to consider the submission focusing on its: - __scope__: is the paper on-topic w.r.t. the workshop’s track or theme? - __significance__: is the idea proposed in the paper meaningful for the LoStaN community? Are the results relevant? - __soundness__: is the approach proposed in the paper correct and robust? Are experiments (if any) well-designed? - __clarity__: is the paper clear and well organized? Is the discussion complete? - __contextualization__: is relevant background and related literature being adequately referenced? - __novelty__: is the contribution novel either from a conceptual or technical perspective? Furthermore, the reviews will take into account the specific characteristics of the submitted contributions (see submission types above) to obtain their final decision. ## Proceedings The workshop will publish its own proceedings onto Scopus-indexed repositories, such as [CEUR-WS](https://ceur-ws.org/). > __Important note:__ Accepted papers will be included in the workshop > proceedings __only if__ the authors explicitly agree to do so. In this way, works which are not mature enough for publication, as well as works which have already been published elsewhere can be presented and discussed at the workshop. ## 📖 Special Issue A selection of the best papers presented at the workshop will be invited to submit an extended version to a special issue in a journal (TBD). We are negotiating with several journals, and we will update this CfP as soon as we have more information. -- [LOGIC] mailing list, provided by DLMPST More information (including information about subscription management) can be found here: http://dlmpst.org/pages/logic-list.php
