Publications

• Genovese D., Ploner P., Scardapane S., Giagu S., Aarrestad T. K. (2026).
  Local Diagnostics of Learned Event Representations for Anomaly Detection at the HL-LHC.
  arXiv preprint arXiv:2610.13437 (Submitted to SciPost).

  Evaluation of supervised and self-supervised neural embeddings on the 750-million-event COLLIDE-2V dataset for autoencoder-based anomaly detection under High-Luminosity LHC conditions, with local diagnostics to recover physical interpretability of the learned representations.
  Download Paper (PDF)

• Fontanari T. V., Rulli M. E., Alvetreti F., Genovese D., Scardapane S. (2026).
  LRCC: Generalizing Low-Rank Compression with Conditional Computation.
  arXiv preprint arXiv:2610.08858.

  Low-Rank Conditional Computation (LRCC): lightweight per-block routers select among nested low-rank paths, adding token-dependent compute to compressed pretrained LLMs (Llama, Qwen).
  Download Paper (PDF)

• Capriotti S., Genovese D., Devoto A., Badreshany K., Braekmans D., Mignardi S., Scardapane S., Medeghini L. (2026).
  Comparative evaluation of CNN and ViT architectures for the petrographic classification of Levantine ceramic micrographs.
  ACTA IMEKO, 15(3).
  DOI: 10.21014/tc26-2025.045

  Comparative evaluation of Convolutional Neural Networks (CNN) and Vision Transformers (ViT) for petrographic classification of thin-section Levantine ceramics.
  Download Paper (PDF)

• Oddi L., Scardapane S., Sugimoto T., Genovese D. (2026).
  Authorship attribution and aesthetic evaluation of AI poetry: a case study with Haiku.
  arXiv preprint arXiv:2609.15511 (Accepted at SCIS&ISIS 2026).

  Empirical study on human authorship perception versus aesthetic evaluation of Japanese haiku generated by modern Large Language Models (LLMs).
  Download Paper (PDF)

• Bernárdez G., Telyatnikov L., Papillon M., Montagna M., Theiler R., Cornelis L., Mathe J., Ferriol M., Vasylenko P., Van Looy J.-W., Testa L., Neri B., Genovese D., Weber M., Wei A., Devoto A., et al. (2026).
  Topological Deep Learning Challenge 2025: Expanding the Data Landscape.
  Proceedings of the 1st Conference on Topology, Algebra, and Geometry in Data Science (TAG-DS 2025), PMLR 321:4-14.

  Summary and key findings of the 2025 TDL Challenge, introducing 44 new dataset loaders, benchmark tasks, and scalable data pipelines for topological deep learning.
  Download Paper (PDF)

• Genovese D., Devoto A., Carmignani J., Sebastiani C., Scardapane S., D’Onofrio M. (2026).
  Graph transformer and mixture of experts for rare signal detection in calorimeter data.
  Machine Learning: Science and Technology.
  DOI: 10.1088/2632-2153/ae8e31

  Application of graph transformer architectures and mixture-of-experts for rare signal detection in calorimeter data.
  Download Paper (PDF)

• Genovese D., Sgroi A., Devoto A., Valentine S., Wood L., Sebastiani C., Scardapane S., D’Onofrio M., Giagu S. (2025).
  Mixture-of-experts graph transformers for interpretable particle collision detection.
  Scientific Reports, 15(1), 27906.
  DOI: 10.1038/s41598-025-12003-9

  A novel approach combining Graph Transformer and Mixture-of-Expert layers for explainable high-energy physics data analysis.
  Download Paper (PDF) | Download Slides (PDF)