### 2025
**The Result Model under Inconsistent Knowledge: Theory and Experiments**
*Yoann Morello, Agata Ciabattoni, Morgan Gray*
To appear in: Proceedings of JURIX 2025 – The 38th International Conference on Legal Knowledge and Information Systems, Turin, Italy
[PDF link: https://www.logic.at/staff/agata/jurix2025.pdf]
**A Bayesian View of the Result Model**
*Yoann Morello, Agata Ciabattoni*
In: **Digital Humanism – Research and Practice** (DIGHUM-25), Springer.
We provide a Bayesian reinterpretation of the result model for case-based reasoning in law, allowing us to quantify the reliability of precedents under inconsistent knowledge.
[Chapter link: https://link.springer.com/chapter/10.1007/978-3-032-11108-1_4]
---
### 2024
**Exploring Task Affinities through NTK Alignment and Early Training Dynamics in Multi-Task Learning**
*Yoann Morello, Emilie Grégoire, Sam Verboven*
Poster at the **Mathematics of Modern Machine Learning (M3L)** workshop, NeurIPS 2024.
We express several gradient-based task-similarity measures in terms of the Neural Tangent Kernel (NTK) and propose ways to average out initialization noise so that useful task affinities can be computed without full training.
[Workshop page: https://neurips.cc/virtual/2024/103491]
[OpenReview PDF: https://openreview.net/pdf?id=HxT9EuHdXW]