Papers by Davide Bacciu
JuriFindIT: an Italian legal retrieval dataset (2026.findings-eacl)
Copied to clipboard
| Challenge: | Statutory article retrieval (SAR) targets retrieval of legislative provisions relevant to a natural language question. |
| Approach: | They propose a pipeline that integrates dense encoders with an heterogeneous legislative graph . they propose statutory article retrieval (SAR) is the first SAR dataset for the italian legal domain . |
| Outcome: | The proposed pipeline improves over existing approaches. |
Learning from Non-Binary Constituency Trees via Tensor Decomposition (2020.coling-main)
Copied to clipboard
| Challenge: | a binarisation procedure changes the structure of constituency trees, furthering constituents that are not binary. |
| Approach: | They propose a binarised approach to binarise constituency trees by tensor-based models . they propose 'trunk-LSTM' model which exploits such a rich structure . |
| Outcome: | The proposed model performs well on different NLP tasks. |
Self-generated Replay Memories for Continual Neural Machine Translation (2024.naacl-long)
Copied to clipboard
| Challenge: | Neural Machine Translation systems exhibit strong performance in several different languages, but their ability to learn continuously is limited by catastrophic forgetting. |
| Approach: | They propose a method that leverages a key property of encoder-decoder Transformers, i.e. their generative ability, to continuously learn Neural Machine Translation systems. |
| Outcome: | The proposed approach can counteract catastrophic forgetting without explicit memorization of training data. |