Roberto Dessì
Source-listed: Research Scientist · Not Diamond
Record snapshot: Sep 3, 2026. Coverage and affiliations may be incomplete or historical.
Selected work
5- Toolformer: Language Models Can Teach Themselves to Use ToolsSaved credit: Contributor · 2023Explore people connected to this work →
- Augmented Language Models: a SurveySaved credit: Contributor · 2023Explore people connected to this work →
- Cross-Domain Image Captioning with Discriminative FinetuningSaved credit: First author · 2023Explore people connected to this work →
- Communication breakdown: On the low mutual intelligibility between human and neural captioningSaved credit: First author · 2022Explore people connected to this work →
- Interpretable agent communication from scratch (with a generic visual processor emerging on the side)Saved credit: First author · 2021Explore people connected to this work →