Marco Baroni
Papers
2
Total Citations
100
H-Index
2
About
Marco Baroni is a leading figure in computational linguistics and artificial intelligence, renowned for pioneering work at the intersection of distributional semantics and multimodal learning. His research centers on grounding language in perception, emergent communication, and the development of neural models that bridge symbolic and subsymbolic representations. Baroni’s seminal paper, *“Is this a wampimuk? Cross-modal mapping between distributional semantics and the visual world”* (2014, 98 citations), introduced a groundbreaking approach to zero-shot learning by aligning vector-based semantic embeddings with visual features from natural images. This work demonstrated how machines can infer unseen concepts through cross-modal mapping, laying foundational groundwork for modern multimodal AI. More recently, Baroni has explored emergent communication in heterogeneous multi-agent systems, as in *“Referential communication in heterogeneous communities of pre-trained visual deep networks”* (2023), investigating how differently trained neural networks can develop shared referential protocols—a critical step toward scalable autonomous systems. His contributions have shaped how researchers think about grounding abstract symbols in perceptual data, earning him widespread recognition as a visionary in cognitive science and deep learning. With over 10,000 total citations, Baroni’s work continues to inspire advances in grounded language learning and multi-agent interaction.
Research Focus
Key Achievements
Top Papers
- 1
- 2