Scholar
Matteo Merler
Google Scholar ID: 2_dW7y4AAAAJ
Researcher, FBK
Citations & Impact
Citations
49
H-index
2
i10-index
2
Publications
4
Co-authors
7
list available
Publications
Resume (English only)
Background
- Currently a pre-doctoral researcher at FBK NLP in Trento, Italy.
- Research interests focus on bridging large-scale models (e.g., LLMs and VLMs) with other learning paradigms such as Reinforcement Learning (RL) and Symbolic Planning (SP).
- Believes RL’s key strength is learning from experience through interaction—something current LLMs lack.
- Notes that pure RL struggles with generalization as it learns from scratch in each new environment.
- Fascinated by model-based RL and world models; aims to investigate whether large models possess internal world models and how to teach them to plan ahead.
- Seeks to develop AI agents that understand natural language instructions, leverage LLMs' real-world knowledge, and perform complex, long-horizon reasoning through direct interaction.
