Valdinei Freire da Silva
Universidade de São Paulo, Universidade Politecnica, Universidade Cidade de São Paulo
Papers
6
Total Citations
21
H-Index
3
About
Valdinei Freire da Silva is a researcher whose work lies at the intersection of reinforcement learning and robotics, with a particular emphasis on developing algorithms that enable robots to navigate and learn efficiently in complex, unknown environments. His major contributions center on the creation of abstract, transferable policies that allow robotic agents to reuse knowledge across different tasks, significantly reducing the time and computational cost of learning from scratch. For instance, his "Compulsory Flow Q-Learning" algorithm (2009) introduced macro-states and partial-policy concepts to accelerate robot navigation learning, while his "Stochastic Abstract Policies" (2011) and "Memoryless Probabilistic Relational Policies" (2012) advanced the field of inter-task knowledge transfer. His work on "Risk-Aware Stochastic Abstract Policies" (2014) further integrated safety considerations into learning, a critical step for real-world deployment. Although his most-cited papers each garner around 2–5 citations, their collective impact is evident in the foundational ideas they provide for scalable, generalizable robotic learning. Da Silva’s research is particularly notable for bridging the gap between reactive navigation and generalized learned control, offering a pathway from hand-coded behaviors to adaptive, intelligent systems.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3
- 4
- 5
- 6