Luckeciano Carvalho Melo

Zihuiwen Ye, Luckeciano Carvalho Melo, Younesse Kaddar et al.

Complex multi-step reasoning tasks, such as solving mathematical problems, remain challenging for large language models (LLMs). While outcome supervision is commonly used, process supervision via process reward models (PRMs) provides intermediate rewards to verify step-wise correctness in solution traces. However, as proxies for human judgement, PRMs suffer from reliability issues, including susceptibility to reward hacking. In this work, we propose leveraging uncertainty quantification (UQ) to enhance the reliability of step-wise verification with generative reward models for mathematical reasoning tasks. We introduce CoT Entropy, a novel UQ method that outperforms existing approaches in quantifying a PRM's uncertainty in step-wise verification. Our results demonstrate that incorporating uncertainty estimates improves the robustness of judge-LM PRMs, leading to more reliable verification.

Luckeciano Carvalho Melo, Marcos Ricardo Omena Albuquerque Maximo, Adilson Marques da Cunha

Controlling a high degrees of freedom humanoid robot is acknowledged as one of the hardest problems in Robotics. Due to the lack of mathematical models, an approach frequently employed is to rely on human intuition to design keyframe movements by hand, usually aided by graphical tools. In this paper, we propose a learning framework based on neural networks in order to mimic humanoid robot movements. The developed technique does not make any assumption about the underlying implementation of the movement, therefore both keyframe and model-based motions may be learned. The framework was applied in the RoboCup 3D Soccer Simulation domain and promising results were obtained using the same network architecture for several motions, even when copying motions from another teams.