Deep reinforcement learning-driven ensemble framework for multi-target prediction of mechanical properties in additively manufactured parts using a small dataset

Key Points

• Accurate predictions of mechanical properties can be achieved with deep reinforcement learning methods, showing a clear advantage over traditional techniques.
• Key evidence indicates that the model outperforms conventional methods on a small dataset with at least three mechanical properties evaluated.
• Ensemble framework approach combines multiple predictive models to achieve enhanced accuracy and robustness in predictions.
• Implications include potential advancements in manufacturing processes, though further validation with larger datasets is needed.