EXTENDED ABSTRACT: The convergence of machine learning (ML) and materials science heralds a new era in material design and optimization. Despite the promise of accelerated discovery and enhanced performance, aligning ML techniques with the intricate realities of material synthesis and characterization poses significant challenges. In this talk, Dr Liu, Founder and CEO of DeepVerse, will delve into how advanced ML methodologies can be effectively integrated with material design process to tackle these challenges. We will explore the limitations of current generative models and computational methods in predicting material properties. He will discuss how to leverage ML, computational proxy, and high-throughput experimentation to not only propose novel material structures but also ensure their practical applicability. • Bridging the Gap between ML Models and Experiments: Leveraging experimental data to refine ML models and accelerate the material discovery process. • Multi-Fidelity Optimization Techniques: Balancing computational predictions with experimental validation to improve • Case Studies and Applications: Real-world examples where aligning ML with material design has led to significant breakthroughs in material performance and industrial impact. By aligning machine learning with the practical demands of material design and optimization, this presentation aims to provide a roadmap for researchers and industry professionals seeking to harness the full potential of AI in materials informatics.

Keywords: Artiffcial Intelligence;Machine Learning; Materials Deisgn; Multi-Fidelity;