HOMER.PTML Heuristic Optimization of Molecules, Emsembles, and Reactions by Perturbation Theory, and Machine Learning from Chiral Catalysts to Polymers, Nanoparticle for Drug Release Systems and Personalized Medicine

• Autores: Shan He
• Directores de la Tesis: Humberto González Díaz (dir. tes.), Sonia Arrasate Gil (codir. tes.)
• Lectura: En la Universidad del País Vasco - Euskal Herriko Unibertsitatea ( España ) en 2025
• Idioma: inglés
• Programa de doctorado: Programa de Doctorado en Química Sintética e Industrial por la Universidad del País Vasco/Euskal Herriko Unibertsitatea y la Universidad Pública de Navarra
• Menciones:
  • Internacional
• Calificación: Sobresaliente Cum Laude
• Enlaces
  • Tesis en acceso abierto en: ADDI
• Resumen

  • Designing and optimizing complex systems, which may include multiple interconnected subsystems, is inherently challenging due to the high number of input and output variables—ranging from continuous to categorical—and the wide array of experimental conditions. Moreover, the limited availability of data often requires researchers to compile information from various sources, adding to the complexity. A major hurdle in handling such large datasets is managing the "5V + C" framework, which encompasses Volume, Velocity, Variety, Veracity, Value, and Complexity. To address these issues, this thesis introduced the HOMER.PTML algorithm (Heuristic Optimization of Molecule, Ensemble, and Reactions by Perturbation Theory and Machine Learning). This method leverages a multi-output, multi-label machine learning approach, structured around three core phases: (1) Information Fusion (IF): Combines data from multiple scientific sources to create a comprehensive dataset. (2) Perturbation Theory (PT): Analyzes how structural variables deviate from expected values under different categorical conditions. (3) AI/ML Algorithm Development: Focuses on modeling and predicting biological properties in complex systems, with applications in chemical reactivity, nanoscience, and personalized medicine. Thisintegrated approach aims to streamline the optimization of complex systems while addressing data scarcity and variability, reducing time and resources in wet-laboratory.