Due to the rapid advancement in computer hardware and algorithms, molecular modeling, with its main branches of molecular dynamics (MD) and Monte Carlo (MC), has firmly established itself as a third essential pillar alongside theory and experimentation. Substantial progress has been made in overcoming challenges such as accuracy, system size, and time range, with improvements in force fields and ab initio MD methods. While parallel computing has enhanced system size capabilities, it plays a more limited role in extending the time range due to the sequential nature of time evolution. Despite advancements, studying rare molecular transitions remains a time-intensive process.
Path sampling simulations offer an efficient solution, providing dynamics of rare events in an exact manner, exponentially faster than conventional MD simulations. Yet, even advanced path sampling simulations for large systems often require months to a year of wall time. Accelerating simulations will greatly benefit industries, meeting the demand for fast computational tools in predicting chemical and biomolecular processes. A fast and accurate computational technique holds significance in catalyst design and drug development, streamlining screening processes and reducing time and resource requirements.
This workshop aims to deepen expertise in path sampling and Machine Learning, enhancing understanding and manipulation of chemical reactions. Activities include unifying path sampling codes, exploring highly parallelizable path sampling algorithms like ∞RETIS, analyzing reaction mechanisms, and developing reactive force fields through Ab Initio-level path sampling and ML.
