The machine learning life cycle and the cloud: implications for drug discovery

Abstract

Introduction: Artificial intelligence (AI) and machine learning (ML) are increasingly used in many aspects of drug discovery. Larger data sizes and methods such as Deep Neural Networks contribute to challenges in data management, the required software stack, and computational infrastructure. There is an increasing need in drug discovery to continuously re-train models and make them available in production environments.Areas covered: This article describes how cloud computing can aid the ML life cycle in drug discovery. The authors discuss opportunities with containerization and scientific workflows and introduce the concept of MLOps and describe how it can facilitate reproducible and robust ML modeling in drug discovery organizations. They also discuss ML on private, sensitive and regulated data. Expert opinion: Cloud computing offers a compelling suite of building blocks to sustain the ML life cycle integrated in iterative drug discovery. Containerization and platforms such as Kubernetes together with scientific workflows can enable reproducible and resilient analysis pipelines, and the elasticity and flexibility of cloud infrastructures enables scalable and efficient access to compute resources. Drug discovery commonly involves working with sensitive or private data, and cloud computing and federated learning can contribute toward enabling collaborative drug discovery within and between organizations.

Machine Learning Operations: A Survey on MLOps Tool Support

“…Data can be open-source or private and collected using surveys or experiments. Due to the inaccuracy and redundancy of data, they should be cleaned and preprocessed before using for training [15]. Then feature engineering techniques are applied to extract and identify vital informative features for the design of the ML models [16].…”

“…The DataRobot MLOps platform supplies a single place to deploy, monitor, manage models in productions regardless of how they were created, when and where they were deployed 15 . It has a model registry to store and manage all production deployed models.…”

Applying DevOps to Machine Learning Processes: A Systematic Mapping

“…O estudo A12, de [Liu et al 2020], introduz o ciclo de vida de MLOps e investiga processos e ferramentas de CI/CD para implantac ¸ão de modelos de ML, propondo a adoc ¸ão de frameworks de código aberto para esse fim. Por fim, A15, dos autores [Spjuth et al 2021], apresentam uma discussão em torno da aplicac ¸ão de aprendizado de máquina na descoberta de novos medicamentos, e como MLOps pode contribuir para a criac ¸ão e implementac ¸ão de um modelo robusto de ML que atenda a esse cenário.…”

Fink: Early supernovae Ia classification using active learning

“…The basic idea of AL is to allow the algorithm to identify which alerts should be labeled and included in the training set, according to their potential at improving the classifier. This philosophy is at the core of many of the recommendation systems currently in place, from pharmaceutical studies (Spjuth et al 2021) to urban planning (Abernethy et al 2018). In astronomy, it has proven to be effective in a few different scenarios (e.g., Solorio et al 2005;Richards et al 2012;Vilalta et al 2017;Walmsley et al 2020), including the classification of astronomical transients (e.g., Ishida et al 2019;Kennamer et al 2020) and anomaly detection (Ishida et al 2021).…”