Stanislav Belyaev

Daniil Polykovskiy, Alexander Zhebrak, Benjamin Sanchez-Lengeling et al.

Generative models are becoming a tool of choice for exploring the molecular space. These models learn on a large training dataset and produce novel molecular structures with similar properties. Generated structures can be utilized for virtual screening or training semi-supervised predictive models in the downstream tasks. While there are plenty of generative models, it is unclear how to compare and rank them. In this work, we introduce a benchmarking platform called Molecular Sets (MOSES) to standardize training and comparison of molecular generative models. MOSES provides a training and testing datasets, and a set of metrics to evaluate the quality and diversity of generated structures. We have implemented and compared several molecular generation models and suggest to use our results as reference points for further advancements in generative chemistry research. The platform and source code are available at https://github.com/molecularsets/moses.

Saveli Goldberg, Stanislav Belyaev, Vladimir Sluchak

The article proposes a new type of AI system that does not give solutions directly but rather points toward it, friendly prompting the user with questions and adjusting messages. Models of AI human collaboration can be deduced from the classic literary example of interaction between Mr. Holmes and Dr. Watson from the stories by Conan Doyle, where the highly qualified expert Mr. Holmes answers questions posed by Dr. Watson. Here Mr. Holmes, with his rule-based calculations, logic, and memory management, apparently plays the role of an AI system, and Dr. Watson is the user. Looking into the same Holmes-Watson interaction, we find and promote another model in which the AI behaves like Dr. Watson, who, by asking questions and acting in a particular way, helps Holmes (the AI user) make the right decisions. We call the systems based on this principle "Dr. Watson-type systems." The article describes the properties of such systems and introduces two particular: Patient Management System for intensive care physicians and Data Error Prevention System.