System Platform

A new way of organizing and discovering information based on systems

At System, we believe that it is time to evolve the way we organize data and knowledge: from silos to systems. So starting from first principles we invented and patented a new way of organizing and discovering information — based primarily on measured statistical relationships between things in the world.

First Principles

We designed the platform from first principles.

• Retrieve and encode statistical relationships at the highest possible resolution to preserve fidelity.

• Be as objective, impartial, precise, and transparent as possible in how we represent information.

• Edge-first. Design schemas from the edge out. Relationships are first order objects

• Integrate seamlessly with other parts of the data and information stack and the infrastructure of open knowledge.

• Maximize node reusability and reuse to map the world as one system.

• Minimize user data storage and personal data use.

Data

The essence of the platform is an evidence-based relationship.

• We use AI to extract statistical relationships from peer-reviewed studies, datasets, and models. You can read about the steps we take to ensure accuracy in extraction here.

• We have extracted the statistical results from millions of peer-reviewed studies, starting in health and life sciences. To the best of our knowledge, this is the first time statistical information retrieval has been achieved at this scale.

• We have also extracted and layered in causal and mechanistic statements for things like genes and proteins using rules-based models.

• The platform recognizes over 100 types of statistical associations and algorithms (e.g. Odds Ratio, Pearson R, Hazard Ratio, etc.) and stores this information alongside measures of significance (e.g. p-value) and context (population, sample size, control variables, etc.).

• We compute additional information to enrich a raw statistical association, like its strength, direction, and sign.

• Using AI, we ground the statistical results we extract in known scientific terms, with a focus on grounding to ontologies that are interoperable with other biomedical data sources.

• Today, the resulting data powers the syntheses on System Pro and is the basis of our graph.

Graph

We organize evidence-based relationships in a large-scale graph.

• Our graph is organized into nodes and edges. Nodes are measurable variables in the world (like body mass index or concentration of fine particulate matter) and the larger concepts they measure (like obesity or air pollution). Edges indicate a statistical relationship between two nodes in a specific context.

• Notably, unlike knowledge graphs, the edges here are not purely semantic. Instead, they are constructed around quantified or mechanistic evidence of association between things in the world. A causes B vs. A is a type of B.

• We organize, normalize, and link these millions (in the future, billions) of relationships so that anything in the world can be connected to everything else across studies, disciplines, fields, etc.

• Today, the graph powers the maps on System and System Pro and the recommendations on System Pro.