simMachines is a similarity-based ML company focused on making predictions actionable to Decision Makers. Our Predictions come with an intuitive and actionable justification. We provide ultimate data sculpting flexibility; represent data that is not limited to Euclidean vectors or tensors. We can represent data in its native form: 3D molecules, Hierarchical trees, multivariate time series.DIFFERENTIATOR #1: THE WHY
ML algorithms in the industry can predict the future with great accuracy but they will not tell you what will cause an event nor why. Also, some predictions provided by Machine Learning algorithms must be audited by governments or other third parties and we help explain a prediction in a clear and actionable way to auditors of different backgrounds and experiences.DIFFERENTIATOR #2: THE SHAPE
Current technologies represent data in Euclidean vectors or tensors. However, if a user want to represent a Tree, a Graph, a Molecule or a multivariate time-series they are severely limited. Embeddings bring distortion and therefore they must be used carefully. Our approach is based on Metric and Non-metric spaces which means a user can represent any kind of data shape needed without distorting it with embeddings.