The era of Big Data requires the ability to analyze vast amounts of complex data, discover interconnectedness in the data and to make predictions and deliver insight. Thriving amidst this complexity requires new analytic tools that go beyond the simplifications of established analytic methods and basic human
simMachines has addressed these requirements by creating a
revolutionary Discovery & Predictive Analytics tool. simMachines is a similarity-based machine learning company focused on making predictions actionable to Decision Makers. Each prediction we provide comes with an intuitive and actionable justification. Read more on why this is important. For data scientists we provide ultimate data sculpting flexibility: the representation capabilities of our technology are 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”
Machine Learning algorithms in the industry can predict the future with great accuracy but they will not tell you what will cause an event nor why. Furthermore, certain predictions provided by Machine Learning algorithms must be audited by governments or other third parties and in order to explain a prediction in a clear and actionable way to auditors of different backgrounds and experiences.
Our technology is uniquely positioned to address this use case.
DIFFERENTIATOR #2: “THE SHAPE”
Current technologies represent data in Euclidean vectors or tensors. There is nothing wrong with this representation but it is limiting. If you want to represent a Tree, a Graph, a Molecule or a multivariate time-series you are severely limited. Embeddings bring distortion and therefore they must be used carefully. Our approach is based on Metric and Non-metric spaces and that means you can represent any kind of data shape you need without distorting it with embeddings.