An Automated Model Building Assistant for the Data Scientist

Build Better Machine Learning Models with Genetic Algorithms and Deep Learning

SparkCognition has developed the next stage in machine intelligence with Darwin, a platform that automates the model building process and amplifies the potential of a data scientist. Darwin replicates the mind of a data scientist by automating the cleaning of data, performing latent relationship extraction, and determining the optimal model to solve a given problem.
Darwin is a major advancement from traditional machine learning, which operates via manual data cleaning, feature generation, and parameter tuning. These traditional approaches run into several problems: They take time to design and implement, often struggle to scale across large operations, and can be limited by edge cases that occur under extreme or unusual operating parameters. By contrast, Darwin creates dynamic models that are highly accurate and take less time to develop.

Darwin automates a variety of functions, including but not limited to

Preprocessing and Feature Generation

  • Scaling for normalization of data
  • Imputation for replacing missing data with substituted values
  • Handling of categorical variables
  • Balancing of datasets
  • Generation of risk index
  • Windowing and feature creation

Feature Selection and Model Building

  • Classification
  • Regression
  • Anomaly detection
  • Clustering
Darwin finds the best models for a situation significantly faster than a person could. It augments even the most skilled data scientists and lets them spend more time on solving a problem, not dealing with tools. By leveraging the best of genetic algorithms and deep learning, Darwin provides the next step into the future of artificial intelligence.

Find out how Darwin can augment your data scientists