SparkCognition's Release of Darwin™ 2.0 Brings Increased Speed and Scalability to Automated Model Building
SparkCognition, a global artificial intelligence (AI) company, today announced version 2.0 of its automated machine learning (AutoML) product, Darwin. The new version significantly improves the user's experience and capabilities within the platform by expediting the data preparation process and automating the model building process with intuitive workflows. "The release of Darwin v2.0 marks a significant milestone in our product roadmap and our endeavor to accelerate automated extraction of insights from data and improve the productivity of data scientists," said Sridhar Sudarsan, CTO at SparkCognition. "This iteration delivers extensive enhancements on a strong foundation by adding data transformation and additional modeling approaches with an immersive user experience to solve complex problems at scale." AutoML is changing data science, as it automates the process of building, testing, deploying, and maintaining models for a dataset. Darwin provides an intuitive environment that takes users quickly from data to meaningful results, which enables organizations to scale the adoption of data science across teams, and the implementation of machine learning applications across operations, becoming data-driven enterprises. "We recently announced that we would be the partner to deliver SparkCognition's Darwin product to Japan and other Asian countries," said Masahiro Taniguchi, President of Hitachi High-Tech Solutions. "I am excited that Hitachi High-Tech Solutions' data scientists are using Darwin and our cloud service to support our customers to solve problems and enhance business efficiency. We chose SparkCognition's automated model building product, Darwin, over the competition because the product stands above others in the market and enables invaluable business insights." Darwin v2.0 features include: Automated data quality checks that highlight problems in a dataset and offers solutions to any errors encountered Improved neuro-evolutionary engine that enables quickly-built custom models from scratch Increased control over the model building process that preserves the accuracy and performance of models State-of-the-art genetic time series forecasting algorithms and sequence-to-sequence models capable of predicting multiple time horizons Enhanced ability to capture complex relationships over time and make predictions using long short-term memory (LSTM) and temporal convolutional network (TCN) architectures With these new and improved features, Darwin enables data scientists, data and business analysts, software developers, and SMEs to prepare their datasets and custom-build models from scratch much faster than before. These models are tailored to the given dataset by considering its complex relationships and intricacies, and quickly evaluating hundreds of generations of model architectures to optimize from.
Darwin 2.0 Press Release_FINAL (2).docx