December 4, 2019
While AI may be the new electricity significant challenges remain to realize AI potential. Here we examine why data scientists and teams can’t rely on software engineering tools and processes for machine learning.
NEW YORK, NY–comet.ml, the industry-leading machine learning (ML) experimentation platform that enables data scientists to track, compare, monitor, and optimize model development, announced today an additional $4.5 million investment from their existing investors, Trilogy Equity Partners, Two Sigma Ventures, and Founder’s co-op. The new funding will be used to expand Comet Enterprise and continue bringing state of the art meta learning capabilities to the market.
Since the launch of comet.ml in 2018, millions of models have been trained using the platform across multiple industries spanning autonomous vehicles, financial services, technology, bioinformatics, satellite imagery, fundamental physics research, and much more. Comet Enterprise customers include Ancestry, General Electric, The National Institute of Health and Fortune 100 companies, with active users at companies such as Boeing, Google, and Uber.
“Professionals from the best companies in the world choose Comet which allows them to build ML models that bring significant business value. In such a fast growing space Comet is well positioned to become the de-facto Machine Learning development platform. As someone who led Microsoft’s developer tools division I’m excited to see Comet follow a similar growth trajectory.” said Yuval Neeman, partner at Trilogy Equity Partners.
“Comet provides a central place for my team to track their ML experiments and models so we can seamlessly compare and share experiments, debug and stop underperforming models. Comet has improved our efficiency as data scientists and as a team,” said Carol Anderson, Staff Data Scientist, Ancestry.
On top of industry-leading tracking, comparison, and workflow management features, Comet Enterprise offers predictive early stopping, a meta learning functionality not seen in any other experimentation platforms. Building on top of millions of public models, Comet’s predictive early stopping product can predict the performance of experiments before they finish training. This improves model training time by 30% irrespective of the underlying infrastructure and stops underperforming models automatically, which reduces cost and carbon footprint by 30%.
In addition to the new meta learning features, Comet Enterprise offers self-hosted and on-premise deployments, greater security and control for teams and organizations, and 24/7 multi-channel support. The enterprise offering also includes Comet’s flagship visualization engine, which allows users to visualize, explain, and debug model performance and predictions, and a state of the art parameter optimization engine.
“We founded Comet with the goal of helping companies gain business value from machine learning. With Comet Enterprise, data scientists can build reliable models faster and organizations can gain visibility into their AI/ML efforts and progress,” said Gideon Mendels, Co-founder/CEO, comet.ml. “Comet Enterprise already supports some of the biggest and most impactful machine learning teams in the world, and we plan on helping more organizations do the same.”
Comet also recently announced two partnerships. The first with Uber AI to extend Ludwig, a low code deep learning toolbox. The second with Dell EMC, allowing users to leverage the Dell EMC kubernetes architecture with Comet. Learn more about Comet Enterprise here.
Comet provides a self-hosted and cloud-based meta machine learning platform allowing data scientists and teams to track, compare, explain, and optimize experiments and models. Backed by thousands of users and multiple Fortune 100 companies, Comet provides insights and data to build better models faster while improving productivity, collaboration and visibility across teams. Learn more at: www.comet.com