Advertisement
New Zealand markets closed
  • NZX 50

    12,105.29
    +94.63 (+0.79%)
     
  • NZD/USD

    0.5961
    -0.0044 (-0.74%)
     
  • ALL ORDS

    8,153.70
    +80.10 (+0.99%)
     
  • OIL

    81.88
    +0.53 (+0.65%)
     
  • GOLD

    2,217.40
    +4.70 (+0.21%)
     

Continuum Industries Chooses Iterative to Optimize Civil Infrastructure Design with Evolutionary Computing

Iterative reduced Continuum’s setup and runtime from 48 hours to just three using DVC and CML

SAN FRANCISCO, Jan. 18, 2022 (GLOBE NEWSWIRE) -- Iterative, the MLOps company dedicated to streamlining the workflow of data scientists and machine learning (ML) engineers, today announced Continuum Industries, which provides AI tools for engineering professionals to rapidly design linear infrastructure projects, has chosen Iterative-backed open source projects DVC and CML to optimize evolutionary computing optimization workflows and reduce time to market.

Continuum Industries works with large amounts of geospatial data with evolutionary computing algorithms to optimize the design of infrastructure like railways and roads. While Continuum Industries do not use machine learning algorithms, they face a number of the same problems that MLOps aims to resolve. They were looking for a way to have that data sync with the code and be versioned together. After considering a custom build using basic ML tools offered by Amazon Web Services (AWS), Continuum chose Iterative tools for their Optioneer product because they allowed it the freedom to freely integrate various ML tools from other vendors into their workflows (like GitHub Actions for CI/CD in training their models), and begin working on test cases immediately.

“With Iterative, we were able to get started right away without having to maintain it ourselves,” said Ivan Chan, AI engineer at Continuum Industries. “Given the incredible time savings it has already provided, we are planning on expanding our use of DVC to also set up our development and testing environment also to experiment versioning and more.”

ADVERTISEMENT

With Iterative, Continuum Industries is now able to version everything beyond code, including data, and ML pipelines, and experiments, with DVC, run frequent algorithm tests with reproducible results through Continuous Machine Learning (CML), as well as slash support time. The developer time spent on maintaining Continuum's suite of algorithm tests has been reduced from five hours every three weeks down to virtually no time at all. Due to the time savings, the team can invest more resources on model development and optimization.

While initially drawn in by Iterative’s DVC tool, Continuum has also found great success with CML for continuous integration and testing for their algorithms. CML connects ML model testing phase that runs through CI/CD with source code and project management system GitHub and automatically generates reports right next to the proposed modeling code changes which are called Pull Request in GitHub. Besides automating CI/CD processes for model training in the cloud, CML also provides summaries of performance and metrics and informs the team of any potential bugs or issues as model training progresses.

“With the recently passed Infrastructure Investment and Jobs Act, Continuum is helping build the infrastructure of the future using machine learning,” said Dmitry Petrov, founder and CEO of Iterative. “Because they use enormous amounts of data to evaluate construction engineering decisions, DVC and CML help automate their MLOps, resulting in faster time to market for ML models for Continuum's clients.”

DVC brings agility, reproducibility, and collaboration into the existing data science workflow. DVC provides users with a Git-like interface for versioning data and models, bringing version control to machine learning and solving the challenges of reproducibility. DVC is built on top of Git, allowing users to create lightweight metafiles and enabling the system to handle large files, rather than storing them in Git. It works with remote storage for large files in the cloud or on-premise network storage.

CML is an open-source library for implementing continuous integration and delivery (CI/CD) in machine learning projects. Users can automate parts of their development workflow, including model training and evaluation, comparing ML experiments across their project history, and monitoring changing datasets. CML will also auto-generate reports with metrics and plots in each Git pull request.

Together, CML and DVC provide ML Engineers a number of features and benefits that support data provenance, machine learning model management and automation. DVC and CML are open-source tools available for free. Iterative also provides a commercial offering of a collaboration service DVC Studio. To schedule a demo, visit www.iterative.ai.

About Continuum Industries
Early decisions in infrastructure projects have major consequences, however there is not enough time or resources to consider them in sufficient detail.

Continuum Industries mission is to build the global solution for infrastructure planning, allowing projects to complete faster.

Continuum Industries' solution, Optioneer, automates design processes and uses a powerful AI-driven engine to explore millions of possible design options. Optioneer reduces the planning time from days and weeks to minutes and hours, by combining analysis of engineering, cost, GIS, social and design data.

Continuum Industries technology allows you to explore more options, in greater detail, earlier in projects so that infrastructure projects can complete faster. For more information visit https://continuum.industries/.

About Iterative
Based in San Francisco, Iterative.ai is the company behind the development of DVC and CML, open-source tools to streamline the workflow of data scientists. Iterative.ai integrates ML workflows into current practices for software development instead of creating a separate AI platform. For more information visit www.iterative.ai.

Media Contact:
Joe Eckert, jeckert@eckertcomms.com
Ray George, ray@eckertcomms.com