Algorithmia Tackles Machine Learning Model Risks with New Governance Capabilities


Receive instant alerts when news is published about your stocks. Claim your one week free trial for StreetInsider Premium here.


SEATTLE, March 03, 2021 (GLOBE NEWSWIRE) — algorithma leader in machine learning (ML) operations and management software, has launched new advanced reporting tools to help IT managers and insider risk managers manage the use of ML models in IT environments. production.

According to Algorithmia’s 2021 Machine Learning Enterprise Trends Report, the primary ML challenge facing organizations today is governance. 56% of IT leaders responding to Algorithmia’s survey ranked governance, security, and auditability issues as a top concern; 67% of all respondents said they had to comply with multiple regulations. The effects of a model failure may not be known for some time, perhaps after bad credit decisions, fraud detection decisions, or customer visible decisions have been made. Current model governance approaches are not sufficient or are not applied appropriately to machine learning operations (MLOps).

In most organizations, governance and risk management of the ML model is primarily focused on validating and testing models and inspecting documentation before deploying the model. As ML adoption has accelerated over the past year, IT managers, line of business managers, CIOs and risk managers have realized that what happens after deployment of a model is even more important than testing and validation before deployment. Operational risk is now the most important analytical risk.

Advanced reporting and governance features

Today’s launch of Algorithmia’s advanced reporting capabilities for governance complements the compliance and auditing capabilities of its Enterprise product. It also augments Algorithmia’s existing capabilities regarding explainability and performance monitoring (available in Algorithmia Insights), model cataloging, repository, and security.

Algorithmia’s Enterprise product now provides the following reporting and governance features:

  • Cost and usage reports on the infrastructure, storage and compute consumption within Algorithmia to understand and manage the overall cost of maintaining the platform.
  • Improved chargeback and chargeback reporting for monthly storage costs, CPU and GPU consumption, and usage billing.
  • Algorithm Usage Reports with details of the algorithm used, so that organizations can charge users for their use.
  • Improved reports and audit logs so that reviewers and auditors can review model results, change history, and a record of data errors or past model failures and actions taken.
  • Advanced report panel for Algorithmia administrators that provides an overview of all available metrics and usage reports, the ability to create reports and export reports and metrics to systems of record.

“We are still in the early days of ML governance, and organizations lack a clear roadmap or prescriptive guidance to implement it effectively in their own unique environments,” said Diego Oppenheimer, CEO of Algorithmia. . “Regulations are undefined and a changing and ambiguous regulatory landscape leads to uncertainty and the need for businesses to invest significant resources to maintain compliance. Those who can’t keep up risk losing their competitive advantage. Moreover, the existing solutions are manual and incomplete. Even organizations implementing governance today do so with a patchwork of disparate tools and manual processes. Not only do such solutions require constant maintenance, but they also risk critical gaps in coverage.

About the algorithm

For machine learning leaders who need to bring ML models to production faster, more securely, and more cost-effectively as part of their existing business processes, Algorithmia is an MLOps software that manages all stages of the ML life. Unlike inefficient, expensive, and insecure MLOps management solutions that lock users into specific technology stacks, Algorithmia automates ML deployment, optimizes collaboration between operations and development, leverages existing SDLC and CI/CD processes, and provides a advanced security and governance. More than 120,000 engineers and data scientists have used Algorithmia’s platform to date, including the United Nations, government intelligence agencies, and Fortune 500 companies. For more information, visit www.algorithmia.com.

Media contact: Kevin Wolf TGPR (650) 483-1552 [email protected]

Source: Algorithmy