How MLOps Can Help Federal Agencies Maximize Returns on Their Artificial Intelligence Investments

0

The Centers for Disease Control and Prevention is leveraging machine learning to predict the spread of COVID-19. The Postal Service uses technology to expedite the delivery of packages. The Department of Transportation is piloting the use of ML to predict the structural safety of highway bridges. The MoD is testing it to visualize terrain and “see” around obstacles.

In fact, a growing number of federal agencies are exploring ML and other forms of artificial intelligence to further…

READ MORE

The Centers for Disease Control and Prevention is leveraging machine learning to predict the spread of COVID-19. The Postal Service uses technology to expedite the delivery of packages. The Department of Transportation is piloting the use of ML to predict the structural safety of highway bridges. The MoD is testing it to visualize terrain and “see” around obstacles.

In fact, a growing number of federal agencies are exploring ML and other forms of artificial intelligence to further their missions or improve services to citizens. In doing so, they need proven ways to ensure they are implementing ML effectively and getting positive returns from their ML investments.

Enter Machine Learning Operations (MLOps), a set of practices intended to optimize the reliability and efficiency of ML design, development, and execution. Influenced by the widely adopted DevOps approach to building and operating custom applications, MLOps combines methodologies and technologies to accelerate the deployment of ML models. There are great tools available today from software vendors and the open source community, such as cnvrg.io, c3.ai, Databricks, and SAS; with open source tools such as KubeFlow, MetaFlow, Kedro and MLFlow.

MLOps is already proven in government. In fact, documenting and applying MLOps makes agencies twice as likely to meet AI goals and three times more likely to be prepared for AI-related risks. That’s according to Deloitte’s analysis of government respondents to its 2021 survey on the state of AI in business.

MLOps is used by data scientists and other AI experts. But what makes it particularly appealing is that it can allow agencies with less specialized technical skills to benefit from ML models. The main benefits are rapid innovation, creation of repeatable models and workflows, creation of structure and management flow for the entire machine learning lifecycle, all focused on sharing, learning and fastest response time.

Addressing ML Challenges

ML involves complicated math. Training and maintaining ML models can be complex and time-consuming. Agencies sometimes struggle to attract and retain experienced data scientists or secure budget approvals for expensive and time-consuming ML pilot projects.

MLOps solves these problems. The approach can help simplify, streamline, and even automate parts of ML development and operations. This can reduce the number of ML experts required and the time and cost needed to move from business requirement to operational solution.

MLOps provides an iterative process for designing, developing, and operating ML models. It starts with a standardized approach to determining the business requirements of the ML use case. It then streamlines the identification of necessary data inputs, the appropriate ML algorithm, and ML model requirements. Finally, it automates the reintroduction of outputs back into the model for continuous improvement, replacing the time-consuming manual approach of repeatedly training and testing the model.

Many agencies will be able to implement MLOps without investing in new hardware or other infrastructure. But some technological tools can help. For example, if you want to leverage large datasets in your ML use case, you might need hardware that supports persistent memory. If you implement a lot of ML, you might benefit from MLOps orchestration software to manage your end-to-end ML pipeline.

But a growing number of data analysis and visualization products incorporate MLOps functionality. Some can even walk you through the process, from identifying business needs, to specifying the dataset, to selecting the ML algorithm and creating the ML model.

Additionally, major public cloud service providers offer tool suites that include ML algorithms. You can use MLOps methodologies to identify your use case algorithm and test an ML model to see if it meets your needs.

Follow MLOps best practices

MLOps specifies ML best practices. But there are also best practices for getting the most out of MLOps. It always starts with having a clear picture of the problem you want to solve, whether it’s identifying structural problems in critical infrastructure or reporting interactions between prescription drugs.

Next, review your existing software inventory to see if you already have any tools that offer MLOps functionality. Your data management software or public cloud service may provide a library of relevant datasets or ML algorithms. The same goes for your department or your partner organizations. The DoD, for example, is building marketplaces of ML algorithms and models. You can take advantage of these resources to accelerate your ML deployment.

Don’t neglect cybersecurity. If you are working with sensitive data such as personally identifiable information (PII) or classified information, you will need hardware that can encrypt ML data both when it is at rest in a database and when are used in computer memory.

Finally, consider the potential for bias in datasets and ML models. There are documented cases where facial analysis programs offered by major computer vendors were biased based on gender and skin color. As a result, their outputs could be extremely inaccurate. Using a biased dataset or model from an ML library just perpetuates these errors. The good news is that MLOps’ documentation and enforceable processes help ensure more transparent and reliable AI.

From planning supply networks to making public health recommendations, from predicting critical infrastructure maintenance to isolating anomalies in cyber warfare, ML will find more use cases among more federal agencies. . MLOps can help your agency establish best practices for efficiently and cost-effectively designing, developing, testing, deploying, and maintaining ML solutions. The result will be ML results that better serve citizens and support your mission.

Gretchen Stewart is Chief Data Scientist for Intel’s Public Sector.

Share.

Comments are closed.