Our great sponsors
-
seldon-core
An MLOps framework to package, deploy, monitor and manage thousands of production machine learning models
-
InfluxDB
Power Real-Time Data Analytics at Scale. Get real-time insights from all types of time series data with InfluxDB. Ingest, query, and analyze billions of data points in real-time with unbounded cardinality.
-
AIF360
A comprehensive set of fairness metrics for datasets and machine learning models, explanations for these metrics, and algorithms to mitigate bias in datasets and models.
While tools in the model experimentation space normally include diagnostic charts on a model's performance, there are also specialised solutions that help ensure that the deployed model continues to perform as they are expected to. This includes the likes of seldon-core, why-labs and fiddler.ai.
Besides the existing monitoring solution mentioned in the section above, we were also took inspiration from continuous integration and continuous delivery (CI/CD) testing tools like Jenkins and Circle CI, on the engineering front, and existing fairness libraries like Microsoft's Fairlearn and IMB's Fairness 360, on the machine learning side of things.
We explored various solutions in the machine learning space as well as in the neighbouring software development space for inspiration and learning. Some notable ones include Git and Github, Google Model Cardsand IBM's Factsheets. One of the main design decisions we faced was whether to go with a plain-text git-based solution or a structured schema approach. Eventually, we decided to build on top of Google Model Card (structured protobuf schema). This sacrifices immediate readability of the file, but makes it easy to be processed across different systems in a predictable manner.
Besides the existing monitoring solution mentioned in the section above, we were also took inspiration from continuous integration and continuous delivery (CI/CD) testing tools like Jenkins and Circle CI, on the engineering front, and existing fairness libraries like Microsoft's Fairlearn and IMB's Fairness 360, on the machine learning side of things.
Besides the existing monitoring solution mentioned in the section above, we were also took inspiration from continuous integration and continuous delivery (CI/CD) testing tools like Jenkins and Circle CI, on the engineering front, and existing fairness libraries like Microsoft's Fairlearn and IMB's Fairness 360, on the machine learning side of things.
Interested readers should check out the VerifyML website, docs or Github code. Feel free to create a Github issue and drop any suggestions or feedback over there! VerifyML is proudly open-sourced and created by a small tech startup. I would like to think that more companies are going to see responsible AI as a comparative advantage or requirement and having an ecosystem of solutions that are not controlled by the interests of large tech companies would be key in driving the sector forward. I look forward to improving the user experience and integration with more machine learning tools over the next year, as well as sharing more thoughts in the space.