What is the difference between Vertex ML Metadata and Audit logs in Vertex AI's context ?

[gauravpant]

Vertex ML Metadata is local to Vertex AI , on the other hand Audit logs are kind of characteristic of the GCP itself . 

Both the features can essentially tell us about our ML workflow .

To know about a certain Machine Learning workflow outside of Google Cloud Console (  using nodejs apis ) , which of the feature among the 2 should be used and why ?

[nceniza]

If you are looking for Vertex AI workflows I would suggest visiting this document. 

Vertex AI uses a standard machine learning workflow:

1. Gather your data: Determine the data you need for training and testing your model based on the outcome you want to achieve.
2. Prepare your data: Make sure your data is properly formatted and labeled.
3. Train: Set parameters and build your model.
4. Evaluate: Review model metrics.
5. Deploy and predict: Make your model available to use.

But before you start gathering your data, you need to think about the problem you are trying to solve. This will inform your data requirements.

---

For audit logs these helps you answer "Who did what, where, and when" as described in the documentation here. It's more on administrator logs than the actual machine learning. 

For analysis related to Machine Learning, I would recommend the Vertex ML Meta data since it has recording both your observations and the parameters of an experiment.

Reference: 

https://cloud.google.com/vertex-ai/docs/ml-metadata/introduction#overview_of