Embeddings are vector representations of data. Embeddings are everywhere in modern deep learning, such as transformers, recommendation engines, layers of deep neural networks, encoders, and decoders.
Data drift in unstructured data like images is complicated to measure. The measures typically used for drift in structured data do not extend to unstructured data. The general challenge with measuring unstructured data drift is that you need to understand the change in relationships inside the unstructured data itself.
Arize supports 2 approaches - bring your own embeddings or have Arize generate them for you.
Embeddings are foundational because:
- 1.They can represent images, audio signals, and even large chunks of structured data.
- 2.They provide a common mathematical representation of your data
- 3.They compress your data
- 4.They preserve relationships within your data
- 5.They are the output of deep learning layers providing comprehensible linear views into complex non-linear relationships learned by models
Example of Image, image embedding, and the embedding projection to 2-D
Check out our tutorials on how to send embeddings to Arize for different use cases.
Learn more about embeddings and troubleshooting with Arize: