Overview

Explore the core ideas and principles behind Superlinked's functionality..

1. Describe your data using Python classes with the @schema decorator.

2. Describe your vector embeddings from building blocks with Spaces.

3. Combine your embeddings into a queryable Index.

4. Define your search with dynamic parameters and weights as a Query.

5. Load your data using a Source.

6. Define your transformations with a Parser (e.g.: from pd.DataFrame).

7. Run your configuration with an Executor.

Colab notebooks explaining the concepts

Accessing stored vector parts

Use query interface to return vector-parts based on our needs

Categorical Embeddings

Efficiently represent and compare categorical data in vector space for similarity searches.

Combine Multiple Embeddings

Merge different types of embeddings to create a unified representation for complex objects.

Custom Spaces

Create and manage custom vector spaces for specialized similarity searches.

Dynamic Parameters

Adjust query parameters dynamically to fine-tune search results.

Event Effects

Model and apply the impact of events on vector representations over time.

Hard Filtering

Apply strict criteria to narrow down search results before similarity ranking.

Image Embedding

Embed text or images into a multi-modal vector space.

Natural Language Querying

Perform similarity searches using natural language queries instead of vector representations.

Number Embedding Minmax

Embed numerical values within a specified range for effective similarity comparisons.

Number Embedding Similar

Embed numbers to find similar values based on relative closeness rather than exact matches.

Optional Schema Fields

Define optional fields in your schema.

Query Result

Customize the result object for your queries.

Query Time Weights

Adjust the importance of different embedding components during query execution.

Querying Options

Customize search behavior with various querying options for refined results.

Recency Embedding

Incorporate time-based relevance into vector representations for up-to-date search results.

Text Embedding

Convert text data into vector representations for semantic similarity searches.

Vector Sampler

Generate diverse vector samples to explore and understand the embedding space.