The number of searches performed on the internet each day is staggering. Facebook alone sees over 1.5 billion search requests from users each day. Search has become part of just about every online experience — consider booking a hotel, buying a pair of shoes or looking for a movie to stream. They all require search.
When we put to one side tech giants like Google or Amazon, the vast majority of search experiences in websites and applications are built on lexical search. For example, when you search (e.g “blue shirt”), it matches results that contain the words “blue” and “shirt”. This is problematic because end users usually interact with language in human ways — if the user makes a typo, searches using a synonym or phrases their query as a question they are unlikely to retrieve relevant results. This is a long-known short-coming of lexical search and where tensor search can help.
Tensors allow us to use neural networks to structure documents, images and other data in such a way that it can be searched with human-like understanding. This provides typo tolerance, natural language understanding and multi-modal functionality (such as image search) out of the box.
Using Marqo, developers can build and deploy tensor search experiences in a few lines of code. It’s designed for the cloud, horizontally scalable and provides a query DSL language for efficiently filtering results. Marqo allows developers to seamlessly navigate from prototype to production with a single solution.
What is tensor search?
Tensor search uses deep-learning to transform documents, images and other data into collections of vectors called “tensors”. Representing data as tensors allows us to match queries against documents with human-like understanding of the query and document’s content. Tensor search can power a variety of use cases such as:
Check out Marqo
We’re live on github: https://github.com/marqo-ai/marqo
Read more at: https://www.marqo.ai/
We’re hiring engineers
Head to https://jobs.gohire.io/marqo-n6oehhnh/ with your Github/Experience/links to past projects.