4 ms·
Here's a (kinda) ELI5: you would use a language model to create "embeddings" of the text, which you can think of as a set of numbers representing the "meaning"
by bigzyg33k 3y ago
Here's a (kinda) ELI5: you would use a language model to create "embeddings" of the text, which you can think of as a set of numbers representing the "meaning" of a set of characters.
These numbers can be plotted as points in a space, and embeddings of things with similar meanings are plotted close to each other. So things like "exam preparation" would have embeddings close to things like "top study tips".
Say you have created embeddings for a large corpus of text (in this case all youtube captions) once. If you create embeddings for a user query, you can search for embeddings close to it, and these will be "semantically" similar to the query.
The advantage is that unlike traditional full-text search, the user doesn't need a query that includes words present in the text.
- cced 3y agoDo you have any resources that might guide one on doing something like this from scratch?
- djhn 3y agoHere's a 6 minute speed run of something like that on weviate https://youtu.be/mBcBoGhFndY https://youtu.be/mBcBoGhFndY
- password4321 3y agohttps://news.ycombinator.com/item?id=34684593#34689275 https://news.ycombinator.com/item?id=34684593#34689275
- rahimnathwani 3y agohttps://news.ycombinator.com/item?id=36012299 https://news.ycombinator.com/item?id=36012299
- fzliu 3y agoHere's one that may be helpful for you: https://zilliz.com/blog/using-vector-database-search-white-house-speeches https://zilliz.com/blog/using-vector-database-search-white-h...