rkt08

3 ms·

1. 1.▲
   by rkt08 5y ago

   Very good point. This is an active area of research, there's been some work in using other similarity measures like MI,Kendall's Tau etc in embeddings as well. And they have been found to perform pretty well. I found something for
2. 2.▲
   by rkt08 5y ago

   How about a vector oriented 'database' instead? Pinecone( https://www.pinecone.io/ ) does both exact and approx search and it's fully managed so you don't have to worry about reliability,availability etc.
3. 3.▲
   by rkt08 5y ago

   That is definitely an interesting use case. You can look for methods that combine multiple vectors to create a combined vector or you may want to filter on metadata and vectors both. Who knows maybe you might be able to come with something
4. 4.▲
   by rkt08 5y ago

   Any search domain is good for vector search if your method for producing vectors is good! Seems like it should be possible with Recipes.Are you performing search based on recipe text? It shouldn't take long to try out a simple POC with
5. 5.▲
   by rkt08 6y ago

   Looking at the VGGish paper itself, I see they use spectrograms as inputs, they show results where they can identify instrument types. I'm not too sure how specific embeddings from these models can be. Do we know if spectrograms can di