Transformer Memory as a Differentiable Search Index: memorizing thousands of random doc ids works!?
Neural Search Talks — Zeta Alpha
00:00
Using the Model as a Way to Store the Documents
With this method, there's this whole sort of universe of efficiency and scaling things that we don't know and don't have intuitions about because it's a pretty new way of doing materials. A lot of these things are hard to make sense of at least in my head. So I'm super interested in seeing where other people take this further,. Because I see so many ways that this can be explored further or applied to other data sets"
Transcript
Play full episode
