Search papers, labs, and topics across Lattice.
Imperial College London
1
0
3
Continual learning is achieved without any parameter updates by performing in-context Bayesian inference over a learned latent concept space.
