Search papers, labs, and topics across Lattice.
Courant Institute of Mathematical Sciences, School of Engineering and Applied Sciences, New York University, Harvard University, Google Research, Google Research Tel-Aviv University
Google Research2
0
4
Graph transformers can be fundamentally limited by their tokenization strategy, as some tokenizations provably preclude efficient learning of structural representations realizable with other tokenizations.
Autoregressive neural surrogates can now simulate dynamical systems for infinitely long horizons, thanks to a novel self-refining diffusion model that avoids error compounding.
