Search papers, labs, and topics across Lattice.
Yale University, CT, USA, These authors contribute equally as task contributors
4
0
8
6
Current AI agents excel in structured tasks but falter at generating novel insights and tackling open-ended scientific challenges.
INFUSER outperforms a frozen 32B model with just an 8B co-evolving generator, showcasing the power of adaptive question generation in self-evolution.
Neural networks can achieve exponential convergence to structured representations, revealing a surprising link between training dynamics and representation theory.
Two-layer neural nets solve modular addition via a "flawed indicator function" built from noisy single-frequency Fourier features, where phase symmetry enables a majority-voting noise cancellation scheme.
