Search papers, labs, and topics across Lattice.
Affiliation:
2
0
3
0
Reward hacking is mitigated and sample quality preserved in few-step discrete diffusion models by replacing intractable KL penalties with a novel inverse-distillation regularization.
An AI agent can now autonomously design functional antibodies from text prompts with a 67% lab-validated success rate, compressing weeks of expert work into hours.
