Improving antibody optimization ability of generative adversarial network through large language model

A comprehensive overview of recent advances in generative models for antibodies

“…AbGAN-LMG is used to generate high-quality antibody libraries. Different from traditional GAN, it uses language models as input and combines the representation ability of language models to improve the effect of generating high-quality antibodies [57] . GANs inherently have a certain degree of randomness, which allows them to explore a wider space in the process of sequence generation [58] .…”