Search papers, labs, and topics across Lattice.
The paper introduces iLoRA, a Bayesian graph-conditioned LoRA framework that infers a latent interaction graph from the input to generate input-conditioned LoRA updates. This allows for joint learning of prediction and latent interaction structure, addressing the limitations of static low-rank updates in standard LoRA. Applied to microbiome diagnosis, iLoRA demonstrates improved performance over LoRA and Bayesian baselines in both interactive QA with human-annotated graphs and multi-cohort IBD diagnosis, while also recovering meaningful interaction graphs.
Discovering hidden relationships between microbes could unlock faster, more accurate disease diagnoses, thanks to iLoRA's ability to learn both prediction and interaction structure simultaneously.
Parameter-efficient adaptation has made LLMs practical for domain prediction, but standard LoRA still relies on a static low-rank update and does not expose the latent interactions that often drive scientific labels. We introduce iLoRA. To our knowledge, it is the first Bayesian graph-conditioned LoRA framework. It infers a latent interaction graph from the input and uses it to generate input-conditioned LoRA updates. As a result, iLoRA learns prediction and latent interaction structure jointly, rather than training a predictor and applying interaction analysis only post hoc. We instantiate this idea for microbiome diagnosis, where disease state can depend on both species-level abundance and microbe-microbe cross-talk, and evaluate it in two complementary settings: interactive QA with human-annotated graphs, which tests latent structure recovery, and multi-cohort IBD diagnosis, which tests biomedical utility. Across both settings, iLoRA improves over strong LoRA and Bayesian adaptation baselines, recovers graphs aligned with human annotations and cohort-level microbiome associations, and provides calibrated uncertainty with moderate graph-branch overhead.
