Search papers, labs, and topics across Lattice.
The paper introduces SME-HGT, a Heterogeneous Graph Transformer framework, to predict the likelihood of Small Business Innovation Research (SBIR) Phase I awardees progressing to Phase II funding. SME-HGT constructs a heterogeneous graph of companies, research topics, and government agencies, leveraging their relational structure to predict SME potential. The model achieves an AUPRC of 0.621 on a temporally-split test set, outperforming MLP and R-GCN baselines, and demonstrates high precision (89.6%) at a screening depth of 100 companies.
A heterogeneous graph transformer can predict high-potential SMEs with nearly 90% precision from publicly available data, offering a scalable alternative to traditional due diligence.
Small and Medium Enterprises (SMEs) constitute 99.9% of U.S. businesses and generate 44% of economic activity, yet systematically identifying high-potential SMEs remains an open challenge. We introduce SME-HGT, a Heterogeneous Graph Transformer framework that predicts which SBIR Phase I awardees will advance to Phase II funding using exclusively public data. We construct a heterogeneous graph with 32,268 company nodes, 124 research topic nodes, and 13 government agency nodes connected by approximately 99,000 edges across three semantic relation types. SME-HGT achieves an AUPRC of 0.621 0.003 on a temporally-split test set, outperforming an MLP baseline (0.590 0.002) and R-GCN (0.608 0.013) across five random seeds. At a screening depth of 100 companies, SME-HGT attains 89.6% precision with a 2.14 lift over random selection. Our temporal evaluation protocol prevents information leakage, and our reliance on public data ensures reproducibility. These results demonstrate that relational structure among firms, research topics, and funding agencies provides meaningful signal for SME potential assessment, with implications for policymakers and early-stage investors.
