Reproduce “Gradient Boosting Neural Networks: GrowNet” on OpenResearch

Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.

Universal app for Apple silicon and Intel. Requires macOS 11 or later.

About this paper

Sarkhan Badirli, Xuanqing Liu, Zhengming Xing, Avradeep Bhowmik, Khoa Doan, Sathiya S. Keerthi

GrowNet builds structured-data predictors from shallow neural networks, adding learners sequentially and using a corrective step to update earlier predictors. On MSLR-WEB 10K, its pairwise-loss NDCG@10 score is 0.5203 versus XGBoost’s 0.4858. GrowNet also exceeds listed comparators on both ranking datasets, but on Higgs with 1% of data its AUC of 0.8180 falls below XGBoost’s 0.8304.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.