Reproduce “Seahorse: A Unified Benchmarking Framework for Spatiotemporal Event Modeling” 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
Yahya Aalaila, Gerrit Großmann, Sebastian Vollmer
SEAHORSE provides a common framework for comparing spatiotemporal point-process models under fixed data splits and held-out likelihood reporting in original coordinates. It preserves model-specific architectures and training objectives while recording likelihood methods. On HawkesNest’s hardest entanglement level, NSMPP has lower test NLL than AutoSTPP, but AutoSTPP and DeepSTPP better correlate with the known intensity surface; findings are limited to tested regimes.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.