Szczegóły publikacji
Opis bibliograficzny
The landscape of foundation models for molecular chemistry / Mateusz PRASKI // W: CIKM'25 : proceedings of the 34th ACM international conference on Information and Knowledge Management : November 10 - 14, 2025, Seoul, Republic of Korea / eds. Meeyoung Cha, [et al.]. — New York : Association for Computing Machinery, Inc. (ACM), cop. 2025. — ISBN: 979-8-4007-2040-6. — S. 6805–6808. — Wymagania systemowe: Adobe Reader. — Bibliogr. s. 6808, Abstr. — Publikacja dostępna online od: 2025-11-10
Autor
Słowa kluczowe
Dane bibliometryczne
| ID BaDAP | 165918 |
|---|---|
| Data dodania do BaDAP | 2026-02-09 |
| Tekst źródłowy | URL |
| DOI | 10.1145/3746252.3761663 |
| Rok publikacji | 2025 |
| Typ publikacji | materiały konferencyjne (aut.) |
| Otwarty dostęp |  |
| Creative Commons | |
| Wydawca | Association for Computing Machinery (ACM) |
| Konferencja | ACM International Conference on Information and Knowledge Management 2025 |
Abstract
Pre-trained neural networks have recently emerged as powerful tools for molecular data mining, offering an alternative to classical approaches. However, these models are often evaluated on limited datasets with narrow baselines, leaving their benefits unclear. We present the first large-scale benchmark comparing pre-trained molecular embedding models across 20 public datasets spanning classification and regression tasks. Our evaluation covers text-based, graph-based, and multimodal architectures, all tested under a unified methodology. The results show that the classical fingerprint-based models remain highly competitive. Only a few models consistently exceeded the baseline. We also highlight key factors influencing model performance, offering practical guidance for model selection and future improvements in molecular embeddings.
