Research map / Label-efficient and robust learning / Recognition and long-tail
MapLE: Matching Molecular Analogues Promptly with Low Computational Resources by Multi-Metrics Evaluation (Student Abstract)
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's full text. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Proposes MapLE, a multi-metric molecular analogue matching framework with inverted-list indexing and progressive top-k evaluation to speed up search under low computational resources.
Problem
Matching molecular analogues quickly over large chemical libraries is costly; existing methods focus on accuracy and slow down with large databases, which is impractical under low computational resources.
Approach
Combine multiple pharmacological, structural (topological sub-metrics) and chemical similarity metrics via weighted Tanimoto fusion, store features in inverted lists, and progressively output top-k results using a threshold so that only the top of each metric's ranked list is traversed rather than the whole database.
Key results
- On CASF-2016, MapLE's query time grows with k and is markedly lower than the baseline that accesses all ligands, especially for small k (top-1 is fastest).
- The paper reports fewer candidate molecules accessed by MapLE than the exhaustive baseline, with the gap largest at small k.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use MapLE’s progressive prompt evaluation instead of accessing all ligand molecules | MapLE · CASF-2016 | query time | Helped |
| Use MapLE’s progressive prompt evaluation instead of accessing all ligand molecules | MapLE · CASF-2016 | candidates accessed | Helped |
Limitations
- Evaluated only on a single dataset (CASF-2016) with a single baseline
- No accuracy/effectiveness numbers reported in the extract; efficiency is the main focus
- Hardware and quantitative results are only shown in figures
Experimental setup
Datasets: CASF-2016
Benchmarks: CASF-2016
Metrics: number of candidates accessed query time for different k
Baselines: exhaustive baseline accessing all ligands
Builds on: Arthor RoundTable algorithm
Closest papers by meaning
- UniMatch: Universal Matching from Atom to Task for Few-Shot Drug Discovery (ICLR 2025 · similarity 0.84)
- Offline Model-based Optimization for Real-World Molecular Discovery (ICML 2025 · similarity 0.84)
- Atomas: Hierarchical Adaptive Alignment on Molecule-Text for Unified Molecule Understanding and Generation (ICLR 2025 · similarity 0.83)
- Representing Molecules as Random Walks Over Interpretable Grammars (ICML 2024 · similarity 0.83)
- Hierarchical Graph Tokenization for Molecule-Language Alignment (ICML 2025 · similarity 0.83)
- Efficient Evolutionary Search Over Chemical Space with Large Language Models (ICLR 2025 · similarity 0.83)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.