liveBook · Manning

Exact-match accuracy

As explained in Build AI Drug Discovery Pipelines

Exact match rate is the proportion of reconstructed strings that are identical to their reference strings.

Exact-match accuracy is the proportion of reconstructed strings that are identical to their reference strings. For molecular autoencoders, it checks whether the predicted SMILES reproduces the input exactly, rather than merely producing a valid or structurally similar molecule. section 10.2.4

Interpret the score carefully: exact matching is meaningful only when both strings use the same canonicalization. Different valid SMILES can represent the same molecule, so non-canonical or randomized strings can produce mismatches despite chemical equivalence. section 10.2.4

MOSES reconstruction result

A BasicAutoencoder achieves an exact match rate of 0.847 on canonical MOSES SMILES, meaning 84.7% of reconstructed strings exactly reproduce their canonical inputs. Compare this with valid reconstruction rate and Tanimoto similarity: exact matching is stricter than chemical validity or structural resemblance. section 10.2.4

How Build AI Drug Discovery Pipelines covers it

How deep it goes
Defines the metric, explains canonicalization limits, and contrasts it with validity and structural similarity.
Examples
Uses a BasicAutoencoder on the MOSES benchmark, reporting an exact match rate of 0.847.
What it assumes
Assumes familiarity with SMILES strings and distinguishes exact textual identity from chemical validity and structural similarity.
What you can do after
Evaluate/understand reconstruction fidelity while checking whether reference and output strings share canonicalization.
How it connects
Connects exact matching with SMILES, canonicalization, valid reconstruction rate, token accuracy, and Tanimoto similarity.

Also explained in

Compare how 8 books explain exact-match accuracy →

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Build AI Drug Discovery Pipelines. The section links point to the parts of the book it draws on.