liveBook · Manning

Greedy algorithm

As explained in Build AI Drug Discovery Pipelines

Greedy decoding generates a sequence by selecting the highest-probability token at every step.

Greedy decoding generates a sequence by choosing the highest-probability token at each step. During SMILES generation, the decoder uses its own predictions until it produces an end-of-sequence token. section 10.4.3

This strategy is fast, but it can produce suboptimal sequences. The choice between greedy decoding and temperature sampling affects the quality and diversity of generated molecules. section 10.4.3

How Build AI Drug Discovery Pipelines covers it

How deep it goes
Defines highest-probability token selection and notes speed and sequence-quality tradeoffs without detailing its implementation.
Examples
Contrasts greedy decoding with temperature sampling when generating SMILES from a latent vector.
What you can do after
Evaluate generation strategies by recognizing that greedy decoding is fast but may produce suboptimal sequences.
How it connects
Connects SMILES generation with decoder inference and temperature sampling.

Also explained in

Compare how 16 books explain greedy algorithm →

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.