@misc{sumner_he_thakkar_engkvist_bjerrum_2020, 
  title={Levenshtein Augmentation Improves Performance of SMILES Based Deep-Learning Synthesis Prediction}, 
  DOI={10.26434/chemrxiv.12562121.v2}, 
  abstractNote={<jats:p>&lt;p&gt;SMILES
randomization, a form of data augmentation, has previously been shown to
increase the performance of deep learning models compared to non-augmented
baselines. Here, we propose a novel data augmentation method we call "Levenshtein
augmentation" which considers local SMILES sub-sequence similarity between
reactants and their respective products when creating training pairs. The performance
of Levenshtein augmentation was tested using two state of the art models -
transformer and sequence-to-sequence based recurrent neural networks with
attention. Levenshtein augmentation demonstrated an increase  performance over non-augmented, and
conventionally SMILES randomization augmented data when used for training of
baseline models. Furthermore, Levenshtein augmentation seemingly results in
what we define as &lt;i&gt;attentional gain &lt;/i&gt;– an
enhancement in the pattern recognition capabilities of the underlying network
to molecular motifs.&lt;/p&gt;</jats:p>}, 
  publisher={American Chemical Society (ACS)}, 
  author={Sumner, Dean and He, Jiazhen and Thakkar, Amol and Engkvist, Ola and Bjerrum, Esben Jannik}, 
  year={2020}, 
  month={Jul}
  }