Candidate fusion: Integrating language modelling into a sequence-to-sequence handwritten word recognition architecture

Sequence-to-sequence models have recently become very popular for tackling handwritten word recognition problems. However, how to effectively integrate an external language model into such recognizer is still a challenging problem. The main challenge while training deal with the corpus which usually different one used system. Thus, bias between both corpora leads incorrectness on transcriptions, providing similar or even worse performances task. In this work, we introduce Candidate Fusion, novel way sequence-to-sequence architecture. Moreover, it provides suggestions from knowledge, as new input recognizer. Hence, Fusion two improvements. On hand, has flexibility not only combine information itself and model, but also choose importance of provided by model. other ability adapt learn most common errors produced Finally, conducting comprehensive experiments, proves outperform state-of-the-art tasks.