We investigate the representation and retrieval of noisy texts in general and digitized historical manuscripts in particular. The noise originates from several sources, these include imperfect text recognition, spelling variation, non-standardized grammar, in addition to user-side confusion due to her/his limited knowledge of the underlying language and/or the searched text. Manual correction or normalization are very time consuming and resource-demanding tasks and are thus out of the question. Furthermore, external resources such as thesauri are not available for the older, lesser-known languages. We present several methods that provide a low-cost yet highly-effective text representation to limit the negative impact of recognition errors and the variable orthography and morphology.

The presentation will be given in French with English slides.

La présentation sera donnée en français avec des transparents en anglais.