Annotation tools for artificial intelligence training

More and more often, automatic entity recognition, i.e. NER (Named Entity Recognition), is required when performing machine learning tasks. These entities can be pre-defined, such as names of geographical locations, organisations, dates and times, or entities that are relevant to a specific sector.

But should a company need to profile its users on the basis of, say, the interests mentioned on social networks, specific entities would be required, for instance books, films or song titles. The fields of application of NER are very diverse, ranging from the extraction of entities from legal, financial and medical documents to the automatic classification of news content and the implementation of sophisticated recommendation systems. While automatic annotation is increasingly necessary to customise your business, the annotation tools available on the market are not always easy to use and often slow down the workflow.

Presago has focused on maximising the user experience by developing an annotator that allows large amounts of text to be labelled very quickly and easily, and that can be confidently used by non-experts too.

The annotation tool allows to define a customised list of entities to be used for tagging, and to select the documents to be annotated in a random way, in order to avoid the emergence of biases that may affect the performance of machine learning systems. Labels can be applied with a simple click by selecting a word or block of text. Each label is assigned a different colour. This makes switching from one tag to another quicker and more intuitive, and above all, it makes the control of the labelling easier.

Finally, the tool allows you to set targets and download the annotated documents to feed them to machine learning engines to produce the desired classification models.

If you need to annotate large amounts of data with a fast and reliable tool, contact us for a free consultation. We can suggest the best solution for your needs.