Teacher Student Curriculum Learning applied to OCR

If you arrive here, you probably come from khipu!

The poster

        

You can download it here.

Contact

Let’s keep in touch! my e-mail is rodrigo.laguna@fing.edu.uy I’m also in Twitter as @ro_laguna_ or in Linkedin.

Contribute to Luisa

You can contribute to the LUISA (Leyendo Unidos para Interpretar loS Archivos) project, labeling some instances here.

LUISA is a collaborative project with the objective of transcribing texts from the period prior to, and during, the Uruguayan dictatorship, which for several reasons may be difficult to transcribe automatically. The project seeks to contribute to the search for truth and justice regarding the acts of State terrorism.

References

• [1] Yoshua Bengio et al. “Curriculum learning”. In: Proceedings of the 26th annual international conference on machine learning. 2009, pp. 41–48.
• [2] Felipe Chavat Pérez. “Modelos Seq2Seq para la transcripción de documentos del Archivo Berrutti”. In: (2022).
• [3] Jeffrey L Elman. “Learning and development in neural networks: The importance of starting small”. In: Cognition 48.1 (1993), pp. 71–99.
• [4] Tambet Matiisen et al. “Teacher-student curriculum learning”. In: IEEE transactions on neural networks and learning systems (2019).

The experiments presented in this poster were carried out using ClusterUY.