### November 9, 2020

**Luc Dartois**(LACL)

Deterministic two-way transducers define the robust class of regular functions which is, among other good properties, closed under composition. However, the best known algorithms for composing two-way transducers cause a triple exponential blow-up in the size of the inputs.

In this talk, I will present the class of reversible transducers, which are machines that are both deterministic and co-deterministic. This class enjoys polynomial composition complexity, even in the two-way case.

Although this class is not very expressive in the one-way scenario, I will show that any two-way transducer can be made reversible through a single exponential blow-up. As a consequence, the composition of two-way transducers can be done with a single exponential blow-up in the number of states, enhancing the known algorithm from the 60s.