"The regular use of an adequate hearing aid increases the chance of keeping hearing impaired patients communicatively, socially and economically active. Many profoundly deaf patients are potential candidates for cochlear implantation. The optimal use of such a device requires that the cochlear implant speech processor be adjusted so that sounds perceived by the patient are representational and at a comfortable level. In prior collaborative research, new fitting processes have been developed to optimize the patient’s hearing in a more efficient and accurate way by means of an assisted or (semi-)automated fitting procedure in which a large number of the cochlear implant speech processor parameters may be adjusted, based on measured psycho-acoustic feedback from the implant user. Such an assisted fitting process drastically reduces the number of man-hours of fitting during the lifetime of the device with qualitatively better outcomes. However, the state-of-the-art still has a number of short-comings: (i) the self-learning character of the currently used fitting model; and (ii) the limited input data with respect to speech perception testing in very young child populations.The main objectives of the proposed research project are to enhance researchers’ knowledge (i) of fitting models based on Bayesian networks to reinforcement learning models such as partially-observable Markov decision processes (POMDPs); (ii) of evaluation tools to measure functional hearing capacities in ‘difficult’ listeners such as young children or elderly adults.Due to its multi-disciplinary and combined fundamental and applied nature, the proposed research crucially depends on the transfer of knowledge to create a common framework for hearing scientists bringing together insights from different disciplines targeting current issues in speech perception such as linguistics, biomedical physics, mathematics and audiology."