BIASLAB

Bayesian Intelligent Autonomous Systems

About

We are an academic research team in the Signal Processing Systems group (EE department) at Eindhoven University of Technology. Our main research interest is in developing autonomous agents that teach themselves to behave intelligently through interactions with the environment. Typical application areas include (medical) device personalization, automated vehicle control and robotics.

Mission

Learning through environmental interactions also goes by the name Reinforcement Learning. Our research on reinforcement learning is strongly inspired by developments in Bayesian Inference and Computational Neurosciences. Ultimately, we aim to develop a fully Bayesian reinforcement learning framework that is realized by message passing algorithms in highly modular Graphical Model networks.

Approach

Our research projects are usually selected by our quest to develop (reinforcement learning) technology to support in-situ hearing aid personalization. We ‘eat our own dogfood’ by immersing ourselves in complex acoustic situations and use our own tools to optimize audio processing algorithms under in-situ conditions. This leads to new problems that drive our next research steps.

Meet the team

Bayesian Pure-tone Audiometry

The pure-tone hearing threshold is usually estimated from responses to stimuli at a set of standard frequencies. In this project we develop a probabilistic approach to the estimation problem in which the hearing threshold is modeled as a smooth continuous function of frequency using a Gaussian process. This allows sampling at any frequency and reduces the number of required measurements. The Gaussian process is combined with a probabilistic response model to account for uncertainty in the responses. The resulting full model can be interpreted as a two-dimensional binary classifier for stimuli, and provides uncertainty bands on the estimated threshold curve. The optimal next stimulus is determined based on an information-theoretic criterion. This leads to a robust adaptive estimation method that can be applied to fully automate the hearing threshold estimation process.

Probabilistic Hearing Loss Compensation

Hearing Aid (HA) algorithms need to be tuned (fitted) to match the impairment of each specific patient. The lack of a fundamental HA fitting theory is a strong contributing factor to an unsatisfying sound experience for about 20% of hearing aid patients. We propose here a probabilistic modeling approach to the design of HA algorithms. Our method relies on a generative probabilistic model for the hearing loss problem and provides for automated inference of the corresponding (1) signal processing algorithm, (2) the fitting solution as well as a principled (3) performance evaluation metric. All three tasks are realized as message passing algorithms in a factor graph representation of the generative model, which in principle allows for fast implementation on hearing aid or mobile device hardware. The methods are theoretically worked out and simulated with a custom-built factor graph toolbox for a specific hearing loss model.