Module 54: Applying Reinforcement Learning to Control Theory

Faculty Contact: Yu-Xiang Wang

Research Areas:

We would like to be able to parameterize a state space model by using an LSTM, much in the same way that was done by Rangapuram in “Deep State Space Models for Time Series Forecasting” Our original idea was to use the deep state space model as a method of dialog state tracking, such as in the paper by Wen et. al., “Latent Intention Dialog Models,” presented at ICML 2017.


Active Quarters:

  • Winter 2019: Alon Albalak