March 19, 2018
The lab has piled up three publications this month:
-
Our work on modelling and predicting Spatially Spreading Processes, in this case Forest Wildfires, using raw image data. Both these papers highlight different components of our approach of using Reinforcement Learning to automatically learnÌýagent-based models of fire spreading on a landscape from readily avaialble satellite and weather data. the first paperÌýbuilds on outÌýwork from RLDM last year. We will be presenting this work at this year's in Toronto, May 8-11, 2018.:
- Subramanian, S.G. & Crowley, M., 2018.ÌýUsing Spatial Reinforcement Learning to Build Forest Wildfire Dynamics Models from Satellite Images.ÌýFrontiers in ICT: Environmental Informatics.Ìý
- Subramanian, S.G. & Crowley, M., 2018.ÌýCombining MCTS and A3C for Prediction of Spatially Spreading Processes in Forest Wildfire Settings. InÌýCanadian Conference on Artificial Intelligence. Toronto, Ontario, Canada: (to appear).
-
This paper was a collaborative work from three of my students growing out of a course project to build a very impressive multi-layer system forÌýDecision AssistanceÌýfor self-driving cars. Decision assistance is more subtle thanÌýautonomous driving, so take a look at the paper to find out the difference and the new approach weÌýfound that worked well.
- Subramanian, S.G., Singh, J., Ghojogh, B.,ÌýCrowley, M., 2018.ÌýDecision Assist For Self-Driving Cars. InÌýCanadian Conference on Artificial Intelligence. Toronto, Ontario, Canada: (to appear).
Ìý
