Solving Learn-to-Race Autonomous Racing Challenge by Planning in Latent Space
Authors:
Shivansh Beohar,
Fabian Heinrich,
Rahul Kala,
Helge Ritter,
Andrew Melnik
Abstract:
Learn-to-Race Autonomous Racing Virtual Challenge hosted on www<dot>aicrowd<dot>com platform consisted of two tracks: Single and Multi Camera. Our UniTeam team was among the final winners in the Single Camera track. The agent is required to pass the previously unknown F1-style track in the minimum time with the least amount of off-road driving violations. In our approach, we used the U-Net archite…
▽ More
Learn-to-Race Autonomous Racing Virtual Challenge hosted on www<dot>aicrowd<dot>com platform consisted of two tracks: Single and Multi Camera. Our UniTeam team was among the final winners in the Single Camera track. The agent is required to pass the previously unknown F1-style track in the minimum time with the least amount of off-road driving violations. In our approach, we used the U-Net architecture for road segmentation, variational autocoder for encoding a road binary mask, and a nearest-neighbor search strategy that selects the best action for a given state. Our agent achieved an average speed of 105 km/h on stage 1 (known track) and 73 km/h on stage 2 (unknown track) without any off-road driving violations. Here we present our solution and results.
△ Less
Submitted 5 July, 2022; v1 submitted 4 July, 2022;
originally announced July 2022.
Multimodal Meta-Learning for Time Series Regression
Authors:
Sebastian Pineda Arango,
Felix Heinrich,
Kiran Madhusudhanan,
Lars Schmidt-Thieme
Abstract:
Recent work has shown the efficiency of deep learning models such as Fully Convolutional Networks (FCN) or Recurrent Neural Networks (RNN) to deal with Time Series Regression (TSR) problems. These models sometimes need a lot of data to be able to generalize, yet the time series are sometimes not long enough to be able to learn patterns. Therefore, it is important to make use of information across…
▽ More
Recent work has shown the efficiency of deep learning models such as Fully Convolutional Networks (FCN) or Recurrent Neural Networks (RNN) to deal with Time Series Regression (TSR) problems. These models sometimes need a lot of data to be able to generalize, yet the time series are sometimes not long enough to be able to learn patterns. Therefore, it is important to make use of information across time series to improve learning. In this paper, we will explore the idea of using meta-learning for quickly adapting model parameters to new short-history time series by modifying the original idea of Model Agnostic Meta-Learning (MAML) \cite{finn2017model}. Moreover, based on prior work on multimodal MAML \cite{vuorio2019multimodal}, we propose a method for conditioning parameters of the model through an auxiliary network that encodes global information of the time series to extract meta-features. Finally, we apply the data to time series of different domains, such as pollution measurements, heart-rate sensors, and electrical battery data. We show empirically that our proposed meta-learning method learns TSR with few data fast and outperforms the baselines in 9 of 12 experiments.
△ Less
Submitted 2 November, 2021; v1 submitted 5 August, 2021;
originally announced August 2021.
