Liangping Ma

Liangping Ma, John Kaewell

Monte Carlo dropout may effectively capture model uncertainty in deep learning, where a measure of uncertainty is obtained by using multiple instances of dropout at test time. However, Monte Carlo dropout is applied across the whole network and thus significantly increases the computational complexity, proportional to the number of instances. To reduce the computational complexity, at test time we enable dropout layers only near the output of the neural network and reuse the computation from prior layers while keeping, if any, other dropout layers disabled. Additionally, we leverage the side information about the ideal distributions for various input samples to do `error correction' on the predictions. We apply these techniques to the radio frequency (RF) transmitter classification problem and show that the proposed algorithm is able to provide better prediction uncertainty than the simple ensemble average algorithm and can be used to effectively identify transmitters that are not in the training data set while correctly classifying transmitters it has been trained on.

Tianyi Xu, Liangping Ma, Gregory Sternberg

In IEEE 802.11, the retry limit is set the same value for all packets. In this paper, we dynamically classify video teleconferencing packets based on the type of the video frame that a packet carries and the packet loss events that have happened in the network, and assign them different retry limits. We consider the IPPP video encoding structure with instantaneous decoder refresh (IDR) frame insertion based on packet loss feedback. The loss of a single frame causes error propagation for a period of time equal to the packet loss feedback delay. To optimize the video quality, we propose a method to concentrate the packet losses to small segments of the entire video sequence, and study the performance by an analytic model. Our proposed method is implemented only on the stations interested in enhanced video quality, and is compatible with unmodified IEEE 802.11 stations and access points in terms of performance. Simulation results show that the performance gain can be significant compared to the IEEE 802.11 standard without negatively affecting cross traffic.