CS 244 ’18: Recreating and Extending Pensieve

Paul Crews, Hudson Ayers

Original Paper: Mao, Hongzi, Ravi Netravali, and Mohammad Alizadeh. “Neural adaptive video streaming with pensieve.” Proceedings of the Conference of the ACM Special Interest Group on Data Communication. ACM, 2017.

For our CS 244 final project, we present the recreation of two results from the paper Neural Adaptive Streaming with Pensieve. First, we present a recreation of Figure 11, which compares Pensieve to other ABR algorithms across three different network types. Next, we present a recreation of Figure 13, which compares Pensieve algorithms trained on multiple simulated videos to Pensieve algorithms trained on a single test video. In order to gain additional insight into the robustness and flexibility of Pensieve ABR algorithms, we extended the experiments conducted by the original authors. Specifically, we tested each algorithm in Figure 11 on additional network types, and we trained and evaluated Pensieve algorithms using additional test videos beyond the lone video discussed in the original paper.


Final Report.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s