> For the complete documentation index, see [llms.txt](https://xzhu0027.gitbook.io/blog/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://xzhu0027.gitbook.io/blog/video/index/short-summaries.md).

# Short Summaries

### [**Potluck: Cross-Application Approximate Deduplication for Computation-Intensive Mobile Applications**](https://www.cs.yale.edu/homes/guo-peizhen/files/potluck-asplos18.pdf) - Guo et al., ASPLOS' 18

This paper proposes a caching mechanism to enable cross-application approximate deduplication. The feature vector of input(e.g, a frame) serves as the key. Thus, a lookup attempt is made with the key and the name of the function called by matching the input key to any existing key within a given similarity threshold. To strike the trade-off between performance speedup from reusing previous results and the accuracy of the results, Potluck adaptively tunes the similarity threshold. One thing to note is that Potluck only saves the result of the model and does not save any intermediate results within a model.

### [DeepCache: Principled Cache for Mobile Deep Vision](https://dl.acm.org/doi/10.1145/3241539.3241563) - Xu et al., MobiCom' 18

This paper aims to cache CNN's final results as well as its intermediate results, exploiting the temporal locality among a video stream. In other words, consecutive video frames often have substantial similar overlapped regions.

For a CNN model, DeepCache maintains a cache, covering the model’s input as well as its internal layers. The cache stores recent video frames for the model input, and recent feature maps for the internal layers. The cache keys are equal-sized, fine-grained regions on the cached input frames. The cache values are the cached feature maps produced by the layers.

Note that, Deep cache is inspired by video compression techniques and aims to cache results for a single application.

###


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:

```
GET https://xzhu0027.gitbook.io/blog/video/index/short-summaries.md?ask=<question>&goal=<endgoal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.