# Index

### Algorithm

* [**A Few Useful Things to Know About Machine Learning**](https://homes.cs.washington.edu/~pedrod/papers/cacm12.pdf) - Domingos CACM '12 \[[Summary](https://xzhu0027.gitbook.io/blog/ml-system/ml)]
* [**Hidden Technical Debt in Machine Learning Systems**](https://papers.nips.cc/paper/5656-hidden-technical-debt-in-machine-learning-systems.pdf) - Sculley et al, NIPS '15 \[[Summary](https://xzhu0027.gitbook.io/blog/ml-system/hidden-technical-debt-in-machine-learning-systems)]
  * Identified the technical debt of Machine Learning system and provided some common mitigation strategy.
* [**Don't Decay The Learning Rate, Increase The Batch Size**](https://openreview.net/pdf?id=B1Yy1BxCZ)\* - Smith et al, ICLR '18 
  * Showed that we can increase the batch size, instead of decreasing the learning rate, to get equivalent test accuracies after the same number of training epochs, but with fewer parameter updates, leading to greater parallelism and shorter training times.
* [**Accelerating Deep Learning by Focusing on the Biggest Losers**](https://arxiv.org/abs/1910.00762) - Jiang et al., arXiv '19 \[[Summary](https://xzhu0027.gitbook.io/blog/ml-system/accelerating-deep-learning-by-focusing-on-the-biggest-losers)]
  * Proposes an algorithm that accelerates the training of deep neural networks (DNNs) by prioritizing examples with high loss at each iteration. 

### Faster Inference

* [**End-to-end Learning of Action Detection from Frame Glimpses in Videos**](https://arxiv.org/pdf/1511.06984.pdf) - Yeung et al., CVPR '16
* [**Watching a Small Portion could be as Good as Watching All: Towards Efficient Video Classification**](https://www.ijcai.org/Proceedings/2018/0098.pdf) - Fan et al., IJCAI '18
* [**AdaFrame: Adaptive Frame Selection for Fast Video Recognition**](https://arxiv.org/abs/1811.12432) - Wu et al., CVPR '19
* [**Multi-Agent Reinforcement Learning Based Frame Sampling for Effective Untrimmed Video Recognition**](https://arxiv.org/abs/1907.13369) - Wu et al., ICCV  '19
* [**Listen to Look: Action Recognition by Previewing Audio**](https://arxiv.org/abs/1912.04487) - Gao et al., CVPR '20
* [**FastBERT: a Self-distilling BERT with Adaptive Inference Time**](https://www.aclweb.org/anthology/2020.acl-main.537.pdf) - Liu et al., ACL '20

### Misc.

* [**Beyond Accuracy: Behavioral Testing of NLP Models with CheckList** ](https://homes.cs.washington.edu/~marcotcr/acl20_checklist.pdf)- Ribeiro et al., ACL '20


---

# Agent Instructions: 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:

```
GET https://xzhu0027.gitbook.io/blog/machine-learning/untitled.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
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.