《JavaScript深度學習》資源連結清單

為方便讀者查詢，本文彙總了《JavaScript深度學習》一書中用到的部分網路資源連結。連結中內容可能隨時間變化，請讀者知悉。

關於本書

• 本書GitHub程式碼地址：https://github.com/tensorflow/tfjs-examples

第1章

表1-1中參考文獻

• a. Kaiming He et al., “Deep Residual Learning for Image Recognition,” Proc. IEEE Conference Computer Vision and Pattern Recognition (CVPR), 2016, pp. 770–778, http://mng.bz/PO5P.
• b. Christian Szegedy et al., “Going Deeper with Convolutions,” Proc. IEEE Conference Computer Vision and Pattern Recognition (CVPR), 2015, pp. 1–9, http://mng.bz/JzGv.
• c. Large Scale Visual Recognition Challenge 2017 (ILSVRC2017) results, http://image-net.org/challenges/LSVRC/2017/results.
• d. Yunpeng Chen et al., “Dual Path Networks,” https://arxiv.org/pdf/1707.01629.pdf.
• e. Yonghui Wu et al., “Google’s Neural Machine Translation System: Bridging the Gap between Human and Machine Translation,” submitted 26 Sept. 2016, https://arxiv.org/abs/1609.08144.
• f. Chung-Cheng Chiu et al., “State-of-the-Art Speech Recognition with Sequence-to-Sequence Models,” submitted 5 Dec. 2017, https://arxiv.org/abs/1712.01769.
• g. Volodymyr Mnih et al., “Playing Atari with Deep Reinforcement Learning,” NIPS Deep Learning Workshop 2013, https://arxiv.org/abs/1312.5602.
• h. David Silver et al., “Mastering Chess and Shogi by Self-Play with a General Reinforcement Learning Algorithm,” submitted 5 Dec. 2017, https://arxiv.org/abs/1712.01815.
• i. Varun Gulshan et al., “Development and Validation of a Deep Learning Algorithm for Detection of Diabetic Retinopathy in Retinal Fundus Photographs,” JAMA, vol. 316, no. 22, 2016, pp. 2402–2410, http://mng.bz/wlDQ.

第2章

• 本章CodePen程式碼：https://codepen.io/collection/Xzwavm/

波士頓房價資料

• https://storage.googleapis.com/tfjs-examples/multivariate-linear-regression/data/train-data.csv
• https://storage.googleapis.com/tfjs-examples/multivariate-linear-regression/data/train-target.csv
• https://storage.googleapis.com/tfjs-examples/multivariate-linear-regression/data/test-data.csv
• https://storage.googleapis.com/tfjs-examples/multivariate-linear-regression/data/test-target.csv

第5章

• 5.1.3節音訊資料集：http://mng.bz/POGY

第6章

• 圖6-3的CodePen程式碼：https://codepen.io/tfjs-book/pen/MLQOem
• 圖6-4的CodePen程式碼：https://codepen.io/tfjs-book/pen/JxpMrj
• 加利福尼亞州住房資料集說明：https://developers.google.com/machine-learning/crash-course/california-housing-data-description

第7章

• 7.1.1節CodePen示例：https://codepen.io/tfjs-book/pen/BvzMZr
• 7.1.2節氣象資料集：https://www.kaggle.com/pankrzysiu/weather-archive-jena
• 7.3節的參考文獻：Marco Tulio Ribeiro, Sameer Singh, and Carlos Guestrin, “Why Should I Trust You? Explaining the Predictions of Any Classifier,” 2016, https://arxiv.org/pdf/1602.04938.pdf.
• 7.3節TensorFlow.js的tSNE庫（https://github.com/tensorflow/tfjs-tsne）

第8章

• 表8-1中批次標準化（batch normalization）主要的可調引數：https://js.tensorflow.org/api/latest/#layers.batchNormalization

第9章

• 資訊欄9-2中的Embedding Projector工具：https://projector.tensorflow.org
• GloVe：https://nlp.stanford.edu/projects/glove/

9.4 延伸閱讀

• Chris Olah, “Understanding LSTM Networks,” blog, 27 Aug. 2015, http://mng.bz/m4Wa.
• Chris Olah and Shan Carter, “Attention and Augmented Recurrent Neural Networks,” Distill, 8 Sept. 2016, https://distill.pub/2016/augmented-rnns/.
• Andrej Karpathy, “The Unreasonable Effectiveness of Recurrent Neural Networks,” blog, 21 May 2015, http://mng.bz/6wK6.
• Zafarali Ahmed, “How to Visualize Your Recurrent Neural Network with Attention in Keras,” Medium, 29 June 2017, http://mng.bz/6w2e.
• Jason Brownlee, “How to Implement a Beam Search Decoder for Natural Language Processing,” 5 Jan. 2018, https://machinelearningmastery.com/beam-search-decoder-naturallanguage-processing/.
• Stephan Raaijmakers, Deep Learning for Natural Language Processing, Manning Publications, in press, https://www.manning.com/books/deep-learning-for-natural-language-processing.

第10章

• 谷歌Magenta專案中使用的Performance-RNN模型：https://magenta.tensorflow.org/performance-rnn
• David Ha和Douglas Eck的Sketch-RNN專案：http://mng.bz/omyv
• 克勞德·夏農1951年發表的論文：http://mng.bz/5AzB
• 10.3節開頭，英偉達的StyleGAN模型網站：https://thispersondoesnotexist.com
• Hao-Wen Dong等人的MuseGAN專案：https://salu133445.github.io/musegan/
• 10.3.4節提前準備好的示例程式：http://mng.bz/4eGw
• 延展閱讀中的GAN Lab：https://poloclub.github.io/ganlab/

第11章

• 11.1節，OpenAI Gym網站中的“Humanoid environment”問題：https://gym.openai.com/envs/Humanoid-v2/
• 11.2.2節，MIT OpenCourseWare網站上由Russ Tedrake編寫的MIT控制論公開課（6.832 Underactuated Robotics, Spring 2009）的教案：https://ocw.mit.edu/courses/electrical-engineering-and-computer-science/6-832-underactuated-robotics-spring-2009/readings/MIT6_832s09_read_ch03.pdf
• 延展閱讀，倫敦大學學院的David Silver編寫的強化學習課堂筆記“UCL Course on RL”：http://www0.cs.ucl.ac.uk/staff/d.silver/web/Teaching.html

第12章

• 12.2.2節表12-3的相關程式碼：https://github.com/tensorflow/tfjs/tree/master/tfjs/integration_tests/
• 12.3.6節，TensorFlow.js微信小程式外掛：https://github.com/tensorflow/tfjs-wechat

12.4 延展閱讀

• Denis Baylor et al.,“TFX: A TensorFlow-Based Production-Scale Machine Learning Platform,” KDD 2017, https://www.kdd.org/kdd2017/papers/view/tfx-a-tensorflow-based-production-scale-machine-learning-platform.

• Raghuraman Krishnamoorthi,“Quantizing Deep Convolutional Networks forEfficient Inference: A Whitepaper,” June 2018, https://arxiv.org/pdf/1806.08342.pdf.

• Rasmus Munk Larsen and Tatiana Shpeisman,“TensorFlow Graph Optimization,” https://ai.google/research/pubs/pub48051.

第13章

13.2.3節

• @tensorflow-models/coco-ssd：https://www.npmjs.com/package/@tensorflow-models/coco-ssd
• face-api.js：https://github.com/justadudewhohacks/face-api.js
• handtrack.js：https://www.npmjs.com/package/handsfree
• @tensorflow-models/posenet：https://www.npmjs.com/package/@tensorflow-models/posenet
• @tensorflow-models/speech-commands：https://www.npmjs.com/package/@tensorflow-models/speech-commands
• @tensorflow-models/toxicity：https://www.npmjs.com/package/@tensorflow-models/toxicity
• @tensorflow-models/universal-sentence-encoder：https://www.npmjs.com/package/@tensorflow-models/universal-sentence-encoder
• @magenta/music：https://www.npmjs.com/package/@magenta/music

13.4.3 探索TensorFlow.js生態

• 使用TensorFlow.js的主要參考資料是官方的線上文件：https://www.tensorflow.org/js/。最新的詳細API文件位於https://js.tensorflow.org/api/latest/。

• 你可以在Stack Overflow平臺上用“tensorflow.js”標籤諮詢關於TensorFlow.js的問題：https://stackoverflow.com/questions/tagged/tensorflow.js。

• 關於TensorFlow.js的一般討論，推薦使用這個谷歌網上論壇（Google Groups）：https://groups.google.com/a/tensorflow.org/forum/#!forum/tfjs。

附錄A

A.1 在Linux上安裝tfjs-node-gpu

• NVIDIA Developer網站頁面“推薦開發者使用的GPU”：https://developer.nvidia.com/zh-cn/cuda-gpus
• Node.js下載網址：https://nodejs.org/en/download/
• 下載CUDA工具包（CUDA Toolkit）：https://developer.nvidia.com/cuda-downloads
• 從NVIDIA Developer網站下載CuDNN：https://developer.nvidia.com/zh-cn/cudnn

A.2 在Windows上安裝tfjs-node-gpu

• 安裝前須知：https://docs.nvidia.com/cuda/cuda-installation-guide-microsoft-windows/index.html#system-requirements

附錄B

• TensorFlow.js支援大量的張量運算，可以在TensorFlow官方文件中檢視相關介紹：https://js.tensorflow.org/api/latest