跟李沐学AI 02 深度学习介绍【动手学深度学习v2】

02 深度学习介绍【动手学深度学习v2】

https://www.bilibili.com/video/BV1J54y187f9?p=1

https://www.bilibili.com/video/BV1J54y187f9?p=2

 

论坛 https://discuss.d2l.ai/c/chinese-version/16

 

https://www.image-net.org/

ImageNet is an image database organized according to the WordNet hierarchy (currently only the nouns), in which each node of the hierarchy is depicted by hundreds and thousands of images. The project has been instrumental in advancing computer vision and deep learning research. The data is available for free to researchers for non-commercial use.

 

物体检测和分割 https://github.com/matterport/Mask_RCNN

样式迁移   https://github.com/StacyYang/MXNet-Gluon-Style-Transfer

 

人脸合成   

Progressive Growing of GANs for Improved Quality, Stability, and Variation

Final result video of our paper, published at ICLR 2018. We strongly recommend viewing at maximum quality (1080p @ 60).

https://research.nvidia.com/sites/default/files/pubs/2017-10_Progressive-Growing-of/karras2018iclr-paper.pdf

 

文字生成图片 https://openai.com/blog/dall-e/

Creating Images from Text

We’ve trained a neural network called DALL·E that creates images from text captions for a wide range of concepts expressible in natural language.

 

文字生成 https://en.wikipedia.org/wiki/GPT-3

Using text on the internet, GPT-3 is trained to generate realistic human text. GPT-3 has been used to create articles, poetry, stories, news reports and dialogue using just a small amount of input text that can be used to produce large amounts of quality copy.

 

无人驾驶

 

案例研究，广告点击

点击搜索触发， 点击率预估，  排序(点击率*竞价)

预测与训练

特征提取(广告主，产品描述，产品图片)--> 模型-->点击率预测

训练数据(过去广告展现和用户点击)-->特征用户和点击-->模型

 

完整的故事

领域专家 ：点击 <-- 展现  模型控制广告展现，这些又用来训练新的模型。

数据科学家：数据-->模型

AI专家： 提升模型精度和性能。