AI Studio recently found on a lot of high-quality open-source project depth learning, deep learning from entry to advanced, covering CV, NLP, generated against the network, reinforcement learning more research, as well as the latest motion picture, are the NoteBook out of direct revenue, and AI Studio also offers a free operator force, direct online operation to run training mission, recommended depth learning developers and learners collections and research.
Operators force get the link: https://aistudio.baidu.com/aistudio/questionnaire?activityid=667
A, deep learning Getting Started Collection
This collection is divided into four sections, from zero introduces the basics of the depth of learning, including machine learning, artificial intelligence, such as the relationship between the depth of the three learning; entry required mathematical knowledge, including mathematics, linear algebra, probability theory, information theory refining and other essential pre-knowledge; Python Quick Start Guide and Quick Start PaddlePaddle, teach you learn Python, and give you step by step with the code shows the basic usage PaddlePaddle, house prices predicted and digital handwriting recognition demo, so that complete zero-based students can start this entry through knowledge-based collection to learn, learn - into the back of the white necessary, immediately learn ~
Ø Getting Started Lesson - What is the depth of learning? https://aistudio.baidu.com/aistudio/projectdetail/150131
Ø Getting Started Lesson - necessary mathematical knowledge https://aistudio.baidu.com/aistudio/projectdetail/150017
Ø Getting Started Lesson --Python Quick Start Guide https://aistudio.baidu.com/aistudio/projectdetail/149800
Ø Getting Started Lesson --PaddlePaddle Quick Start https://aistudio.baidu.com/aistudio/projectdetail/150026
B, Computer Vision Project Collection
AI Studio in October featured on the new project, add a lot more fit the actual project applications, including flow density estimation, aircraft identification, etc., in addition to various models reproduce the foreword algorithm, including: PyramidBox, MobileNet, MobileNet-v2, DisResNet, SE_ResNeXt and so on, I seriously study the implementation of face recognition algorithm which PyramidBox classic classic with PaddlePaddle, recognize the effect is really very good, indeed the three materials championship algorithm, strong recommend it to everyone ~
Ø achieve classic classic with PaddlePaddle face recognition algorithm PyramidBox https://aistudio.baidu.com/aistudio/projectDetail/169468
In addition, there are a series of projects CV writing direction is also very good, covering the latest mainstream network, combined with the actual problem to explain, I released directly linked - we can experience directly fork click Run ~
Ø crowd density achieved with PaddlePaddle estimate https://aistudio.baidu.com/aistudio/projectDetail/169478
Ø implement aircraft identification PaddlePaddle https://aistudio.baidu.com/aistudio/projectDetail/169475
Ø achieve -DPN image classification algorithm PaddlePaddle https://aistudio.baidu.com/aistudio/projectDetail/169389
Ø 用PaddlePaddle实现图像分类-MobileNet https://aistudio.baidu.com/aistudio/projectDetail/169427
Ø 用PaddlePaddle实现图像分类-MobileNet-v2 https://aistudio.baidu.com/aistudio/projectDetail/169398
Ø 用PaddlePaddle实现图像分类-DistResNet https://aistudio.baidu.com/aistudio/projectDetail/169503
Ø 用PaddlePaddle实现图像分类-SE_ResNeXt https://aistudio.baidu.com/aistudio/projectDetail/169410
C、 自然语言处理项目合集
在NLP精选项目中，AI Studio不仅推出了基于PaddlePaddle的对话通用理解模块DGU来实现对话系统，使用ERNIE+Bi-LSTM实现对话情绪识别，还有各个实际应用中的场景，譬如在在客服对话中，识别客户情绪的好坏；根据用户的搜索语句，为用户推荐相似问题；从快递单中抽取关键信息等等~都是NLP领域的上佳学习项目哦。我想主推下以下结合实际场景的四大任务，这四个项目在AI Studio的NLP专区里也有，听说都是NLP大佬们亲自写的~
Ø 分类任务：如何在客服对话中，识别客户情绪的好坏 https://aistudio.baidu.com/aistudio/projectdetail/121630
Ø 匹配任务：如何根据用户的搜索语句，为用户推荐相似问题 https://aistudio.baidu.com/aistudio/projectdetail/125034
Ø 序列标注任务，如何从快递单中抽取关键信息 https://aistudio.baidu.com/aistudio/projectDetail/131360
Ø 生成任务：语言的中外翻译，你应该知道的NLP生成任务 https://aistudio.baidu.com/aistudio/projectdetail/120044
另外这两个项目稍微要复杂一些，质量也非常不错~再来两个小手一点，马上学它~
Ø 基于PaddlePaddle的对话通用理解模块DGU https://aistudio.baidu.com/aistudio/projectDetail/169436
Ø 基于PaddlePaddle的对话情绪识别ERNIE+BI-LSTM https://aistudio.baidu.com/aistudio/projectDetail/169473
D、 生成对抗网络项目合集
生成对抗网络一直是研究的热门领域，我惊喜的发现 AI Studio近期给大家推出了一些生成对抗网络的精选项目，马上收藏学习~
有我非常喜欢的InfoGAN的算法实现，这是一种基于互信息优化实现无监督训练的GAN，能够在无标签的数据中寻找发现数据见的标签类别；
Ø 基于PaddlePaddle的InfoGAN算法实现 https://aistudio.baidu.com/aistudio/projectDetail/169476
还有Conditional GAN 、StarGAN、AttGAN、STGAN等模型复现。此类项目属于高阶任务类型的项目~适合有一定深度学习基础的小伙伴学习，当然小白也可以点击fork了解一下哦~
Ø 基于PaddlePaddle的StarGAN，AttGAN，STGAN算法 https://aistudio.baidu.com/aistudio/projectDetail/169439
Ø 基于PaddlePaddle的强化学习算法Conditional GAN https://aistudio.baidu.com/aistudio/projectDetail/169443
E、 强化学习项目合集
让我更加吃惊的是，AI Studio还推出了基于主流的强化学习算法DQN、DoubleDQN、DuelingDQN的模型复现项目，我给大家整合了一下~对强化学习感兴趣的同学可以fork并学习哦~
Ø 基于PaddlePaddle的强化学习算法DQN https://aistudio.baidu.com/aistudio/projectDetail/169455
Ø 基于PaddlePaddle的强化学习算法DoubleDQN https://aistudio.baidu.com/aistudio/projectDetail/169438
Ø 基于PaddlePaddle的强化学习算法DuelingDQN https://aistudio.baidu.com/aistudio/projectDetail/169453
E、动态图项目精选合集
在以往的精选项目中，AiStudio主要以静态图为主，近期特地推出了一批动态图的实现案例，让大家体验到各种算法的动态图实现的优势。在九月精选推荐发出后，我也听到了很多用户的呼声，想看看PaddlePaddle动态图到底效果怎么样~我可以告诉大家，非常好用！！！
我学习了以下几个动态图的项目，相比于静态图版本，动态图的训练操作更加便捷，可立即获得执行结果，无需等待计算图的全部构建完成。主推在公开的蔬菜数据集上进行了模型训练以及验证的DPN算法 ~
Ø 用PaddlePaddle实现图像分类DPN算法（动态图版） https://aistudio.baidu.com/aistudio/projectDetail/169391
当然还有MobileNet、MobileNet-v2、DisResNet、SE_ResNeXt等各种前沿算法，让大家一次体验个够！那么如果你还不知道PaddlePaddle动态图怎么用，学它、学它、学它吧
Ø 用PaddlePaddle实现图像分类-MobileNet（动态图版） https://aistudio.baidu.com/aistudio/projectDetail/169429
Ø 用PaddlePaddle实现图像分类-MobileNet-v2（动态图版） https://aistudio.baidu.com/aistudio/projectDetail/169403
Ø 用PaddlePaddle实现图像分类-DisResNet（动态图版） https://aistudio.baidu.com/aistudio/projectDetail/169406
Ø 用PaddlePaddle实现图像分类-SE_ResNeXt（动态图版） https://aistudio.baidu.com/aistudio/projectDetail/169412
以上只是给大家列出了部分10月优质项目，还有更多项目大家可以在下面AI Studio官方合集链接中找到，现在正好有一个活动：只要您fork其中不同合集的三个项目并运行（ABCDEF六个合集中任选三个合集分别fork一个项目即可），然后在楼层内回复，均有机会获得百度定制数据线、移动电源和Tesla V100 100小时算力卡哦~~
Ps：现在AI Studio上的项目支持评论了～这个升级点不错，也可以点击头像或昵称去查看他人的主页查看更多学习内容～
活动戳这里：https://ai.baidu.com/forum/topic/show/956768
活动时间：11.12~12.30（分两次抽奖~~每个用户限参与一次）
希望大家一起参与，也希望大家多多贡献高质量的项目，期待11月的精选月刊里能看到你的项目~