- Those who are interested in code and programming can add my official account < Old K Play Code > to communicate with me!
“Learning Python, half of the friends are for deep learning.
Self-study is good, but someone still needs guidance.
Are there any online communities or platform tools that friends who study deep learning will gather?
In fact, there are. If you want to know, just click on the old K play code above and follow me~!
- Old K checked the recommendations of the netizens on Zhihu on this issue, organized them into articles, and recommended them to you!
- Start the text below~!
1. Quora recommendation: 4.5/5.0 stars ★★★★☆
https://www.quora.com/
The following are some representative comments on "Quora": Excerpted from the answers of netizens under the knowledge of related questions:
“
- Reddit and Quora are more inclined to information interaction, there is less discussion about solving specific technical problems, and they are more methodological. The last three tend to solve specific problems, many of which are very specific. Quora and stackexchange are full of bigwigs.
- The machine learning forum with Chinese as the main body does not seem to be particularly famous. According to my observation, the relatively high level at this stage is actually...Zhihu (machine learning, deep learning), which is very similar to a combination of Reddit, Quora and StackOverlow. There are also some good articles on CSDN, but It seems that the update frequency is not high, and the article is too old.
- As for the Q&A community, data science on stackOverflow should not become the main body on behalf of others. Exchange activity has always been relatively average. As for Quora, the traffic difference is still quite large compared to Zhihu.
2. Reddit recommendation: 4.5/5.0 stars ★★★★☆
https://www.reddit.com/
The following are some representative comments on "Reddit": Excerpted from the answers of netizens under the knowledge of related questions:
“Reddit and Quora are more inclined to information interaction, there is less discussion about solving specific technical problems, and they are more methodological. The last three tend to solve specific problems, many of which are very specific. Quora and stackexchange are full of bigwigs.
- The machine learning forum with Chinese as the main body does not seem to be particularly famous. According to my observation, the relatively high level at this stage is actually...Zhihu (machine learning, deep learning), which is very similar to a combination of Reddit, Quora and StackOverlow. There are also some good articles on CSDN, but It seems that the update frequency is not high, and the article is too old.
- If it’s a forum format for posting more specialized and detailed questions, and getting quick answers, in fact Tieba is more suitable. After the two years of Baidu Tieba’s sale, it seems to be quite cold now, Reddit may be even worse. As you usually imagine.
3. BigML recommendation: 4.5/5.0 stars ★★★★☆
https://bigml.com/
The following are some representative comments on "BigML": excerpted from the answers of netizens under the knowledge of related questions:
“The BigML platform is designed to help customers create, quickly experiment, fully automate and manage machine learning workflows. BigML provides a good visualization function, and has algorithms for solving classification, regression, clustering, and anomaly detection. The product can be subscribed on a monthly, quarterly, and annual basis, and a free version is also available (the data set of the free version is limited to 16 MB).
- BigML's platform, private deployment and rich tool set will continue to help our customers create, quickly experiment, fully automate and manage machine learning workflows to provide the best smart applications in their class. Based on the graphical user interface, BigML provides the following functions:
4. DataRobot recommendation: 4.5/5.0 stars ★★★★☆
https://www.datarobot.com/ The
following are some representative comments on "DataRobot": Excerpted from the answers of netizens under the knowledge of related questions:
“DataRobot (DR) is a highly automated machine learning platform created by the world's best Kagglers, including Jeremy Achin, Thoman DeGodoy and Owen Zhang. On the DR official website, we can see the sentence, "Data science requires mathematics, statistics, programming skills and business cognition. With DataRobot, you can easily obtain data and business insights, and we are responsible for the rest."
5. fast.ai recommendation: 4.0/5.0 stars ★★★★
https://www.fast.ai/
Here are a few "fast.ai" more representative comments: Taken at netizen know almost answered under the relevant questions:
“http://forums.fast.ai fast.ai forum, there are still many people discussing courses or technical issues outside of the courses every day.
6. Recommended by mathexchange : 4.0/5.0 stars ★★★★
https://math.stackexchange.com/ The following are some representative comments on "mathexchange": excerpted from the answers of netizens under the knowledge of related questions:
“前面的回答基本上都涵盖了比较好的去处了吧。就个人而言，英文主要是reddit的ml板块，推特上关注的一些行业大牛，或者hashtag（比如nlp就是#NLProc），quora，stackoverflow和mathexchange。中文的话知乎和微博也可，关键在于怎么构建自己的timeline
7. Driverless AI 推荐：4.0/5.0星★★★★
www.h2o.ai/driverless-ai/
以下是关于“Driverless AI”的几个比较有代表性的评论：摘自网友在知乎相关问题下的回答:
“Driverless AI是由开源机器学习平台h2o.ai推出的最新产品，这一工具可以帮助非技术人员准备数据，审校参数，确定最优算法，进而用机器学习来解决专门的企业问题。该产品旨在降低数据科学在企业环境下运行的门槛。
- 这些是Driverless AI的一些令人振奋的功能：
8. hashtag 推荐：4.0/5.0星★★★★
以下是关于“hashtag”的几个比较有代表性的评论：摘自网友在知乎相关问题下的回答:
“前面的回答基本上都涵盖了比较好的去处了吧。就个人而言，英文主要是reddit的ml板块，推特上关注的一些行业大牛，或者hashtag（比如nlp就是#NLProc），quora，stackoverflow和mathexchange。中文的话知乎和微博也可，关键在于怎么构建自己的timeline
9. stack exchange 推荐：4.0/5.0星★★★★
https://stackexchange.com/
以下是关于“stack exchange”的几个比较有代表性的评论：摘自网友在知乎相关问题下的回答:
“stack exchange统计版（Cross Validated ），更多是统计相关
10. paperweekly 推荐：4.0/5.0星★★★★
https://www.paperweekly.site/
以下是关于“paperweekly”的几个比较有代表性的评论：摘自网友在知乎相关问题下的回答:
“https://www.paperweekly.site/ paperweekly，推荐论文的专业论坛，上面有许多用户推荐的论文。
- 现在一般的微信群和qq群都讨论对理解深入貌似没多大作用，感觉更多的是传递名词功能，个人感觉真正的技术方面的可以看看某些博士们写的博客，大部分不会存在误导，写的内容也值得推敲，这个要自己根据方向去找。有个paperweekly上面会有很多好的论文推荐，部分论文有代码供复现，你可以去了解一下。
11. mxnet 推荐：4.0/5.0星★★★★
https://discuss.gluon.ai/
以下是关于“mxnet”的几个比较有代表性的评论：摘自网友在知乎相关问题下的回答:
“mxnet深度学习框架，有配套的视频，论坛，学习文档，简直应有尽有。论坛里面有大佬们组织的kaggle经典训练集的比赛，很多牛X的小伙伴论坛里讨论方法，每次都有人拿下相关竞赛前几名。
12. medium 推荐：4.0/5.0星★★★★
https://medium.com/
以下是关于“medium”的几个比较有代表性的评论：摘自网友在知乎相关问题下的回答:
“medium：很多人代码在github，教程在medium。
13. github 推荐：4.0/5.0星★★★★
https://github.com/
以下是关于“github”的几个比较有代表性的评论：摘自网友在知乎相关问题下的回答:
“当然还有github，github是讨论技术问题的最好地方，但问题在于假如你关注的一个技术问题并没有任何的repo，或者star的人很少，自然就没什么人讨论。
- medium：很多人代码在github，教程在medium。
14. IBM Watson Studio 推荐：4.0/5.0星★★★★
https://www.ibm.com/cloud/watson-studio
以下是关于“IBM Watson Studio”的几个比较有代表性的评论：摘自网友在知乎相关问题下的回答:
“如何从这个清单中排除IBM？这个世界上最知名的IT品牌之一。IBM Watson Studio为构建和部署机器学习和深度学习模型提供了一个出色的平台。借助Watson Studio，你可以轻松完成数据准备工作、使用RStudio等熟悉的开源工具、访问最流行的库、训练深度神经网络等。对于机器学习的入门者来说，IBM提供了一系列的教程视频帮助你入门Watson Studio。
15. stackoverflow 推荐：4.0/5.0星★★★★
https://stackoverflow.com/
以下是关于“stackoverflow”的几个比较有代表性的评论：摘自网友在知乎相关问题下的回答:
“Personally, English is mainly the ml section of reddit, some industry leaders followed on Twitter, or hashtags (such as nlp is #NLProc), quora, stackoverflow and mathexchange. Zhihu and Weibo are also available in Chinese, the key is how to build your own timeline
- For more interesting code information, follow the official account < Old K Play Code >, or add my WeChat kevinchaos to play with me
- If you feel this post helpful to you, please click easily endorse or look oh!
