Editorial

Thoughts From Your Humble Curators

This week we cover Google Dataset Search Engine and other topics.

As always, if you find this newsletter useful, feel free to share it with your friends/colleagues.


This newsletter is published by Waikit Lau and Arthur Chan. We also run Facebook's most active A.I. group with 172,000+ members and host an occasional "office hour" on YouTube. To help defray our publishing costs, you may donate via link. Or you can donate by sending Eth to this address: 0xEB44F762c58Da2200957b5cc2C04473F609eAA65.

Join our community for real-time discussions here - Expertify

Artificial Intelligence and Deep Learning Weekly

News


Deals

Artificial Intelligence and Deep Learning Weekly

Blog Posts


Open Source

Member's Question

How come my post get so few clicks?

Lately we have been asked this question often:

"I have this super-duper posts with super-fine details on why deep neural network is going to revolutionize humanity. My post talks about loss function and back-propagation. There is NO MATH! So I think it is very good for the community. And I think it is the first time we see this kind of blog post on AIDL!"

"But How come I only get 4 likes? My post should go VIIIIRAAALLL!"

We rephrased of course. But we think viewership has a lot of to do with qualities of our content. So the question deserves a good answer.

The first answer perhaps is just "Well your post is not that interesting....." But why is it the case? And in particular why is it the case at AIDL?

You should understand, by now, knowledge about deep learning is not entirely new. And AIDL has been there for almost 3 years. So most beginner topics you try to cover, has been covered more than once. For example, (again I rephrase),

  • "Write your own neural network in less than 5 characters"
  • "The detailed, simple, easy, beginner's, expert's, mathematical-but-tons-of-mistakes, mathematical-but-the-notation-was-inconsistent, mathematical-and-correct-but-boring guide of back propagation"
  • "The very detailed guide of convolutional networks with a wrong notion of convolutions"
  • "How do you start machine/deep learning and AI if you don't only know how to add and subtract."
  • "Why AI is/isn't an imminent danger for humanity (Disclaimer: I have no other statistics in other human calamities to compare. But I feel like writing one.)"

So my take is you should ask yourself, "am I really writing something new?" Now that's a tough requirement. Writing something new means you have already read up a body of literature yourself. And after studying the nuances of important literature, you then write something which people never read before. Yet these new reading also has to be convincing and interesting.....

In a nutshell, technical blogging is not easy. Some statistics for you: out of 50 blog posts submitted to AIDL. we might accept approve around 20-30% of them. And for the published ones, a majority of them will not gather more than 10 likes. That has nothing to do with whether you hashtag or come up with a catchy title. It has to do with whether you are creating genuinely interesting content for AIDL members.

On the other side of the coin, when your content is valuable, you will get attention. Just check out Joey Adrian Rosebrock's posts on computer vision? Or Raymond de Lacaze's regular posts on different papers? Rosebrock's posts usually include non-trivial implementation with working python code. And de Lacaze's posts are explanations of recent papers. Both of them post useful links which we check out from time to time. That's why their posts got clicks, not because they use any type of gimmicks.

One last thing: don't give up - writing about machine learning, just like learning machine learning, takes a long time to master. Also as Adrian later commented: writing a post is really not about getting likes. Writing by itself, it's a learning process. Once you master the knowledge, other things will naturally come.

Artificial Intelligence and Deep Learning Weekly

About Us

This newsletter is published by Waikit Lau and Arthur Chan. We also run Facebook's most active A.I. group with 172,000+ members and host an occasional "office hour" on YouTube. To help defray our publishing costs, you may donate via link. Or you can donate by sending Eth to this address: 0xEB44F762c58Da2200957b5cc2C04473F609eAA65.

Join our community for real-time discussions here: Expertify

Artificial Intelligence and Deep Learning Weekly