Thinking about party attack-like growth tactics using deep learning

Summary

The other day

After talking with

The reply is.

????

Isn't writing an article on Qiita an output? ??

Lobby activities like SIer and see articles with human wave tactics! Does that mean you should do it? ?? Instead, it can be quantified, so even if it becomes a score, it's simply good! If you don't want to get it, I want you to crack it with Masakari. Even if it becomes impossible to recover. ** Otherwise you can't pivot, so you can't grow. ** That's not good.

Well, there is "the opposite of love is not hatred, but indifference [^ 1]". In other words, Masakari may not be the opposite of love (≒ like!). Flame Looking at the above, there are some problems with the content, but it seems that there are many cases where the comment section is more than that.

Masakari is fine, I want a voice. But I don't have the power to borrow. There are no people around me who are interested in the program. Better yet, you can use the power of a computer instead of human power! If you can make a title that you can read the article by machine learning somehow, maybe you can see it and write something. ** **

It smelled like a black hat. Overwhelming feeling of donmai. Lol

and

Automatic Deep Learning text generation (+ Colaboratory Tips and Snippets) that even cats who cannot write code can do

I will use this. It's a 3-layer LSTM. I thought I'd use BERT, but I quickly searched for a Colaboratory that could work. (It seems better not to cut corners there) I should use Malcolm chain because it is a play level, but ** "with the help of a computer" ** I chose tpu + tensorflow because I want to feel better. Lol

Colaboratory doesn't work as it is, so make plotly.plotly chart-studio.plotly, ! pip install tensorflow == Please run 1.13.1 around the beginning. (Runtime restart required)

I extracted the learning material from the following

[2019 review] Qiita likes ranking TOP100 Qiita Various Ranking 2018

Have TPU go around. It was like this. 10 excerpts

(2019 revised edition)
Roass's slimy moving portfolio
VSCode)
[translation]You are in programming
"Machine learning
[April 2019 version] JavaScript commentary
Since I made this, the source and explanation
Deep learning
Ali of the engineer in the world
Get to know Git in an hour

It's not a Zenzen sentence, but in summary, AI says, the titles you can see are

Threatening system Curation system For beginners
Don't use xx xx selection best practice
No one tells me yyyy year Awareness
Why we As of yyyy year mm month The engineer
Still 【xxxx(Affirmative sentence)] * Even if you are inexperienced, you will not be frustrated

That's it! It doesn't generate sentences completely, so it makes sense to use AI! It seems that it will be, but I'm sorry that AI chose the feature amount.

So what happens when you write an article for SIer with a title that incorporates this element? ??

Expect next time! !!

Next time ↓

[Best for beginners] Why do we rely on RPA? Automation best practices that no one can tell us yet (as of March 2020)

Too bad ...

[^ 1]: [It seems that it has nothing to do with Mother Teresa. "Is the" genius "the 1981 AC Japan advertisement?" Navel. Millennials don't know. ]]]

Recommended Posts

Thinking about party attack-like growth tactics using deep learning
About Deep Learning (DNN) Project Management
I tried deep learning using Theano
Under investigation about PYNQ-Let's do deep learning with FPGA using Python-
Image recognition model using deep learning in 2016
Stock price forecast using deep learning (TensorFlow)
About data expansion processing for deep learning
[Deep Learning from scratch] About hyperparameter optimization
Deep Learning
A story about simple machine learning using TensorFlow
An amateur tried Deep Learning using Caffe (Introduction)
An amateur tried Deep Learning using Caffe (Practice)
Stock Price Forecast Using Deep Learning (TensorFlow) -Part 2-
[Causal search / causal reasoning] Execute causal search (SAM) using deep learning
An amateur tried Deep Learning using Caffe (Overview)
Stock price forecast using deep learning [Data acquisition]