/tech/ - Technology

Sentiment Analysis / User classification / Chatbots

1. Topic-specific sentiment i.e. Topic Modelling + Sentiment Analysis

2. User classification by what they say online

3. Neural Networks vs Support Vector Machines vs Naive Bayes

4. Chatbot architecture selection

Audio and Image Generation

1. Tacotron and other text-to-speech system (Lyrebird)

2. Image denoising and scaling (Waifu2x)

3. Image swapping and fabrication (Deepfakes)

Others

1. Facial recognition and classification by Age, Sex, Race etc.

2. Video games and cryptocurrency strategies

>>887042

Do you have any actual code?

>>887712

I am asking /tech/ if anyone here does Machine Learning.

For >>887060 though

S1~S2. https://github.com/MKLab-ITI/reveal-user-classification https://github.com/zafargilani/uc https://github.com/dsquareindia/twitter-user-classification

S3 is a whole can of worms

S4 https://github.com/gunthercox/ChatterBot is the most common toolkit for the job

A1 https://github.com/Kyubyong/deepvoice3 https://github.com/Rayhane-mamah/Tacotron-2 https://github.com/riverphoenix/tacotron2

A2 IT'S WAIFU2X, WHAT DO YOU EXPECT

A3 Straight outta Reddit https://www.reddit.com/r/GifFakes/ https://www.reddit.com/r/SFWdeepfakes/

F1 https://github.com/JostineHo/mememoji https://github.com/dhvanikotak/Emotion-Detection-in-Videos https://github.com/AsoStrife/Leveraging-Cognitive-Computation-Text-Extraction-Tools-to-Detect-Emotions-from-Facial-Expressions (Replace emotions with demographics)

Also https://github.com/oarriaga/face_classification https://github.com/stdcoutzyx/DeepID_FaceClassify https://github.com/dhvanikotak/Emotion-Detection-in-Videos

bampo

>>890017

Then could you please revive >>>/machinecult/ ? And allow us to dump ebooks for it to read?

>>890018

I'll have to fish it out of my old HDD first and then post it here. That is if I can even find it. ebooks would probably make a good resource for the bot to use. I was thinking of just grabbing some books in text format from the Gutenberg Library

>>890020

Since we ARE building Tay, >>>/pdfs/ or http://inclibuql666c5c4.onion.link/

>>890020

>>890017

Keep up the good work

>>890007

gigo

>>890007

What was with all the circle faces?

>>890705

>Not knowing Twatter

Too many newfags

>>887042

>Build our future^W^W^W

>fuck up our future with shitty bugged markoff chain automatic medical instruments, vehicles, and law

FTFY

>>891531

In that case things will get horrible.

Things we want:

Machine Learning shitpost bot

Algorithmic trading (fuck /biz/)

Moar fake porn from Deepfakes

Bampo