Robert Muellers Report Of The Investigation Into Russian Interference In The 2016 Presidential Elections is out. It is heavily redacted so lets if we cant apply natural language Processing (NLP) to fill in the blanks. Lets unredacted it!
If there is code for this video, please refer to YouTube notes
Robert Muellers Report Of The Investigation Into Russian Interference In The 2016 Presidential Elections is out. It is heavily redacted so lets if we cant apply natural language Processing (NLP) to fill in the blanks. Lets unredacted it! MORE: Blog or code: http://www.viralml.com/video-content.html?fm=yt&v=YU3EoS3kj10 Signup for my newsletter and more: http://www.viralml.com Connect on Twitter: https://twitter.com/amunategui The NLP Tool: http://www.viralml.com/mueller Signup for my newsletter and more: http://www.viralml.com Connect on Twitter: https://twitter.com/amunategui Check out my books on Amazon: https://amzn.to/2DpLcW0 Code: https://www.tensorflow.org/tutorials/sequences/text_generation Transcript Hello YouTube Friends, A short video about a complex event. The Report Of The Investigation Into Russian Interference In The 2016 Presidential Elections. And more specifically into what extent people will use todays modeling tools to see what is hiding underneath the redacted docs - but let look at the modeling first before we get to this. Welcome to the ViralML Show, my name is Manuel Amunategui. I am the author of a few books including Monetizing Machine Learning where we port Python ML projects to the web for the world to enjoy and look at paywalls to monetize your work. You can find this book and my other ones on Amazon, simply enter my first and last name in the Amazon search bar. Please signup for my newsletter to get early access to my material. Subscribe to the channel - and give it a big thumbs up, pretty please! - This video will be filed under State of the Industry. I assume most of you have about the same security clearance as me -so youre stuck with this hot potato of a report that cant be fully read - its politically and emotionally charged... its an object of fascination and will undoubtedly be an analyzed to death by people all around the world. theres a lot at stake for a lot of people. Feels like were in a weird movie, smack in the middle and we have no idea where its going. About the redaction - I know we all thought the same thing - whats underneath all that? If they werent there, I probably wouldnt be curious about it - but now that they are... Where she has to check the math on the heavily redacted document and simply holds it to the light - yep I tried that with this one by printing it out but no dice - were going to have to use some other trick So, I did what any data scientist would do an that is break out a model - I went with the hot off the press, and barely legal... Tensorflow 2.0 well, it is legal but its in alpha and ran an RNN with eager execution - I got the data on Kaggle - get it there both original pdf and the pdf-to-CSV conversion that Paul did. I modeled it with a GPU, and found that the model stopped converging around 30 epochs - so - 30 it is.... page 130 even the footnotes can be redacted... page 23 By February 2016, internal IRA documents referred to support for the Trump Campaign and opposition to candidate Clinton. For example, page 26 The IRAs Twitter operations involved two strategies. First, IRA specialists operated certain Twitter accounts to create individual U.S. personas, page 120 During Manaforts prosecution and while the jury was deliberating, the President repeatedly stated that Manafort was being treated unfairly and made it known that Manafort could receive a pardon. Now, am certainly not making any fun of this whole situation. The document represents a lot of work, a lot of political energy and emotions condensed around this form all sides. Its about the 2016 election and about Donal Trump who is the current President of the US. Were like in the middle of a weird, weird movie and we have no idea how it ends. but back to the Recurrent Neural Network Its basically a model that predicts on the character after another, you heard it right - it predicts on the character after the other, now that is unbiased. lets touch that point at the end of the video - such as is redacting and releasing things around it giving us enough of the code to infer the res CATEGORY:DataScience