Trust, but verify
At the age of 19, a Stanford freshman Elizabeth Holmes filed for her first patent application — Medical device for analyte monitoring and drug delivery. If you don’t know anything about the Theranos blood test machine or haven’t read the book Bad Blood yet, you would think she must be Sheldon in medical science.
But hell no. She wrote the patented ideas, more precisely patented day-dream, like writing a term paper for her Sci-fi literature class. Such medical device did NOT exist in the past, does NOT exist now or will NOT exist in near future. But she was granted the patent anyway in the US, Canada, Europe Patent Office, Israel, Japan, South Korea, etc.
This is a well-known secret in US patent application: the United States Patent and Trademark Office has no resource to verify the claims in every patent application. If Uncle Sam is unable to verify it, do you think the patent office in the evil regime can verify the infamous invention — V2ray traffic identification method based on long short-term memory neural network?
"Trust, but verify" is a Russian proverb. It was quoted by President Ronald Reagan many times when dealing with Soviet Unions in the 1980s. In a country where its AI technology is employed heavily to monitor citizens’ every move, I’d prefer to verify the claim by myself. If this turns out to be valid, I can always find a way to defeat it.
In the end, Freedom is not free.
Step 0
If Programs = Data Structure + Algorithm, then Machine Learning = Big Data. Without a proper data set, there is no way to train our neural network. That is one of the barriers to most of data engineers. (Please stop disgracing the word scientist. 99.99% of so-called data scientists are all just engineers who use/modify the existing tools and algorithm to solve the real world problem.)
I don’t have state-sponsored resources to collect network traffic at the country level. But I’m going to collect network traffic from my home router over a long period. I will occasionally use V2ray to mimic the traffic that wants to bypass the Great Fire Wall.
I installed pfSense in my retired beefy PC and plan to use it as a router. It has a dual ports Gigabit Ethernet PCI-E card, a 500GB SSD, i7-2700k CPU. It is powerful enough to live capture packets and perform TCP packet reassembly in real-time.
In terms of hardware, I got it ready. I will post more about my on-going data collection work in this research project.