[sibeshk96]

Kudos for the concept. Intriguing form of counter-surveillance that could actually work in a lot of existing surveillance states. Had a few questions :

1. Who in the system is responsible for maintaining and updating the object detection model? It seems like a centralized point of failure for an otherwise decentralized system. Might want to check out existing techniques for Federated Learning for ways to counter this. (https://www.openmined.org/)

2. How possible is it to mount a Sybil attack on this system? Why did you select PoW considering it's weaknesses? One possible problem could be that early adopters will necessarily have to be co-located for the system to have any value(Hence consolidating value in a certain geographical area, making it very easy to regulate/shut-down).

[david_at]

1) Yeah, I'm not sure how Federated Learning would solve the problem. Federated learning still requires someone to decide on a single model and it's still trained under the coordination of a central server[1]. And how do you ensure the data people are giving the model is good? You need some sort of scheme to determine that. So how do you decide on the scheme? Decentralization? Isn't this decentralised-turtles all the way down?

2) I'll explain in detail how it prevents Sybil attacks later. But It involves signing all the tracklets and digest tuples (see website) as well as a sort of Russian Nesting Doll of hashes. PoW is the only viable one we have. I like to use systems that are battle tested. I don't know what you mean by PoW weaknesses. Could you clarify? You don't need to co-locate. Read what I learned about my neighbourhood with just one node under "Use Cases" on the site. As to people shutting it down, I have clearly delineated on the site's front page what my theorem predicts will happen if they do. They're handing data, unprecedented knowledge of human behaviour and AI dominance to those who don't. So yeah, shut it down. Fill your boots, mate...(https://i.imgur.com/rDJWv5E.gif?noredirect).

[1] https://towardsdatascience.com/whats-new-in-deep-learning-re...