Planking Adversarial Attack
Collaboration with OLEG YUSUPOV
Planking Adversarial Attack
This NFT collection
shows the ‘ideas and thoughts’ of algorithms for detecting and recognizing people, and what they are capable of at the moment.
Thus, we study this area: where it all began, at what point of development we are now, and where we are moving.
Big brother isn’t omnipotent
Dynamic NFT
Interactive Demo
The dynamic works of the NFT collection are presented in an interactive format: you can change the threshold and observe the dynamics.
Threshold is a way to manage the predictions of a neural network. The lower the threshold, the more objects will be detected, but the less likely it will be identified correctly.
PREDICTION THRESHOLD
0.1.2.3.4.5.6.7.8.91
PREDICTION THRESHOLD
0.1.2.3.4.5.6.7.8.91
Models
Investment roadmap
Accompanying a phygital product and digital asset with high-quality promotions, creating collaborations and integrations.
20%
This tool will allow the use of 3DML neural networks in daily creative processes. It promotes the cooperation of humans and artificial intelligence for creating content of the future.
30%
Managing Telegram, Discord, Instagram and Twitter accounts. Conducting Phygital Art Workshop. Organizing Phygital Nights and Phygital Science events. Creating educational courses for in-depth work with AI and 3DML.
40%
The ALGORITHMIC AESTHETICS project is like an expanding fractal. We will create more technological art using the potential of NFT.
>
BACKGROUND STORY_
PROBLEM
The computer vision technology for human recognition is exciting, invisible and already being used — cameras are everywhere. It can be used for good purposes, i.e., ensuring order, investigating, finding criminals.But at the same time, the technology can become an instrument for repression and be used to persecute politically unsuitable people or to introduce rating systems. This fine line between controlling actions and protecting rights and security worries many people today.
COMPUTER VISION & NEURAL NETWORKS
Human recognition is based on computer vision technologies and neural networks. Technological growth begun in 2012, when the first datasets were collected, marked up and applied in machine learning technologies, sometimes even without permission of their owners.
These technologies are massively implemented in cameras on the streets of cities, metro, roads, train stations, airports and other elements of urban infrastructure. Various organizations understand the seriousness of the situation: Meta has officially turned off support for face recognition in social networks, ethical issues are raised at the state level, and one of the Black Mirror episodes has already been implemented in China.
ADVERSARIAL ATTACK: ANALOG & SOFTWARE
While CV being developed, new adversarial attack approaches are born. It represents attacks on computer vision systems, where the addition of one pixel to the image can break the algorithm, and the network does not recognize the object.
All approaches in the field of Adversarial Machine Learning (AML) mean a targeted impact on a neural network that may cause errors in its work. They are ranked as Adversarial attacks.
However, there are “simple” attacks on algorithms that do not require writing code.
BEGINNING: PLANKING ADVERSARIAL ATTACK
My project began in 2008 with a protest against the same boring photographs, where everyone holds the sun in their hand, puts on a smile near attractions, or stands in front of something in a standard pose. My youthful protest was expressed intentionally in a silly way — in the form of planking.
Over time, I realized that it was interesting also because there were simply no photographs of this kind in the datasets. The first algorithms could not find a person in the planking photograph — this was not typical for recognition.
A modern artist, researcher and entrepreneur. Everything that he creates is inspired by combining art, science and business spheres. Oleg promotes phygitalism as a new art direction and studies the Intelligence Amplification concept - the interaction of humans (artists) and artificial intelligence
Oleg Yusupov
Artist, CEO of PHYGITAL+ and PHYGITALISM
PHYGITAL+
we@phygitalism.com