Artificial neural network learned to guess passwords from smartphones
A team of scientists from technological University in Singapore have developed an artificial neural network which is able to guess the pin of a smartphone with an accuracy of 84%.
The results of the experiment published in the journal of Cryptology. The only information that the program needs to find the right password – this is the data from the accelerometer, light sensor and other sensors, which exist today in most devices.
Scientists have developed a mobile app that collects all the necessary information for the neural network. User permissions it does not need. The basic idea is that the position of all keys on the keyboard is known, and therefore, by tilting the smartphone, the change of light and other data you can calculate what a key has been pressed.
Neural network, on the basis of which the program works, is necessary in order to treat the entire volume of collected sensor data and calculate the desired combination of digits. During testing it became clear that on all possible combinations of four digits recognition accuracy in 20 attempts amounted to 83.7%. Accuracy in detection among 50 of the most common pin – 99,5% on the first attempt.
© 2017 – 2019, paradox. All rights reserved.
