 Hello, everyone. This talk is about revisiting related key boomerang attacks on AES using computer-aided tool. This work is joined with Patrick, Marie, Pierre Alon, and Wa. AES is one of the most widely used encryption algorithms in the world. Since it was first introduced, many researchers have studied its design, looking to probe the algorithm's In recent years, boomerang attack was found to be powerful against many block ciphers, especially full-round AES. A few years later, several milk models were introduced to search automatically for boomerang distinguishers and boomerang attacks on block ciphers. But they can only be used when the key schedule is linear. In this work, a new model is introduced to deal with non-linear key schedules, as is the case for AES. This model is more complex and too slow for an exhaustive search. However, when some hints are added to the solver, it found a new boomerang attack against full AES 192. Its complexity is 2 to the power of 52 times lower than the one presented by Biryukov and Kovradovich. And we also recover the attack on AES 256 of Biryukov. For AES 192, we were able to find a better attack. It is like the original version, we use the same lower trail, but for the upper trail, the same but the position of the active bite is different. Our distinguisher has the probability of 2 to the power of minus 108 instead of 2 to the power of minus 110, it is a bit better, but almost has the same probability. We expand to mount a boomerang key recovery attack. In this case, it is much, much better. Here is the zoom on the input of both attacks. As you can see, in our attack, we know much more information on plaintext than in Biryukov's attack. In general, we found the current best related key boomerang attack on AES 192 with 2 to the power of 124 time, 2 to the power of 124 data, and 2 to the power of 79.8 memory complexities, which is better than the one presented by Biryukov and Kovradovich at Asia Crypt 2009 with complexities 2 to the power of 176, 2 to the power of 123, 2 to the power of 152, respectively. This represents a huge improvement in the time and memory complexity, illustrating the power of automatic tools in cryptanalysis. If you are interested, I hope you watch my talk on Wednesday morning, and I am really looking forward to seeing all of you in Asia Crypt next month.