Vicone Corporation

Chih-Sung Chen

Shih-Han Hsu

Yi-Li Cheng

Yao-Tang Chang

Yu-Yu Liao

Pei-Zhong Luo

Yi-Hong Chu

Hao-Min Lin

Jui-Chang Hsu

Weaknesses are detected in binary codes. Vulnerable code snippets are completed into corresponding compilable codes that are in source code form using a generative artificial intelligence (AI) model. The compilable codes are compiled into corresponding binary codes. The binary codes are decompiled into decompiled codes that are in source code form. The decompiled codes are used as training data to train another generative AI model to detect one or more weaknesses that may be present in target decompiled codes. A target binary code is detected to have a weakness when the other generative AI model identifies the weakness in a corresponding target decompiled code of the target binary code.

Detection of weaknesses in binary codes

A large language model is trained by inputting a TARA configuration of an ECU of a connected vehicle, an EEA specification, and IDPS rules of the ECU. Software modules are used to extract feature vectors for input to the LLM. The LLM is fine tuned by further inputting later versions of the ECU and data from other ECUs of the vehicle. IDPS rules for another ECU of the vehicle which has no rules or few rules are generated by inputting the TARA configuration of the other ECU into the trained LLM and prompting the LLM to generate rules. The prompt may include instructions to minimize overhead on the other ECU, reduce rules within the vehicle, etc. The generated rules are output from the LLM and stored within the IDPS engine of the other ECU, stored on a computer, transmitted to another vehicle. Rules for two other ECUs may be generated.

Generation of TARA-based IDPS rules utilizing generative artificial intelligence

Growing community of inventors

Taipei, Taiwan

Chih-Sung Chen