Veracode, Inc.

Anna Bacher

Erich Gstrein

Roman Rudenko

A generative artificial intelligence (AI) driven code fixing pipeline has been created that uses a transformer-based large language model (LLM) to patch flawed program code. A pre-trained LLM is fine-tuned to generate a response that is a modified version of a code fragment in a prompt to the pre-trained model. After fine-tuning, the pre-trained LLM (hereinafter 'code fix model') is integrated into a pipeline that includes a program code cybersecurity scanner and a prompt generator. The scanner generates indications of flaws in program code and weakness types for those flaws. These indications flow into the prompt generator. The prompt generator retrieves reference code pairs based on weakness type and programming language to generate a batch of prompts to run inference on with the code fix model. The responses generated by the code fix model are presented as patching alternatives.

Generative artificial intelligence driven software fixing

A method and apparatus utilize a peer-to-peer network of security nodes collectively adhering to a protocol for inter-node communication. The system is comprised a plurality of first security nodes, at least one second security node, and at least one third security node. The plurality of first security nodes receive at least one of pre-trained detection models and rules, monitor at least one of a blockchain and connected devices for malicious behavior based on the received at least one of pre-trained detection models and rules, and report the malicious behavior. The at least one second security node creates and communicates the at least one of pre-trained detection models and rules to the plurality of first security nodes. The at least one third security node is informed by the at least one second security node of the reported malicious behavior.

Growing community of inventors

Vienna, Austria

Anna Bacher