topic Re: Using custom machine learning models in XSOAR in Cortex XSOAR Discussions

Using custom machine learning models in XSOAR

MareSeestern — Fri, 17 Jan 2025 20:21:16 GMT

Hello everyone,

I am a machine learning scientist exploring ways to integrate ML algorithms with XSOAR.

While I am familiar with XSOAR's AutoML features, my focus is on implementing custom models (e.g., TensorFlow or PyTorch models) within XSOAR workflows. I haven't found a straightforward method to directly integrate these models into a playbook. One idea I’ve considered is using an API to connect external models to XSOAR.

Has anyone worked on similar integrations or used features that align with this approach? If so, I’d greatly appreciate it if you could share papers, resources, or how-tos that might guide me.

I’m also curious to hear about your use cases for integrating custom ML models with XSOAR.

Looking forward to your insights.

Re: Using custom machine learning models in XSOAR

OKaraduman1 — Tue, 21 Jan 2025 14:26:42 GMT

Hi,

Some integrations with external AI engines are available on the XSOAR marketplace, including OpenAI, ChatGPT v3, Anything LLM and Ollama. I believe these integrations will help you with the integration an implementation.

Re: Using custom machine learning models in XSOAR

MareSeestern — Tue, 21 Jan 2025 15:19:33 GMT

Hey,

Thanks for your response! I’ll take a look at these integrations. Additionally, I’ve started experimenting with building my own integration to fetch data from an API that hosts my potential ML model. If anyone has insights or suggestions regarding this approach, I’d greatly appreciate it!