# How AI Learns to Fix IT Problems by Asking for Feedback > This patent describes an AI system that continuously learns to identify and prevent IT issues by building a map of cause-and-effect relationships, getting human feedback, and automatically updating its understanding. - **Patent:** US 12505360 - **Original title:** Continuous knowledge graph generation using causal event graph feedback - **Owner:** Bmc Helix - **Granted:** 2025 - **Status:** Active - **Times cited:** 0 - **Field:** software, ai_ml, telecommunications, consumer_electronics ## What it does The system creates a "causal graph," which is like a map showing how different events in an IT system are connected. It then asks for feedback on this map, for example, by displaying the causal graph and asking for a simple "yes" or "no" (Claim 5) if the connections are correct. This feedback, along with information about when and where events happened ("spatiotemporal context"), is fed into a machine learning model (Claim 1). The model uses this to build a "knowledge graph," a deeper understanding of the IT system. The process repeats: the system generates a new causal graph, gets more feedback, and updates the knowledge graph, even at different levels of detail (Claim 1), ultimately allowing an "Information Technology (IT) landscape manager" to find the root cause of problems and predict future issues to stop them before they happen, all without needing a person to step in (Claim 1). ## What it does NOT cover - Does not cover systems that require manual tuning to determine event cluster boundaries, as the abstract states it does this "without requiring manual tuning." - Does not cover systems that only generate a knowledge graph once without a continuous feedback loop and update mechanism. - Does not cover systems that determine root causes or predict events without using a machine learning model to process feedback and spatiotemporal context. - Does not cover systems where the IT landscape manager requires human intervention to determine root causes or predict future events. - Does not cover systems that only use one level of detail for updating the knowledge graph, as it specifies updating at a "second level" different from the "first level" of feedback. ## The clever bit The novelty lies in the continuous, self-improving loop where a machine learning model constantly refines its understanding of cause-and-effect in an IT system by requesting and processing human feedback on its causal graphs, then using that to update its knowledge graph and predict future problems automatically. ## Real-world examples 1. Automated IT operations (AIOps) platforms 2. Network monitoring and diagnostics tools 3. Cloud infrastructure management systems 4. Predictive maintenance for software systems ## Why it matters In complex IT environments, finding the real cause of a problem can be like finding a needle in a haystack. This patent aims to automate that process, making IT systems more reliable and efficient. By predicting and preventing issues, it could significantly reduce downtime and operational costs for businesses relying on large-scale IT infrastructure. ## Frequently asked questions ### What does How AI Learns to Fix IT Problems by Asking for Feedback cover? This patent describes an AI system that continuously learns to identify and prevent IT issues by building a map of cause-and-effect relationships, getting human feedback, and automatically updating its understanding. ### Who owns patent US 12505360? Bmc Helix owns this patent, granted in 2025. ### When does this patent expire? This patent is expected to expire on September 23, 2042, when the invention enters the public domain. ### What problem does this patent solve? In complex IT environments, finding the real cause of a problem can be like finding a needle in a haystack. This patent aims to automate that process, making IT systems more reliable and efficient. By predicting and preventing issues, it could significantly reduce downtime and operational costs for businesses relying on large-scale IT infrastructure. ### What does this patent NOT cover? Does not cover systems that require manual tuning to determine event cluster boundaries, as the abstract states it does this "without requiring manual tuning." **Full plain-English explainer:** https://patentbrief.org/patent/us/12505360/continuous-knowledge-graph-generation-using-causal-event-graph-feedback **Original patent:** https://patents.google.com/patent/US12505360 --- _Source: PatentBrief — https://patentbrief.org. Patent facts are from public records; the plain-English explanation is PatentBrief's._ ## Related patents Semantically similar inventions in the PatentBrief corpus: - [How AI Connects Different Databases Using Knowledge Graphs](https://patentbrief.org/patent/us/11507851/system-and-method-of-integrating-databases-based-on-knowledge-graph) — This patent describes a server-based method that uses artificial intelligence and two learning models to automatically find and integrate connections between data fields and data values across multiple databases that have different structures. - [How Computers Calculate Probabilities in Large Knowledge Bases](https://patentbrief.org/patent/us/9361579/large-scale-probabilistic-ontology-reasoning) — A method for finding answers in a database of uncertain facts by ignoring probabilities to find a solution first, then calculating how likely that solution is based on the underlying evidence. - [How AI Learns to Control Game Characters Based on Their Surroundings](https://patentbrief.org/patent/us/10607134/artificially-intelligent-systems-devices-and-methods-for-learning-andor-using-an-avatars-circumstances-for-autonomous-avatar-operation) — A system that allows digital characters to automatically perform actions by matching their current environment to previously learned experiences stored in a database. - [How to Automatically Detect and Fix Changes in AI Model Data](https://patentbrief.org/patent/us/10599957/systems-and-methods-for-detecting-data-drift-for-data-used-in-machine-learning-m) — This patent describes a system that automatically notices when the real-world data an AI model sees changes, causing its predictions to become less accurate, and then fixes the model. - [How to Update AI on Small Devices with Slow Internet](https://patentbrief.org/patent/us/20250363357/systems-and-methods-for-deploying-and-updating-neural-networks-at-the-edge-of-a-) — This patent describes a method for efficiently updating artificial intelligence models on small, internet-connected devices, like smart cameras, by sending only the changes, or 'patches,' instead of the entire updated model, which saves bandwidth.