Guideline Recommendation System Using Graph Embeddings
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In large-scale cloud environments, human operators face overwhelming challenges in managing and finding suitable guidelines or labels for managed objects due to the exponential increase in the sheer size and variety of policies and tags, making it difficult to determine if a guideline exists or needs to be created.
Innovation Solution
A system and method that utilize transductive and inductive embedding techniques on a graph of the computing environment to generate embeddings for nodes, which are then used for link classification to recommend guidelines for target managed objects, thereby automating the process of guideline selection and recommendation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If manual assignment of guidelines to managed objects is used, then flexibility and control are maintained, but operator workload and time consumption increase exponentially with system scale
Solution Approach 1:
The system performs self-service by automatically generating guideline recommendations through machine learning models that analyze the computing environment graph, eliminating the need for operators to manually search and assign guidelines, thus reducing operational burden and time consumption
Solution Approach 2:
The manual mechanical process of searching and assigning guidelines is replaced with an automated intelligent system using transductive and inductive embedding techniques combined with link classification algorithms, transforming the operational workflow from manual to automated
2Reliability
If the number of guidelines and policies increases to cover all managed objects, then coverage and compliance improve, but system complexity and difficulty of management increase
Solution Approach 1:
The system introduces an intermediary intelligent recommendation layer between the large number of available guidelines and the managed objects, using machine learning models to filter and prioritize relevant guidelines, thus maintaining comprehensive coverage while simplifying the management interface
Solution Approach 2:
The system changes the parameter of guideline recommendation from exhaustive listing to ranked scoring based on link classification probabilities, transforming the presentation of guidelines from a static comprehensive list to a dynamic prioritized recommendation set
3Productivity
If automated recommendation systems are implemented, then operator workload decreases, but computational complexity and processing requirements increase
Solution Approach 1:
The system performs preliminary action by pre-computing transductive and inductive embeddings for all nodes in the computing environment graph, storing these embeddings for efficient retrieval during recommendation generation, thus reducing real-time computational requirements
Solution Approach 2:
The computational process is segmented into distinct stages: graph construction, transductive embedding generation, inductive embedding generation, and link classification, allowing each component to be optimized and processed independently, reducing overall computational complexity
Data Source
AI summary
System and computer-implemented method for recommending guidelines for managed objects for a computing environment uses a transductive embedding technique on a graph of the computing environment to generate initial embeddings for the nodes of the graph. An inductive embedding technique is then applied on the initial embeddings and features of the nodes of the graph to produce final embeddings for the nodes of the graph, which are used to execute a link classification operation on the final embeddings for at least some nodes of the graph to select a recommended guideline for a target managed object.