Exception Featurization for Automated Code Error Resolution
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing developer tools fail to provide effective information and pointers for resolving errors and exceptions in code, requiring developers to spend significant time searching through irrelevant results, and existing search engines struggle to efficiently find specific error or exception solutions.
Innovation Solution
A computer-implemented method that receives an exception generated from programming code, generates exception features from search logs, and executes a machine learning model trained on extracted exceptions and features to provide relevant information for resolving the exception, using a high-performance model that extracts exceptions from unstructured query text with an F1-score of 0.82.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If developers use traditional web search to find exception solutions, then they can access general information, but they spend significant time sifting through irrelevant results
Solution Approach 1:
The system extracts exception-specific features from search logs and training data, isolating the most relevant characteristics of exceptions (error messages, stack traces, contextual information) to create targeted search queries that filter out irrelevant results and directly retrieve helpful exception solutions
Solution Approach 2:
The system transforms raw exception data into structured features with specific parameters (exception type, severity, contextual metadata) and uses these transformed parameters to query the machine learning model, which then retrieves precisely matched solutions rather than generic search results
2Measurement precision
If existing search engines are used to find exception solutions, then broad information is available, but they struggle to efficiently find specific error solutions
Solution Approach 1:
The system performs preliminary actions by pre-processing search logs to extract and structure exception features, training the machine learning model in advance with labeled exception data, and building a knowledge base of exception patterns before actual exception resolution is needed, enabling rapid and precise matching when exceptions occur
Solution Approach 2:
The machine learning model acts as an intermediary between the raw exception and the search log database, translating exception features into precise search queries and filtering results to return only the most relevant solutions, thereby bridging the gap between general search capabilities and specific exception resolution
3Ease of operation
If developers manually search for exception information, then they can find solutions, but the process requires significant manual effort and time
Solution Approach 1:
The system enables self-service by automatically processing exception information, extracting relevant features, querying the trained machine learning model, and returning targeted solutions without requiring developer intervention in the search process, thereby making exception resolution easier and faster
Data Source
AI summary
A computer implemented method includes receiving an exception generated based on programming code, generating exception features from the received exception, the generated exception features being generated based on a set exception features derived from search logs, and executing a machine learning model on the received exception and generated exception features to provide information from the search logs identified as most helpful to resolve the received exception, wherein the machine learning model was trained on training data comprising extracted exceptions and the set of exception features derived from the search logs.