Neural Network Model Compression via Block Segmentation and Parameter Tuning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Users face difficulties in finding and deploying artificial intelligence models with desired performance on target devices due to the lack of optimized neural network models that can run effectively on these devices.
Innovation Solution
An electronic apparatus that compresses neural network models based on user-input compression methods and device-specific parameters, allowing users to visualize and configure the model structure, identify compressible and non-compressible blocks, and apply compression settings to optimize the model for the target device.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a state-of-the-art neural network model is used to achieve excellent performance, then model accuracy is improved, but the model cannot be operated on a target device due to size and complexity constraints
Solution Approach 1:
The neural network model is divided into multiple blocks, and the compression process is applied block by block. This segmentation allows the system to manage and compress large models in manageable units, reducing memory requirements and enabling deployment on resource-constrained devices while maintaining overall model accuracy
Solution Approach 2:
The system changes parameters such as block compression ratios, quantization precision, and pruning thresholds to optimize the balance between model size and accuracy. By adjusting these parameters, the model can be adapted to run on target devices with limited resources while preserving essential performance characteristics
2Adaptability or versatility
If model compression is applied to reduce model size for device deployment, then device compatibility is improved, but model performance may deteriorate
Solution Approach 1:
The system uses feedback mechanisms to evaluate model performance at each compression stage. By monitoring accuracy metrics and adjusting compression parameters accordingly, the system can achieve maximum compression while maintaining acceptable performance levels on target devices
Solution Approach 2:
The system applies compression selectively to different blocks rather than uniformly across the entire model. By identifying and protecting critical blocks from aggressive compression while applying higher compression to less sensitive blocks, the system maintains overall model performance while achieving size reduction
3Ease of operation
If automated model compression is performed to simplify the process, then ease of operation is improved, but loss of information occurs in compression configuration
Solution Approach 1:
The system automatically analyzes the model structure, identifies compressible blocks, and determines optimal compression parameters without requiring extensive user input. This self-service approach simplifies operation while maintaining detailed configuration information through automated model analysis and evaluation
Data Source
AI summary
Disclosed is a method of compressing a neural network model that is performed by a computing device. The method includes receiving a trained model and compression method instructions for compressing the trained model, identifying a compressible block and a non-compressible block among a plurality of blocks included in the trained model based on the compression method instructions, transmitting a command to a user device that causes the user device to: display a structure of the trained model representing a connection relationship between the plurality of blocks on a first screen such that the compressible block and the non-compressible block are visually distinguished, and display, on a second screen, an input field operable to receive a parameter value entered by a user for compression of the compressible block, and compressing the trained model based on the parameter value entered by the user in the input field.