NeuSemSlice: Towards Effective DNN Model Maintenance via Neuron-level Semantic Slicing

TL;DR

NeuSemSlice introduces a neuron-level semantic slicing framework that enhances deep neural network maintenance tasks by precisely identifying and manipulating critical semantic components, outperforming existing layer-level approaches.

Contribution

The paper presents a novel semantic slicing technique for neuron-level analysis, enabling more effective and fine-grained model restructuring, re-adaptation, and incremental development in DNNs.

Findings

NeuSemSlice significantly outperforms baselines in model restructuring.

The framework effectively preserves critical neurons during model updates.

Semantic slicing improves the precision of neuron manipulation in DNN maintenance.

Abstract

Deep Neural networks (DNNs), extensively applied across diverse disciplines, are characterized by their integrated and monolithic architectures, setting them apart from conventional software systems. This architectural difference introduces particular challenges to maintenance tasks, such as model restructure (e.g., model compression), re-adaptation (e.g., fitting new samples), and incremental development (e.g., continual knowledge accumulation). Prior research addresses these challenges by identifying task-critical neuron layers, and dividing neural networks into semantically-similar sequential modules. However, such layer-level approaches fail to precisely identify and manipulate neuron-level semantic components, restricting their applicability to finer-grained model maintenance tasks. In this work, we implement NeuSemSlice, a novel framework that introduces the semantic slicing technique to effectively identify critical neuron-level semantic components in DNN models for semantic-aware model maintenance tasks. Specifically, semantic slicing identifies, categorizes and merges critical neurons across different categories and layers according to their semantic similarity, enabling their flexibility and effectiveness in the subsequent tasks. For semantic-aware model maintenance tasks, we provide a series of novel strategies based on semantic slicing to enhance NeuSemSlice. They include semantic components (i.e., critical neurons) preservation for model restructure, critical neuron tuning for model re-adaptation, and non-critical neuron training for model incremental development. A thorough evaluation has demonstrated that NeuSemSlice significantly outperforms baselines in all three tasks.