FairCompressAgent framework balances fairness accuracy and compression for FPGA use
FairCompressAgent: An Agentic Framework for Fairness-Aware Model Compression for FPGA Deployment
Artificial IntelligenceHardware Architecture
Summary
Making machine learning models smaller to run efficiently on hardware like FPGAs can sometimes reduce fairness or accuracy. The authors propose FairCompressAgent, a system that intelligently picks ways to shrink models while keeping fairness and accuracy balanced under given limits. It tries different methods and learns from the results, adapting when user needs change. Tests show it can find good compressed models with less trial and error, maintaining or improving fairness and accuracy at the same time.
What this means in practice
- •For ai system developers: Use FairCompressAgent to select model compression settings that maintain fairness and accuracy for FPGA-based AI products under resource constraints.
- •For hardware deployment teams: Adapt machine learning models efficiently for hardware with changing fairness or accuracy requirements using the interactive compression workflow provided by FairCompressAgent.
Tested on one dataset.
Authors
Yuanbo Guo, Yiyu Shi
Abstract
Fairness-aware model compression requires selecting methods and configurations that balance accuracy, fairness, and deployment cost. These decisions become more difficult when compression methods are composed or the user's requirements change. In this paper, we propose FairCompressAgent (FCA), an agentic framework that integrates fairness-aware pruning, incremental quantization, and sparse low-rank factorization through a common operator interface. A language-model planner uses model profiles and measured outcomes to select compression configurations, while an execution layer performs compression, fine-tuning, evaluation, and constraint-based selection. FCA also supports requirement updates and reports the remaining violation when a request cannot be satisfied. Experiments on Fitzpatrick-17k with VGG-11 compare four search methods over 40 measured configurations. Under the accuracy-constrained request, FCA selects a compressed model with 59.54% less inference tensor storage, while validation average precision increases from 0.5141 to 0.5233 and equalized opportunity (EOpp) decreases from 0.2251 to 0.2168. It reaches the same final selection as one-shot planning with 7.33 versus 12 candidate evaluations on average, under their respective stopping policies. Repeated fine-tuning, held-out testing, and online requirement updates characterize the stability and interactive use of this compression workflow. The results demonstrate how measured feedback and explicit constraints support the selection and interactive refinement of fairness-aware compression configurations.