RPIQ: Residual-Projected Multi-Collaboration Closed-Loop and Single Instance Quantization for Visually Impaired Assistance

2026-01-06

arXiv.org

AI Summary

This work addresses the challenge of deploying large language models in assistive devices for the visually impaired, where high memory consumption and inference costs are major bottlenecks. Existing quantization methods often suffer from performance degradation due to inter-block error accumulation, which is typically overlooked. To mitigate this, the authors propose RPIQ, a novel quantization framework that integrates instance-wise calibration, Gauss–Seidel iterative quantization, residual projection, and a multi-agent closed-loop compensation mechanism to effectively suppress error propagation in low-bit quantization. Evaluated on OPT, Qwen, LLaMA, and CogVLM2, RPIQ achieves 4-bit compression with 60%–75% peak memory reduction while maintaining near full-precision inference performance, demonstrating strong results on text understanding and visual question answering tasks.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Computer Vision: Large Vision Models
• Natural Language Processing: (Large) Language Models

Application Category

• Search and Retrieval-Augmented AI: Large language models for search
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation

Abstract

Visually impaired users face significant challenges in daily information access and real-time environmental perception, and there is an urgent need for intelligent assistive systems with accurate recognition capabilities. Although large-scale models provide effective solutions for perception and reasoning, their practical deployment on assistive devices is severely constrained by excessive memory consumption and high inference costs. Moreover, existing quantization strategies often ignore inter-block error accumulation, leading to degraded model stability. To address these challenges, this study proposes a novel quantization framework -- Residual-Projected Multi-Collaboration Closed-Loop and Single Instance Quantization(RPIQ), whose quantization process adopts a multi-collaborative closed-loop compensation scheme based on Single Instance Calibration and Gauss-Seidel Iterative Quantization. Experiments on various types of large-scale models, including language models such as OPT, Qwen, and LLaMA, as well as vision-language models such as CogVLM2, demonstrate that RPIQ can compress models to 4-bit representation while significantly reducing peak memory consumption (approximately 60%-75% reduction compared to original full-precision models). The method maintains performance highly close to full-precision models across multiple language and visual tasks, and exhibits excellent recognition and reasoning capabilities in key applications such as text understanding and visual question answering in complex scenarios. While verifying the effectiveness of RPIQ for deployment in real assistive systems, this study also advances the computational efficiency and reliability of large models, enabling them to provide visually impaired users with the required information accurately and rapidly.

Problem

• visually impaired assistance
• model quantization
• memory consumption
• inference cost
• error accumulation

Innovation

• quantization
• closed-loop compensation
• single instance calibration
• Gauss-Seidel iteration
• large model compression