SOLAR: Communication-Efficient Model Adaptation via Subspace-Oriented Latent Adapter Reparametrization

arXiv:2604.08368v1 Announce Type: cross Abstract: Parameter-efficient fine-tuning (PEFT) methods, such as LoRA, enable scalable adaptation of foundation models by injecting low-rank adapters. However, their communication and storage costs remain a major bottleneck in resource-constrained settings. We propose SOLAR (Subspace-Oriented Latent Adapter Reparameterization), a post-training compression framework that substantially reduces the communication cost (i.e., the number of parameters to transmit or store) of PEFT adapters. SOLAR expresses each PEFT update as a linear combination of basis vectors formed from the foundation model's singular vectors with controlled random perturbations. By exploiting the subspace similarity (the alignment of principal directions) between the foundation model and task-specific fine-tuned updates, SOLAR decouples the adapter size from PEFT structure and ensures compact yet expressive representations. It is model-agnostic and compatible with existing PEFT methods, including LoRA, AdaLoRA, and other adapter modules. We theoretically establish a bound on the reconstruction error. Experiments on language and vision tasks using LLaMA, GPT, and ViT models demonstrate that SOLAR preserves task performance while significantly reducing model representation sizes, offering an effective and communication-efficient solution for deployment in distributed systems and edge devices.

Related

• TalkLoRA: Communication-Aware Mixture of Low-Rank Adaptation for Large Language Models
• LoRA-DA: Data-Aware Initialization for Low-Rank Adaptation via Asymptotic Analysis
• An empirical study of LoRA-based fine-tuning of large language models for automated test case generation
• FLeX: Fourier-based Low-rank EXpansion for multilingual transfer
• A Parameter-Efficient Transfer Learning Approach through Multitask Prompt Distillation and Decomposition for Clinical NLP
• LoFT: Parameter-Efficient Fine-Tuning for Long-tailed Semi-Supervised Learning in Open-World Scenarios

Source: arXiv cs.CL | 2026-04-10