AI 中文总结
EulerLoRA是LoRA的随机扩展，通过共享低秩适配器的秩-1分量采样结构化变化生成预测轨迹，在多数据集上性能相当或优于LoRA-Ensemble，参数量减少约69%，实现参数高效微调与预测不确定性估计。
AI 中文摘要
低秩适配（LoRA）可实现参数高效微调，但标准LoRA仅生成单一确定性模型，无法直接支持预测不确定性估计。我们提出EulerLoRA，它是LoRA的随机扩展，通过在共享低秩适配器的秩-1分量上采样结构化变化生成多条预测轨迹，同时在期望中保留确定性LoRA变换。我们在CIFAR-10、CIFAR-100和HAM10000数据集上对视觉Transformer评估EulerLoRA，并在SVHN上进行分布外检测。在这些基准测试中，EulerLoRA相比强大的LoRA-Ensemble基线取得相当或更优的性能。使用两个秩为20的适配器时，EulerLoRA仅需约300万个可训练适配器参数，而秩为8、16个适配器的LoRA-Ensemble则需要约1000万个，对应可训练参数减少约69%。这些结果表明，可从少量共享适配器中获得有用的预测多样性。
英文摘要
Low-Rank Adaptation (LoRA) enables parameter-efficient fine-tuning, but standard LoRA produces a single deterministic model and does not directly support predictive uncertainty estimation. We introduce EulerLoRA, a stochastic extension of LoRA that generates multiple predictive trajectories by sampling structured variations along the rank-one components of shared low-rank adapters, while preserving the deterministic LoRA transformation in expectation. We evaluate EulerLoRA with vision transformers on CIFAR-10, CIFAR-100, and HAM10000, together with out-of-distribution detection on SVHN. Across these benchmarks, EulerLoRA achieves comparable or improved performance relative to strong LoRA-Ensemble baselines. Using two rank-20 adapters, EulerLoRA requires approximately 3 million trainable adapter parameters, compared with about 10 million for a rank-8, 16-adapter LoRA-Ensemble, corresponding to roughly 69% fewer trainable parameters. These results show that useful predictive diversity can be obtained from a small number of shared adapters.
