ALPaCA vs. GP-based Prior Learning: A Comparison between two Bayesian Meta-Learning Algorithms
This work provides a comparative analysis for researchers in meta-learning, helping them choose between efficiency and performance in Bayesian methods, but it is incremental as it evaluates existing algorithms without introducing new ones.
The paper compares two Bayesian meta-learning algorithms, ALPaCA and PACOH, analyzing their theoretical differences and empirical performance on synthetic and real-world datasets. It finds that ALPaCA is computationally faster due to a linear kernel, while GP-based methods like PACOH offer more flexibility and achieve better results with kernels like the Squared Exponential kernel.
Meta-learning or few-shot learning, has been successfully applied in a wide range of domains from computer vision to reinforcement learning. Among the many frameworks proposed for meta-learning, bayesian methods are particularly favoured when accurate and calibrated uncertainty estimate is required. In this paper, we investigate the similarities and disparities among two recently published bayesian meta-learning methods: ALPaCA (Harrison et al. [2018]) and PACOH (Rothfuss et al. [2020]). We provide theoretical analysis as well as empirical benchmarks across synthetic and real-world dataset. While ALPaCA holds advantage in computation time by the usage of a linear kernel, general GP-based methods provide much more flexibility and achieves better result across datasets when using a common kernel such as SE (Squared Exponential) kernel. The influence of different loss function choice is also discussed.