CIFAR-10 and MNIST datasets when using maximum mean discrepancy (MMD) with learned instead of fixed features. For a related method and in a similar spirit, Santos et al. (2019) show that for image

Random Reshuffling (RR) is an algorithm for minimizing finite-sum functions that utilizes iterative gradient descent steps in conjunction with data reshuffling. Often contrasted with its sibling Stochastic Gradient Descent (SGD), RR is usually faster in practice and enjoys significant popularity in convex and non-convex optimization. The convergence rate of RR has attracted substantial attention recently and, for strongly convex and smooth functions, it was shown to converge faster than SGD if 1) the stepsize is small, 2) the gradients are bounded, and 3) the number of epochs is large.

Firstly, many existing attribution techniques require either computing the gradients [58] or multiple queries to the model [19], both of which are slow and computationally expensive. In contrast, ICL is often applied inreal-time to a large foundation model [12] that necessitates the attribution approaches for ICL to be fast and efficient.

This scaling behavior coincides with the one reportedinTripuranenietal. (2020)usingGaussiancomplexity.

We first present the procedures of Stein path distance calculation in Algorithm 1. The calculation of Stein path distance mainly has three steps. (line 1). A.2 Procedure of multiple-proxies As mentioned in Section 2.3.3, the multiple-proxies algorithm is given by: min We now provide the optimization details on the multiple-proxies algorithm. A.3 Procedure of proxy Stein path loss As we have presented in Section 2.3.3, the proxy Stein path distance is defined as: P We conduct extensive experiments on two popularly used real-world datasets, i.e., Douban [ The details of Douban and Amazon datasets are shown in Table 1 and Table 2. B.2 Visualization Amazon Music are shown in Figure 1.