arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2306.09293cs.LGcs.AI
收藏

[实验与分析] 评估基于采样的技术用于训练多层感知机的可行性

[Experiments & Analysis] Evaluating the Feasibility of Sampling-Based Techniques for Training Multilayer Perceptrons

University of Illinois Chicago(伊利诺伊大学芝加哥分校)

Sana Ebrahimi, Rishi Advani, Abolfazl Asudeh

AI总结:

该研究针对神经网络训练耗时的问题,在有限CPU资源下评估两类基于采样的矩阵乘法近似方法训练多层感知机的可行性,通过理论分析与实验证明前馈近似存在可扩展性障碍,并指出未来研究方向。

摘要:

众所周知,神经网络的训练过程十分耗时,深度架构只会加剧这一问题。该过程主要由矩阵运算构成,其中矩阵乘法是瓶颈。目前已有多种基于采样的技术被提出,通过近似矩阵乘积来加快深度神经网络的训练速度。这些技术分为两类:(i)在每次迭代中采样每个隐藏层的节点子集作为激活节点;(ii)从前一层采样节点子集,利用来自采样节点的边来近似当前层的激活值。在这两种情况下,矩阵乘积都仅使用选定的样本进行计算。本文中,我们在计算资源有限的CPU机器上评估了这些方法的可行性。我们将两个研究方向关联为神经网络背景下近似矩阵乘法的特例,给出了否定性理论分析,表明前馈近似是可扩展性的一大障碍。我们开展了全面的实验评估,揭示了所研究方法面临的最紧迫挑战与局限性。我们观察到,基于哈希的节点选择方法无法扩展到大量层数,印证了我们的理论分析。最后,我们指出了未来的研究方向。

英文摘要:

The training process of neural networks is known to be time-consuming, and having a deep architecture only aggravates the issue. This process consists mostly of matrix operations, among which matrix multiplication is the bottleneck. Several sampling-based techniques have been proposed for speeding up the training time of deep neural networks by approximating the matrix products. These techniques fall under two categories: (i) sampling a subset of nodes in every hidden layer as active at every iteration and (ii) sampling a subset of nodes from the previous layer to approximate the current layer's activations using the edges from the sampled nodes. In both cases, the matrix products are computed using only the selected samples. In this paper, we evaluate the feasibility of these approaches on CPU machines with limited computational resources. Making a connection between the two research directions as special cases of approximating matrix multiplications in the context of neural networks, we provide a negative theoretical analysis that shows feedforward approximation is an obstacle against scalability. We conduct comprehensive experimental evaluations that demonstrate the most pressing challenges and limitations associated with the studied approaches. We observe that the hashing-based node selection method is not scalable to a large number of layers, confirming our theoretical analysis. Finally, we identify directions for future research.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑