Distributed Self-Paced Learning in Alternating Direction Method of Multipliers
This work addresses the problem of inefficient distributed training in self-paced learning for machine learning practitioners, though it is incremental as it adapts an existing framework to a new setting.
The paper tackles the challenge of scaling self-paced learning to large datasets by reformulating it into a distributed setting, proposing a novel method that optimizes model and instance weights in parallel using a consensus alternating direction method of multipliers, and demonstrating superior performance over existing methods in experiments on synthetic and real datasets.
Self-paced learning (SPL) mimics the cognitive process of humans, who generally learn from easy samples to hard ones. One key issue in SPL is the training process required for each instance weight depends on the other samples and thus cannot easily be run in a distributed manner in a large-scale dataset. In this paper, we reformulate the self-paced learning problem into a distributed setting and propose a novel Distributed Self-Paced Learning method (DSPL) to handle large-scale datasets. Specifically, both the model and instance weights can be optimized in parallel for each batch based on a consensus alternating direction method of multipliers. We also prove the convergence of our algorithm under mild conditions. Extensive experiments on both synthetic and real datasets demonstrate that our approach is superior to those of existing methods.