wise Hierarchical clustering can be slow (has to make several merge/split decisions) No clear consensus on which of the two produces better clustering Feng Li (SDU) K-Means December 28, 2021 45 / 46

mislabeling due to human error, data is labeled by multiple human labelers and the label that wins the consensus is assigned to the example. Given all the costs involved, it is imperative to utilize all the training

HyperBand bring the field of HPO closer to the evolutionary approaches which are based on biological mechanisms like mutation and natural selection. The promotion of better performing trials to the next iteration selection or survival of the fittest. Population Based Training4 (PBT) incorporate these biological mechanisms to evolve better models. It spawns a fixed number of trials (referred as population) and trains

through several works27 where the authors propose transformer variants with efficient self-attention mechanisms. These ideas tackle the quadratic complexity at various levels. The simplest idea is to chunk the