Coresets for learning tasks with time series data (Lingxiao Huang)
Abstract
We study several learning tasks (e.g.,
regression or clustering) with time series data. These problems have gained
importance across many fields including biology, medicine, and economics due to
the proliferation of sensors facilitating real-time measurement and rapid drop
in storage costs. We will introduce how to construct coresets, which is a small
subset of the data allowing for fast approximate inference, for the maximum
likelihood objective for these problems, and discuss the differences from
static data. We empirically assess the performance of our coreset with
synthetic data and real-world data.
Time
2021-06-18 15:00-15:30
Speaker
Lingxiao Huang, Huawei TCS Lab
Room
Guangdong Hotel Shanghai
