Dawei Zhou and Yao Zhou, PhD students in computer science, will present the work of iSchool Associate Professor Jingrui He's research group, the iSAIL Lab, at The Web Conference 2020. The conference, which will be held virtually from April 20-24, will address the evolution and current state of the Web through the lens of computer science, computational social science, economics, public policy, and Web-based applications.
Yao Zhou will present "Crowd Teaching with Imperfect Labels." According to the researchers, the need for annotated labels to train machine learning models led to a surge in crowdsourcing—that is, collecting labels from nonexperts. In this paper, He's research group proposes an adaptive scheme that could improve both data quality and workers’ labeling performance, in which "the teacher teaches the workers using labeled data, and in return, the workers provide labels and the associated confidence level based on their own expertise." The researchers demonstrate the proposed framework through experiments on multiple real-world image and text data sets.
Dawei Zhou will present "Domain Adaptive Multi-Modality Neural Attention Network for Financial Forecasting." The paper describes the researchers' work on financial time series analysis, which is a challenging task as the problems are always accompanied by data heterogeneity. For instance, in stock price forecasting, a successful portfolio with bounded risks usually consists of a large number of stocks from diverse domains, and forecasting stocks in each domain can be treated as one task; within a portfolio, each stock is characterized by temporal data collected from multiple modalities, which corresponds to the data-level heterogeneity. To address this problem, He's group proposed a generic time series forecasting framework named Dandelion, which leverages the consistency of multiple modalities and explores the relatedness of multiple tasks using a deep neural network.
He's general research theme is to design, build, and test a suite of automated and semi-automated methods to explore, understand, characterize, and predict real-world data by means of statistical machine learning. She received her PhD in machine learning from Carnegie Mellon University.