Shion Ishikawa

In short

I’m a research scientist at Rakuten Institute of Technology. I have an interest in reinforcement learning, bandit algorithm, off-policy evaluation, causal discovery, bioinformatics.

Research

Paper

1. Position Bias Estimation with Item Embedding for Sparse Dataset (2024, WWW, short)
2. Counterfactual Model Selection in Contextual Bandits (2025, SIGIR, short)

Workshop

1. Dynamic collaborative filtering Thompson Sampling for cross-domain advertisements recommendation (2022, AdKDD)

Article

1. 楽天グループにおける多腕バンディットを用いたMulti-slot型広告の最適化 (2024, 応用数理)

Master’s thesis

Long-read cap-trapping sequencing reveals whole genome associations between transcript start sites usage and splice sites usage

Bachelor’s thesis

Learning causal Bayesian Network structures from lifestyle data and suggesting an improvement for visceral fat.