Artificial Intelligence for Exam Question Answering (AI4Exams) is a 1-day meeting held at The National Institute of Informatics (NII), Tokyo on June 10th. The goal of the workshop is to discuss ongoing research in the area of exam question answering, foster collaboration, increase international awareness of on-going projects, and formalize the task (data and tools) in order to improve the quality of research moving forward.

Organizers:

Yusuke Miyao, NII
Jason Naradowsky, UCL
Sebastian Riedel, UCL


Schedule:

---

10:15-10:30	Reception
10:30-10:50	What is Exam QA? - Sebastian Riedel (UCL)
10:50-11:10	The current state of data-driven QA - Jason Naradowsky (UCL)

Break

Talks I:

11:20-11:45	"Challenge to complex essay questions of university entrance exams"
	Hideyuki Shibuki and Tatsunori Mori (Yokohama National University)
11:45-20:10	"QA as humans do: creating named entity ontology"
	Satoshi Sekine (NYU)
12:10-12:35	"Learning to Generate Textual Data"
	Pontus Stenetorp, (UCL)

Lunch

1:20-2:20	Panel Discussion I: Moderated by Yusuke Miyao
	We will be joined by Ai2's Peter Clark to discuss what we mean by exam QA as a task, how it (and projects like the Todai Robot project) relate to ongoing efforts in general question anwsering and machine reading. Can we abstract over these, can we develop a shared infrastructure for collaboration, and what exciting future directions exist in this line of research?

Break

Talks II:

2:30-2:55	"The SmartReader Project: Automatic Generation of Questions and Answers from Text"
	Teruko Mitamura (CMU)
2:55-3:20	"Exam, exam, exam: history, legal and medical examinations"
	Yoshinobu Kano (Shizuoka University)
3:20-3:45	"Taking the English exam in the Todai Robot project"
	Ryuichiro Higashinaka (NTT)

Break

3:55-4:30	Panel Discussion II: Moderated by Jason Naradowsky
	An organizational discussion where we will discuss the success of the workshop and the possibility for future events of this kind.

4:30	Closing remarks