Human-Centric AI for Collective Knowledge and Wisdom

Interactive AI for Causal Knowledge Elicitation(因果モデル構築AI)

Effective disaster drills and exercises require appropriate scenarios reflecting concrete disaster situations. It is however not easy to manually create such a scenario with enough details and validity, because it is fundamentally difficult to comprehensively predict and assume disaster situations that may occur in various phases through a chain of causality from the primary damage. In order to make a scenario creation easier and more efficient, some support tools are necessary, in particular for predicting what kind of situations will happen through a causal chain from a base disaster assumption. In this paper, we proposed a simple and practical causal model consisting of three elements: cause, precondition, and effect, which can capture indirect causal relationships between two events by introducing the concept of preconditions. We also developed an interactive method with a GUI to elicit causal knowledge about disaster situations based on the model. Users can enter possible events that can occur in a disaster as well as countermeasures against those events by answering the questions presented on the GUI. Then the entered sentences are processed to identify causal elements automatically by a newly developed NLP tech-niques, and finally those elements are integrated into the database. The proposed method still has a room for improvement, however its performance is satisfying and can be expected to be utilized as a technical base for the creation of effective disaster scenarios.


Interactive AI to Collaboratively Elicit Causal Knowledge

GUI for Knowledge Elicitation

Scenario Co-Creation AI for Disaster Exercise(訓練用災害シナリオ作成支援AI)

Widely applied across many sectors, disaster training and exercises pre-embedded within a society can serve as a driving force for strengthening its disaster resilience. To improve the efficiency and effectiveness of this traditional system, which has great potential to enhance social resilience to natural catastrophes, this study proposes a model that can be used to produce and share disaster scenarios, designs a prototype support system for creating disaster scenarios and exercise materials based on the proposed model, and evaluates its potential use.

D-Cotext Model

Collaborative D-Scenario Creation

Interviewer Agent for Cognitive Task Analysis(認知タスク分析用インタビューAI)

A chat-based interviewer agent to elicit know-how of subject matter experts(SMEs) was developed. This agent automatically generates questions and responses to the answers from the human interviewee. This automatic response is based on the shallow intelligence or artificial non-intelligence incorporating an interview method of cognitive task analysis as well as pattern matching to the dialogue characteristics specific to interview dialogues. A four-layered framework was proposed for the evaluation of the interviewer agent. The agent was tested with 6 SMEs and evaluated by the proposed framework. The result showed that the agent effectively elicited experts' knowledge while it still has a limitation in generating natural human-like responses. 

インタビューは他人から情報を引き出すために良く用いられる方法であるが,これを実際に準備,実施するには非常に労力がかかるだけでなく,スキルも必要とされる.エキスパートの状況認識や意思決定などの認知的側面を分析することはCognitive Task Analysis (CTA:認知タスク分析)と呼ばれ,インタビューが頻繁に用いられる.本研究では,このCTAを自動で行うエージェントの開発を行っている.対話エージェントの技術に1)インタビュー技法(Critical Decision Method),会話研究における発話機能や出現パターンに関する知見,3)熟練インタビュアのノウハウ,を実装し,「災害時の看護活動における情報処理行動」に関するインタビュー等の様々な課題に適用して人間に劣らない情報獲得効率を確認している.本研究では単にインタビュアの代行ツールの開発を目的としているだけではなく,対話分析ツールとして用いることも目的としている.エージェントの応答に対する人が感じる違和感に注目,分析することによって,人がなぜエージェントの対話に違和感を抱くのか,自然な対話とは何か,といった対話の暗黙コードに関する研究も行っている. 

Architecture of CTAgent

Transitions of Question Types

Conversation Log.