Data Detectives: A Tabletop Card Game about Training Data
Data Detectives: A Tabletop Card Game about Training Data
复制标题
数据侦探：关于训练数据的桌面纸牌游戏
DOI：
10.1145/3502717.3532128
复制
发表时间：
2022
期刊：
影响因子：
--
通讯作者：
Hammer, Jessica
中科院分区：
文献类型：
--
作者：
Solyst, Jaemarie;Kim, Jennifer;Ogan, Amy;Hammer, Jessica
年轻人经常接触利用监督机器学习的人工智能技术。然而，众所周知，有偏差的训练数据会导致有害的算法偏差。因此，年轻人和家庭理解机器学习中的训练数据非常重要。我们提出了数据侦探，一个关于训练数据的儿童友好的桌面纸牌游戏。基于三个基于研究的设计原则：支持好奇心的低风险实验，促进对话的游戏，以及抽象概念的有形和具体学习，游戏支持学习监督机器学习中训练数据的高级机制，以及练习与算法偏见相关的训练数据的批判性讨论。这款游戏为人工智能扫盲提供了机会，旨在促进有趣的同龄人和孩子与父母的学习。
Youth regularly interface with AI technology that leverages supervised machine learning. However, it is well-known that biased training data can result in harmful algorithmic bias. Thus, it is important that youth and families understand training data in machine learning. We present Data Detectives, a child-friendly tabletop card game about training data. Based on three research-based design principles: low-stakes experimentation to support curiosity, games facilitating conversation, and tangible and embodied learning for abstract concepts, the game supports learning the high-level mechanics of training data in supervised machine learning, as well as practicing critical discussion of training data related to algorithmic bias. Contributing to AI literacy opportunities, this game aims to facilitate playful peer-peer and child-parent learning.
用有形的时钟体现时间的学习
DOI：
--
发表时间：
2017
期刊：
Conference on Interactive Mobile Communication Technologies and Learning
影响因子：
--
作者：
G. Palaigeorgiou;Dimitra Tsapkini;T. Bratitsis;Stefanos Xefteris
通讯作者：
Stefanos Xefteris