arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2202.05983cs.AIcs.CVcs.HCcs.LG
收藏

未校准的模型可以改善人类与AI的协作

Uncalibrated Models Can Improve Human-AI Collaboration

Stanford University(斯坦福大学)

Kailas Vodrahalli, Tobias Gerstenberg, James Zou

摘要:

在人工智能的许多实际应用中,人工智能模型被用作人类用户的决策辅助。人工智能提供建议,人类(有时)将其纳入决策过程。人工智能的建议通常带有某种程度的“信心”,人类可以用它来校准他们对建议的依赖或信任程度。在本文中,我们提出了一个初步的探索,建议显示AI模型比实际更有信心,即使原始AI经过良好校准,也可以提高人类AI性能(衡量为人类在看到AI建议后最终预测的准确性和置信度)。我们首先训练一个模型,使用来自数千次人类与人工智能交互的数据来预测人类对人工智能建议的结合。这使我们能够明确估计如何转换AI的预测置信度,使AI未经校准,以改善最终的人类预测。我们在四个不同的任务中验证了我们的结果-处理图像,文本和表格数据-涉及数百名人类参与者。我们进一步支持我们的研究结果与模拟分析。我们的研究结果表明,联合优化人类-AI系统的重要性,而不是单独优化AI模型的标准范式。

英文摘要:

In many practical applications of AI, an AI model is used as a decision aid for human users. The AI provides advice that a human (sometimes) incorporates into their decision-making process. The AI advice is often presented with some measure of "confidence" that the human can use to calibrate how much they depend on or trust the advice. In this paper, we present an initial exploration that suggests showing AI models as more confident than they actually are, even when the original AI is well-calibrated, can improve human-AI performance (measured as the accuracy and confidence of the human's final prediction after seeing the AI advice). We first train a model to predict human incorporation of AI advice using data from thousands of human-AI interactions. This enables us to explicitly estimate how to transform the AI's prediction confidence, making the AI uncalibrated, in order to improve the final human prediction. We empirically validate our results across four different tasks--dealing with images, text and tabular data--involving hundreds of human participants. We further support our findings with simulation analysis. Our findings suggest the importance of jointly optimizing the human-AI system as opposed to the standard paradigm of optimizing the AI model alone.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑