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 2206.04751cs.CL
收藏

为涌现语言中的组合性辩护

Defending Compositionality in Emergent Languages

Charles University(查理大学);Faculty of Mathematics and Physics(数学与物理学院)

Michal Auersperger, Pavel Pecina

摘要:

传统上,组合性被理解为语言生产力的一个主要因素,更广泛地说,是人类认知的一个主要因素。然而,最近,一些研究开始质疑它的地位,表明人工神经网络即使没有明显的组合行为也擅长泛化。我们认为,其中一些结论是过于强大和/或不完整的。在一个两个智能体通信游戏的背景下,我们表明,组合性确实似乎是成功的推广时,一个适当的数据集进行评估是必不可少的。

英文摘要:

Compositionality has traditionally been understood as a major factor in productivity of language and, more broadly, human cognition. Yet, recently, some research started to question its status, showing that artificial neural networks are good at generalization even without noticeable compositional behavior. We argue that some of these conclusions are too strong and/or incomplete. In the context of a two-agent communication game, we show that compositionality indeed seems essential for successful generalization when the evaluation is done on a proper dataset.

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。

↑