arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2210.09972cs.CLcs.ITcs.LGmath.IT
收藏

论词类比测验中预测项的信息量

On the Information Content of Predictions in Word Analogy Tests

Federal University of Sergipe(塞尔希培联邦大学)

Jugurta Montalvão

摘要:

提出了一种方法来量化,在比特的信息,类比测试中的类比的实际相关性。这种方法的主要组成部分是一个softaccuracy估计,也产生熵估计与补偿偏差。使用预训练的GloVe 300-D向量和两个公共类比测试集获得的实验结果表明,从信息内容的角度来看,在类比测试中,接近提示比类比更相关。因此,一个简单的词嵌入模型被用来预测,类比携带约一位的信息,这是实验证实。

英文摘要:

An approach is proposed to quantify, in bits of information, the actual relevance of analogies in analogy tests. The main component of this approach is a softaccuracy estimator that also yields entropy estimates with compensated biases. Experimental results obtained with pre-trained GloVe 300-D vectors and two public analogy test sets show that proximity hints are much more relevant than analogies in analogy tests, from an information content perspective. Accordingly, a simple word embedding model is used to predict that analogies carry about one bit of information, which is experimentally corroborated.

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。

↑