本頁只刊出中文翻譯與中文說明；英文原文請見下方原文連結。
原文連結
論文資訊
- 類型：已發表論文
- 日期：2001
摘要
We show that the way in which the Shannon 熵 of sequences produced by an 資訊 source converges to the source's 熵 rate can be used to monitor how an intelligent 智能體 builds and effectively uses a predictive model of its environment. We introduce natural measures of the environment's apparent memory and the amounts of 資訊 that must be (i) extracted from observations for an 智能體 to synchronize to the environment and (ii) stored by an 智能體 for optimal prediction. If structural properties are ignored, the missed regularities are converted to apparent randomness. Conversely, using representations that assume too much memory results in false predictability.
※ 此為已發表論文，全文需透過期刊付費取得