SEEN: 以結構化事件增強網路偵測與解釋資訊召回需求

Abstract

在回憶生活經歷時，人們經常忘記或混淆生活事件，所以提供資訊召回的服 務是需要的。而以前關於資訊召回的研究主要是被動式提供，也就是使用者透過 給定生活事件來評估是否需要資訊召回服務。然而，很少有研究涉及由系統主動 偵測人們是否需要資訊召回服務。在本文中，我們透過比較同一作者在兩個不同 時間點、針對同一事件所寫的敘述，來確定用戶在描述他們的過往的生活經歷時 是否遇到困難。因此，我們使用標記者根據個人真實生活經歷組成的資料集來偵 測觸發資訊召回服務的正確時間。此外，我們也提出一個模型–結構化事件增強網 路(SEEN)，它可以檢測到標記者撰寫的生活經歷是否包含不一致、額外新增或 是被遺忘的生活事件。而此模型中還包含我們提出的一種特殊機制，我們透過這 種機制來融合以生活事件為基礎所構建的無向圖和語言模型所產生的文字嵌入向 量。同時，為了進一步提供具解釋性的服務，我們的模型會從生活事件的無向圖 中選擇相關的節點用以當作參考事件。而實驗結果也表明，我們的模型在偵測資 訊召回需求的任務 取得了很好的成果，提取出的參考事件也可以有效作為補充資 訊，提醒用戶他們可能想要召回的生活事件。

When recalling life experiences, people often forget or confuse life events, which necessitates information recall services. Previous work on information recall focuses on providing such assistance reactively, i.e., by retrieving the life event of a given query. What is rarely discussed, however, is a proactive system that is capable of detecting the need for information recall services. In this paper, we propose determining whether users are experiencing difficulty in recalling their life experiences by comparing the events described in two retold stories written at different times. We use a human-annotated life experience retelling dataset to detect the right time to trigger the information recall service. We also propose a pilot model–Structured Event Enhancement Network (SEEN) that detects life event inconsistency, additional information in life events, and forgotten events. A fusing mechanism is also proposed to incorporate event graphs of stories and enhance the textual representations. To explain the need detection results, SEEN simultaneously provides support evidence by selecting the related nodes from the event graph. Experimental results show that SEEN achieves promising performance in detecting information needs. In addition, the extracted evidence can be served as complementary information to remind users what events they may want to recall.