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 2601.21130cs.AI
收藏

你感受的并非他们所见:关于从第三方观察者标签预测自我报告情绪的研究

What You Feel Is Not What They See: On Predicting Self-Reported Emotion from Third-Party Observer Labels

University of Michigan, Ann Arbor, Michigan, USA(密歇根大学)

Yara El-Tawil, Aneesha Sampath, Emily Mower Provost

AI总结:

研究探讨了从第三方观察者标签预测自我报告情绪的可行性,发现个人重要性显著提升模型的情感预测性能。

摘要:

自我报告的情绪标签捕捉内部体验,而第三方标签反映外部感知。这些视角往往存在分歧,限制了第三方训练模型在自我报告情境中的应用。这一差距在心理健康领域尤为重要,因为准确的自我报告建模对于指导干预至关重要。我们提出了首个跨语料库评估第三方训练模型在自我报告上的表现。我们发现激活不可预测(CCC大约为0)而情感中等可预测(CCC大约为0.3)。关键的是,当内容对说话者个人具有重要意义时,模型在情感上的表现显著提高(CCC大约为0.6-0.8)。我们的发现表明,个人重要性是连接外部感知与内部体验的关键路径,并突显了自我报告激活建模的挑战。

英文摘要:

Self-reported emotion labels capture internal experience, while third-party labels reflect external perception. These perspectives often diverge, limiting the applicability of third-party-trained models to self-report contexts. This gap is critical in mental health, where accurate self-report modeling is essential for guiding intervention. We present the first cross-corpus evaluation of third-party-trained models on self-reports. We find activation unpredictable (CCC approximately 0) and valence moderately predictable (CCC approximately 0.3). Crucially, when content is personally significant to the speaker, models achieve high performance for valence (CCC approximately 0.6-0.8). Our findings point to personal significance as a key pathway for aligning external perception with internal experience and underscore the challenge of self-report activation modeling.

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。

↑