We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

eess.AS

Change to browse by:

References & Citations

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Electrical Engineering and Systems Science > Audio and Speech Processing

Title: A Study on the Ambiguity in Human Annotation of German Oral History Interviews for Perceived Emotion Recognition and Sentiment Analysis

Abstract: For research in audiovisual interview archives often it is not only of interest what is said but also how. Sentiment analysis and emotion recognition can help capture, categorize and make these different facets searchable. In particular, for oral history archives, such indexing technologies can be of great interest. These technologies can help understand the role of emotions in historical remembering. However, humans often perceive sentiments and emotions ambiguously and subjectively. Moreover, oral history interviews have multi-layered levels of complex, sometimes contradictory, sometimes very subtle facets of emotions. Therefore, the question arises of the chance machines and humans have capturing and assigning these into predefined categories. This paper investigates the ambiguity in human perception of emotions and sentiment in German oral history interviews and the impact on machine learning systems. Our experiments reveal substantial differences in human perception for different emotions. Furthermore, we report from ongoing machine learning experiments with different modalities. We show that the human perceptual ambiguity and other challenges, such as class imbalance and lack of training data, currently limit the opportunities of these technologies for oral history archives. Nonetheless, our work uncovers promising observations and possibilities for further research.
Comments: Submitted to LREC 2022
Subjects: Audio and Speech Processing (eess.AS); Computation and Language (cs.CL); Sound (cs.SD)
Cite as: arXiv:2201.06868 [eess.AS]
  (or arXiv:2201.06868v1 [eess.AS] for this version)

Submission history

From: Michael Gref [view email]
[v1] Tue, 18 Jan 2022 10:53:07 GMT (471kb,D)

Link back to: arXiv, form interface, contact.