Enhancing Trust in Intelligent Virtual Agents: The Role of Acoustic Cues, Lexical Style, and Listener Characteristics in the Perception of Synthesized Speech

Enhancing Trust in Intelligent Virtual Agents: The Role of Acoustic Cues, Lexical Style, and Listener Characteristics in the Perception of Synthesized Speech This repository supports the research presented in our paper, which has been accepted for publication (details forthcoming). Speech stimuli Question-based speech clips are saved under these three folders: question_humor_audio, question_neutral_audio, question_polite_audio. Recommendation-based speech clips are saved under these three folders: recommendation_humor_audio, recommendation_neutral_audio, recommendation_polite_audio. Note: The audio name: q10_F_saved_audio_files_wav corresponds to text 10, female speaker Code temp_generate_combinations.py and aws_handler.py are main codes to generate speech stimuli Collected ratings collected_ratings includes all collected ratings from 6 experiments. csv file name started with 'processed_df_' has the average of trustworthy ratings from unique 5 raters for each audio clip. csv file name started with 'speaker' has the rater info. If you have any questions about data or code, please reach out to [e-mail]