We present a method to control the emotional prosody of Text to Speech (TTS) systems by using phoneme-level intermediate features (pitch, energy, and duration) as levers. As a key idea, we propose Differential Scaling (DS) to disentangle features relating to affective prosody from those arising due to acoustics conditions and speaker identity. With thorough experimental studies, we show that the proposed method improves over the prior art in accurately emulating the desired emotions while retaining the naturalness of speech. We extend the traditional evaluation of using individual sentences for a more complete evaluation of HCI systems. We present a novel experimental setup by replacing an actor with a TTS system in offline and live conversations. The emotion to be rendered is either predicted or manually assigned. The results show that the proposed method is strongly preferred over the state-of-the-art TTS system and adds the much-coveted “human touch” in machine dialogue. Audio samples from our experiments and the code are available at: https://emtts.github.io/tts-demo/
Investigating brand perception is fundamental to marketing strategies. In this regard, brand image, defined by a set of attributes (Aaker, 1997), is recognized as a key element in indicating how a brand is perceived by various stakeholders such as consumers and competitors. Traditional approaches (e.g., surveys) to monitor brand perceptions are time-consuming and inefficient. In the era of digital marketing, both brand managers and consumers engage with a vast amount of digital marketing content. The exponential growth of digital content has propelled the emergence of pre-trained language models such as BERT and GPT as essential tools in solving myriads of challenges with textual data. This paper seeks to investigate the extent of brand perceptions (i.e., brand and image attribute associations) these language models encode. We believe that any kind of bias for a brand and attribute pair may influence customer-centric downstream tasks such as recommender systems, sentiment analysis, and question-answering, e.g., suggesting a specific brand consistently when queried for innovative products. We use synthetic data and real-life data and report comparison results for five contextual LMs, viz. BERT, RoBERTa, DistilBERT, ALBERT and BART.
This paper presents our system submitted to the EmotionX challenge. It is an emotion detection task on dialogues in the EmotionLines dataset. We formulate this as a hierarchical network where network learns data representation at both utterance level and dialogue level. Our model is inspired by Hierarchical Attention network (HAN) and uses pre-trained word embeddings as features. We formulate emotion detection in dialogues as a sequence labeling problem to capture the dependencies among labels. We report the performance accuracy for four emotions (anger, joy, neutral and sadness). The model achieved unweighted accuracy of 55.38% on Friends test dataset and 56.73% on EmotionPush test dataset. We report an improvement of 22.51% in Friends dataset and 36.04% in EmotionPush dataset over baseline results.