Jungjun Kim

Jungjun Kim, Changjin Han, Gyuhyeon Nam et al.

Most Chinese Grapheme-to-Phoneme (G2P) systems employ a three-stage framework that first transforms input sequences into character embeddings, obtains linguistic information using language models, and then predicts the phonemes based on global context about the entire input sequence. However, linguistic knowledge alone is often inadequate. Language models frequently encode overly general structures of a sentence and fail to cover specific cases needed to use phonetic knowledge. Also, a handcrafted post-processing system is needed to address the problems relevant to the tone of the characters. However, the system exhibits inconsistency in the segmentation of word boundaries which consequently degrades the performance of the G2P system. To address these issues, we propose the Reinforcer that provides strong inductive bias for language models by emphasizing the phonological information between neighboring characters to help disambiguate pronunciations. Experimental results show that the Reinforcer boosts the cutting-edge architectures by a large margin. We also combine the Reinforcer with a large-scale pre-trained model and demonstrate the validity of using neighboring context in knowledge transfer scenarios.

Seokgi Lee, Jungjun Kim

We present the gradual style adaptor TTS (GSA-TTS) with a novel style encoder that gradually encodes speaking styles from an acoustic reference for zero-shot speech synthesis. GSA first captures the local style of each semantic sound unit. Then the local styles are combined by self-attention to obtain a global style condition. This semantic and hierarchical encoding strategy provides a robust and rich style representation for an acoustic model. We test GSA-TTS on unseen speakers and obtain promising results regarding naturalness, speaker similarity, and intelligibility. Additionally, we explore the potential of GSA in terms of interpretability and controllability, which stems from its hierarchical structure.

Jung-Jun Kim, Dong-Gyu Lee, Jialin Wu et al.

Visual question answering requires a deep understanding of both images and natural language. However, most methods mainly focus on visual concept; such as the relationships between various objects. The limited use of object categories combined with their relationships or simple question embedding is insufficient for representing complex scenes and explaining decisions. To address this limitation, we propose the use of text expressions generated for images, because such expressions have few structural constraints and can provide richer descriptions of images. The generated expressions can be incorporated with visual features and question embedding to obtain the question-relevant answer. A joint-embedding multi-head attention network is also proposed to model three different information modalities with co-attention. We quantitatively and qualitatively evaluated the proposed method on the VQA v2 dataset and compared it with state-of-the-art methods in terms of answer prediction. The quality of the generated expressions was also evaluated on the RefCOCO, RefCOCO+, and RefCOCOg datasets. Experimental results demonstrate the effectiveness of the proposed method and reveal that it outperformed all of the competing methods in terms of both quantitative and qualitative results.