Chuang Yu

Chuang Yu

Auto-generated profile, based on publicly available information

Chuang Yu

UCL Interaction Centre, University College London, London, United Kingdom

Research Focus & Metrics

Academic Career Length

11Years (2016-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
26 September 2023 in IEEE Signal Processing Letters
Verified
Journal Article
Open access
YUANBO HOUYUANBO HOU
YUANBO HOU
SSSiyang Song
Siyang Song
CYChuang Yu
Chuang Yu
WWWenwu Wang
Wenwu Wang
DBDick Botteldooren
Dick Botteldooren
Most deep learning-based acoustic scene classification (ASC) approaches identify scenes based on acoustic features converted from audio clips containing mixed information entangled by polyphonic audio events (AEs). However, these approaches have difficulties in explaining what cues they use to identify scenes. This paper conducts the first study on disclosing the relationship between real-life acoustic scenes and semantic embeddings from the most relevant AEs. Specifically, we propose an event-relational graph representation learning (ERGL) framework for ASC to classify scenes, and simultaneously answer clearly and straightly which cues are used in classifying. In the event-relational graph, embeddings of each event are treated as nodes, while relationship cues derived from each pair of nodes are described by multi-dimensional edge features. Experiments on a real-life ASC dataset show that the proposed ERGL achieves competitive performance on ASC by learning embeddings of only a limited number of AEs. The results show the feasibility of recognizing diverse acoustic scenes based on the audio event-relational graph. Visualizations of graph representations learned by ERGL are available here (https://github.com/Yuanbo2020/ERGL) .

7 March 2022
Verified
Conference Paper
Closed access
CYChuang Yu
Chuang Yu
CFChangzeng Fu
Changzeng Fu
RCRui Chen
Rui Chen
Adriana TapusAdriana Tapus
Adriana Tapus
Some robots for human-robot interaction are designed with female or male physical appearance. Other robots are endowed with no gender characteristics, namely genderless robots, such as Pepper and NAO robot. A robot with male or female physical appearance should possess the mapped speech gender style during a natural human-robot interaction, which can be learned from humans' male or female speech. In this paper, we make a new trial to synthesis gender-free speeches for physically genderless robots, which is promising in order to improve a more natural human-robot interaction with genderless robots. Our gender style-controlled speech synthesizer takes the speech text and gender style embedding as inputs to generate speech audio. A speech gender encoder network is used to extract the embedding of the speech gender style with female and male speeches as input. Based on the distribution of the female and male gender style embedding, we explore the gender-free speech style embedding space where we sample some gender-free embedding vectors to generate genderless speech audio. This is a preliminary work where we show how the genderless speech audio wave will be synthesized from text.

11 March 2024
Verified
Conference Paper
Open access
XHXiaoxuan Hei
Xiaoxuan Hei
CYChuang Yu
Chuang Yu
HZHeng Zhang
Heng Zhang
Adriana TapusAdriana Tapus
Adriana Tapus
In situations where both deaf and non-deaf individuals are present in a public setting, it would be advantageous for a robot to communicate using both sign and natural languages simultaneously. This would not only address the needs for diverse users but also pave the way for a richer and more inclusive spectrum of human-robot interactions. To achieve this, a framework for a bilingual robot has been proposed in this paper. The robot exhibits the ability to articulate messages in spoken language, complemented by non-verbal cues such as expressive gestures, all while concurrently conveying information through sign language. The system can generate natural language expressions with speech audio, spontaneous prosody-based gestures, and sign language displayed on a virtual avatar on a robot's screen. The preliminary findings from this research showcase the robot's capacity to seamlessly blend natural language expressions with synchronized gestures and sign language, underlining its potential to revolutionize communication dynamics in diverse settings.

13 March 2023
Verified
Conference Paper
Open access
LLLanmiao Liu
Lanmiao Liu
CYChuang Yu
Chuang Yu
SSSiyang Song
Siyang Song
ZSZhidong Su
Zhidong Su
Adriana TapusAdriana Tapus
Adriana Tapus
Human skeleton-based gesture classification plays a dominant role in social robotics. Learning the variety of human skeleton-based gestures can help the robot to continuously interact in an appropriate manner in a natural human-robot interaction (HRI). In this paper, we proposed a Flow-based model to classify human gesture actions with skeletal data. Instead of inferring new human skeleton actions from noisy data using a retrained model, our end-to-end model can expand the diversity of labels for gesture recognition from noisy data without retraining the model. At first, our model focuses on detecting five human gesture actions (i.e., come on, right up, left up, hug, and noise-random action). The accuracy of our online human gesture recognition system is as well as the offline one. Meanwhile, both attain 100% accuracy among the first four actions. Our proposed method is more efficient for inference of new human gesture action without retraining, which acquires about 90% accuracy for noise-random action. The gesture recognition system has been applied to the robot's reaction toward the human gesture, which is promising to facilitate a natural human-robot interaction.

11 March 2024
Verified
Conference Paper
Open access
HZHeng Zhang
Heng Zhang
CYChuang Yu
Chuang Yu
XHXiaoxuan Hei
Xiaoxuan Hei
Adriana TapusAdriana Tapus
Adriana Tapus
Humor is pervasive in our daily life. It serves not only to build rapport but also to ease tension during interactions and create stronger social interactions. In the realm of Human-Robot Interaction (HRI), humor also plays a vital role in fostering engaging and positive interactions. However, endowing robots with the ability to express humor appropriately is still a challenge. Drawing inspiration from pantomime and sign language humor, our research focuses on the role of semantic gestures in the social robot's expression of humor. In this work, we conducted an experiment in which NAO robot made humorous performances by using a series of semantic gestures. The results of the online survey show that the semantic gesture can significantly enhance the degree of funniness of the robot humor performance. Furthermore, the impact of the semantic gesture is closely tied to both the clarity of its expression and the appropriateness of the chosen semantic words.

13 March 2023
Verified
Conference Paper
Open access
CYChuang Yu
Chuang Yu
BSBaris Serhan
Baris Serhan
MRMarta Romeo
Marta Romeo
Angelo CangelosiAngelo Cangelosi
Angelo Cangelosi