Weiyan Shi
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Weiyan Shi has a total of 34 co-authors. Here are the top co-authors by number of shared publications:
- Kenny Tsu Wei Choo
Singapore University of Technology and Design
8 shared publications - Daqing Zhang
Peking University
3 shared publications - Roy Ka-Wei Lee
Singapore University of Technology and Design
3 shared publications - Simon Tangi Perrault
LTCI, Telecom Paris, IP Paris, Palaiseau, France
3 shared publications - Bryan Chen Zhengyu Tan
Singapore University of Technology and Design
2 shared publications - Gionnieve Lim
Singapore University of Technology and Design, Singapore, Singapore
2 shared publications
Latest Publications
- Open Access
- 23 September 2026
- Published by Association for Computing Machinery (ACM) in Proceedings of the ACM on Human-Computer Interaction
- Vol. 10 (6) , 1-39
Sword and Shield: Uses and Strategies of LLMs in Navigating Disinformation
The emergence of Large Language Models (LLMs) presents a dual challenge in the fight against disinformation. These powerful tools, capable of generating human-like text at scale, can be weaponised to produce sophisticated disinformation, yet they also hold promise for enhancing mitigation strategies. This paper investigates the complex dynamics between LLMs and disinformation in small, localised settings through a communication game based on online forums, inspired by Werewolf , with 25 participants. We analyse how Disinformers, Moderators, and Users leverage LLMs to advance their goals, revealing both the potential for misuse and combating disinformation. Our findings highlight the varying uses of LLMs depending on the participants’ roles and strategies, underscoring the importance of understanding their effectiveness in this context. We conclude by discussing implications for future LLM development and online platform design, advocating for a balanced approach that empowers users and fosters trust while mitigating the risks of LLM-assisted disinformation.
- 23 September 2026
- Published by Association for Computing Machinery (ACM) in Proceedings of the ACM on Human-Computer Interaction
- Vol. 10 (6) , 1-39
Open AccessSword and Shield: Uses and Strategies of LLMs in Navigating Disinformation
The emergence of Large Language Models (LLMs) presents a dual challenge in the fight against disinformation. These powerful tools, capable of generating human-like text at scale, can be weaponised to produce sophisticated disinformation, yet they also hold promise for enhancing mitigation strategies. This paper investigates the complex dynamics between LLMs and disinformation in small, localised settings through a communication game based on online forums, inspired by Werewolf , with 25 participants. We analyse how Disinformers, Moderators, and Users leverage LLMs to advance their goals, revealing both the potential for misuse and combating disinformation. Our findings highlight the varying uses of LLMs depending on the participants’ roles and strategies, underscoring the importance of understanding their effectiveness in this context. We conclude by discussing implications for future LLM development and online platform design, advocating for a balanced approach that empowers users and fosters trust while mitigating the risks of LLM-assisted disinformation.
- Open Access
- 12 June 2026
- Published by Association for Computing Machinery (ACM)
- p. 445-449
When Drawing Is Not Enough: Exploring Spontaneous Speech with Sketch for Intent Alignment in Multimodal LLMs
- 12 June 2026
- Published by Association for Computing Machinery (ACM)
- p. 445-449
Open AccessWhen Drawing Is Not Enough: Exploring Spontaneous Speech with Sketch for Intent Alignment in Multimodal LLMs
- Open Access
- 13 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 1-17
Towards Aligning Multimodal LLMs with Human Experts: A Focus on Parent–Child Interaction
- 13 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 1-17
Open AccessTowards Aligning Multimodal LLMs with Human Experts: A Focus on Parent–Child Interaction
- Open Access
- 13 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 1-5
A Taxonomy of Human–MLLM Interaction in Early-Stage Sketch-Based Design Ideation
- 13 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 1-5
Open AccessA Taxonomy of Human–MLLM Interaction in Early-Stage Sketch-Based Design Ideation
- 27 February 2026
- Published by Springer Nature
- p. 83-97
TalkSketch: Multimodal Generative AI for Real-Time Sketch Ideation with Speech
- 27 February 2026
- Published by Springer Nature
- p. 83-97
TalkSketch: Multimodal Generative AI for Real-Time Sketch Ideation with Speech
- Open Access
- 17 October 2025
- Published by Association for Computing Machinery (ACM)
- p. 264-269
Communication Games to Study Emerging Disinformation Scenarios With LLMs
- 17 October 2025
- Published by Association for Computing Machinery (ACM)
- p. 264-269
Open AccessCommunication Games to Study Emerging Disinformation Scenarios With LLMs
- Open Access
- 14 August 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Data Descriptions
- Vol. 2, 286-295
Descriptor: Coastal Aerial Imagery Dataset for Shoreline Segmentation (CAID)
Shoreline segmentation is a specialized water-land segmentation task that detects the position of the shoreline. Existing datasets for segmentation either do not focus on coastal environments (e.g., DeepGlobe) or are not based on high-resolution imagery (e.g., Sen1Floods11). To fill this gap, we built Coastal Aerial Imagery Dataset (CAID), a high-resolution, manually labeled benchmark dataset built on aerial imagery for shoreline segmentation. CAID consists of 20 689 images of 500 $\times$ 500 pixels across 850 National Agriculture Imagery Program (NAIP) sites across the Great Lakes region. The dataset covers a variety of coastal landscapes, and its representativeness of coastal environments was evaluated using three criteria: 1) water body area; 2) shoreline length ratio; and 3) mean water hue. We also assessed the benchmark performance of existing segmentation models on CAID. While these models achieved high mean intersection over union (mIoU) and mean accuracy (mAcc) scores, accurately positioning shorelines remains a challenge, a limitation that is not fully reflected by existing evaluation metrics. To address this, we introduced mean shoreline covering ratio (mSCR), a novel metric that quantifies each model’s ability to align with ground-truth shorelines. The CAID dataset, along with the associated code for training and evaluating deep learning models, will be publicly released under the CC-BY-4.0 license via GitHub and Zenodo to ensure open-access availability. IEEE SOCIETY/COUNCIL Geoscience and Remote Sensing Society DATA TYPE/LOCATION Aerial Imagery, Segmentation; The Great Lakes region, USA DATA DOI/PID 10.5281/zenodo.15330911
- 14 August 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Data Descriptions
- Vol. 2, 286-295
Open AccessDescriptor: Coastal Aerial Imagery Dataset for Shoreline Segmentation (CAID)
Shoreline segmentation is a specialized water-land segmentation task that detects the position of the shoreline. Existing datasets for segmentation either do not focus on coastal environments (e.g., DeepGlobe) or are not based on high-resolution imagery (e.g., Sen1Floods11). To fill this gap, we built Coastal Aerial Imagery Dataset (CAID), a high-resolution, manually labeled benchmark dataset built on aerial imagery for shoreline segmentation. CAID consists of 20 689 images of 500 $\times$ 500 pixels across 850 National Agriculture Imagery Program (NAIP) sites across the Great Lakes region. The dataset covers a variety of coastal landscapes, and its representativeness of coastal environments was evaluated using three criteria: 1) water body area; 2) shoreline length ratio; and 3) mean water hue. We also assessed the benchmark performance of existing segmentation models on CAID. While these models achieved high mean intersection over union (mIoU) and mean accuracy (mAcc) scores, accurately positioning shorelines remains a challenge, a limitation that is not fully reflected by existing evaluation metrics. To address this, we introduced mean shoreline covering ratio (mSCR), a novel metric that quantifies each model’s ability to align with ground-truth shorelines. The CAID dataset, along with the associated code for training and evaluating deep learning models, will be publicly released under the CC-BY-4.0 license via GitHub and Zenodo to ensure open-access availability. IEEE SOCIETY/COUNCIL Geoscience and Remote Sensing Society DATA TYPE/LOCATION Aerial Imagery, Segmentation; The Great Lakes region, USA DATA DOI/PID 10.5281/zenodo.15330911
- Open Access
Sword and Shield: Uses and Strategies of LLMs in Navigating Disinformation
The emergence of Large Language Models (LLMs) presents a dual challenge in the fight against disinformation. These powerful tools, capable of generating human-like text at scale, can be weaponised to produce sophisticated and persuasive disinformation, yet they also hold promise for enhancing detection and mitigation strategies. This paper investigates the complex dynamics between LLMs and disinformation through a communication game that simulates online forums, inspired by the game Werewolf, with 25 participants. We analyse how Disinformers, Moderators, and Users leverage LLMs to advance their goals, revealing both the potential for misuse and combating disinformation. Our findings highlight the varying uses of LLMs depending on the participants' roles and strategies, underscoring the importance of understanding their effectiveness in this context. We conclude by discussing implications for future LLM development and online platform design, advocating for a balanced approach that empowers users and fosters trust while mitigating the risks of LLM-assisted disinformation.
Open AccessSword and Shield: Uses and Strategies of LLMs in Navigating Disinformation
The emergence of Large Language Models (LLMs) presents a dual challenge in the fight against disinformation. These powerful tools, capable of generating human-like text at scale, can be weaponised to produce sophisticated and persuasive disinformation, yet they also hold promise for enhancing detection and mitigation strategies. This paper investigates the complex dynamics between LLMs and disinformation through a communication game that simulates online forums, inspired by the game Werewolf, with 25 participants. We analyse how Disinformers, Moderators, and Users leverage LLMs to advance their goals, revealing both the potential for misuse and combating disinformation. Our findings highlight the varying uses of LLMs depending on the participants' roles and strategies, underscoring the importance of understanding their effectiveness in this context. We conclude by discussing implications for future LLM development and online platform design, advocating for a balanced approach that empowers users and fosters trust while mitigating the risks of LLM-assisted disinformation.
- Open Access
- 25 April 2025
- Published by Association for Computing Machinery (ACM)
- p. 1-6
Towards Multimodal Large-Language Models for Parent-Child Interaction: A Focus on Joint Attention
- 25 April 2025
- Published by Association for Computing Machinery (ACM)
- p. 1-6
Open AccessTowards Multimodal Large-Language Models for Parent-Child Interaction: A Focus on Joint Attention
- 4 November 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Mobile Computing
- Vol. 24 (3) , 1895-1912
SigCan: Toward Reliable ToF Estimation Leveraging Multipath Signal Cancellation on Commodity WiFi Devices
The widespread deployment of WiFiinfrastructure has facilitated the development of Time-of-Flight (ToF) based sensing applications. ToF estimation, however, is a challenging task due to the complexity of multipath effect. In this paper, we propose a phase difference based method for ToF estimation and uncover the potential of signal cancellation to mitigate the impact of multipath and noise on phase differences among subcarriers. To separate the moving target path from the complex multipath for ToF estimation, we suggest employing specific elimination methods tailored to the characteristics of different signal components. For dynamic multipath, we observe that when a given subcarrier propagates along two paths to the receiver, with path lengths differing by half a wavelength, the phase difference introduced by these two paths cancels each other out. Therefore, we propose two metrics to identify signals that satisfy this condition, utilizing both frequency diversity and spatial diversity. Additionally, we propose leveraging time diversity to eliminate the static multipath component and reduce the impact of noise. We implemented the methods with off-the-shelf WiFidevices and achieved mean errors of 15.36cm and 21.05cm for distance estimation in outdoor and indoor scenarios, outperforming state-of-the-art ToF estimation method by 50% error reduction.
- 4 November 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Mobile Computing
- Vol. 24 (3) , 1895-1912
SigCan: Toward Reliable ToF Estimation Leveraging Multipath Signal Cancellation on Commodity WiFi Devices
The widespread deployment of WiFiinfrastructure has facilitated the development of Time-of-Flight (ToF) based sensing applications. ToF estimation, however, is a challenging task due to the complexity of multipath effect. In this paper, we propose a phase difference based method for ToF estimation and uncover the potential of signal cancellation to mitigate the impact of multipath and noise on phase differences among subcarriers. To separate the moving target path from the complex multipath for ToF estimation, we suggest employing specific elimination methods tailored to the characteristics of different signal components. For dynamic multipath, we observe that when a given subcarrier propagates along two paths to the receiver, with path lengths differing by half a wavelength, the phase difference introduced by these two paths cancels each other out. Therefore, we propose two metrics to identify signals that satisfy this condition, utilizing both frequency diversity and spatial diversity. Additionally, we propose leveraging time diversity to eliminate the static multipath component and reduce the impact of noise. We implemented the methods with off-the-shelf WiFidevices and achieved mean errors of 15.36cm and 21.05cm for distance estimation in outdoor and indoor scenarios, outperforming state-of-the-art ToF estimation method by 50% error reduction.