Yuta Hagio, Makoto Okuda, Marina Kamimura, Yutaka Kaneko, H. Ohmata
This study presents a method for generating utterances for companion robots that watch TV with people, using TV program subtitles. To enable the robot to automatically generate relevant utterances while watching TV, we created a dataset of approximately 12,000 utterances that were manually added to the collected TV subtitles. Using this dataset, we fine-tuned a large-scale language model to construct an utterance generation model. The proposed model generates utterances based on multiple keywords extracted from the subtitles as topics, while also taking into account the context of the subtitles by inputting them. The evaluation of the generated utterances revealed that approximately 88% of the sentences were natural Japanese, and approximately 75% were relevant and natural in the context of the TV program. Moreover, approximately 99% of the sentences contained the extracted keywords, indicating that our proposed method can generate diverse and contextually appropriate utterances containing the targeted topics. These findings provide evidence of the effectiveness of our approach in generating natural utterances for companion robots that watch TV with people.
{"title":"Generating Utterances for Companion Robots using Television Program Subtitles","authors":"Yuta Hagio, Makoto Okuda, Marina Kamimura, Yutaka Kaneko, H. Ohmata","doi":"10.1145/3573381.3596463","DOIUrl":"https://doi.org/10.1145/3573381.3596463","url":null,"abstract":"This study presents a method for generating utterances for companion robots that watch TV with people, using TV program subtitles. To enable the robot to automatically generate relevant utterances while watching TV, we created a dataset of approximately 12,000 utterances that were manually added to the collected TV subtitles. Using this dataset, we fine-tuned a large-scale language model to construct an utterance generation model. The proposed model generates utterances based on multiple keywords extracted from the subtitles as topics, while also taking into account the context of the subtitles by inputting them. The evaluation of the generated utterances revealed that approximately 88% of the sentences were natural Japanese, and approximately 75% were relevant and natural in the context of the TV program. Moreover, approximately 99% of the sentences contained the extracted keywords, indicating that our proposed method can generate diverse and contextually appropriate utterances containing the targeted topics. These findings provide evidence of the effectiveness of our approach in generating natural utterances for companion robots that watch TV with people.","PeriodicalId":120872,"journal":{"name":"Proceedings of the 2023 ACM International Conference on Interactive Media Experiences","volume":"64 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131730428","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Minas Katsiokalis, Elisavet Tsekeri, Aikaterini Lilli, K. Gobakis, D. Kolokotsa, K. Mania
As Extended Reality (XR) media experiences become a commodity, there is a unique opportunity to deploy XR for environmental awareness. Interaction challenges of Augmented Reality (AR) still exist, focused on limited gesture and head tracking affordances. AR technologies should also be seamlessly integrated with sensor data, analytics and ultimately status prediction, to be visualized in an AR experience, rather than merely superimposing visuals onto the real world. This paper presents an innovative, work-in-progress, multimodal AR experience integrating interactive narration, gestures, hands recognition and voice commands, while a citizen is wearing a head-worn AR display, promoting environmental awareness, health, and wellness. By combining AR technologies and a sensor network, GoNature AR provides citizen awareness of real-time, multimodal, air and noise pollution data.
{"title":"GoNature AR: Air Quality & Noise Visualization Through a Multimodal and Interactive Augmented Reality Experience","authors":"Minas Katsiokalis, Elisavet Tsekeri, Aikaterini Lilli, K. Gobakis, D. Kolokotsa, K. Mania","doi":"10.1145/3573381.3597229","DOIUrl":"https://doi.org/10.1145/3573381.3597229","url":null,"abstract":"As Extended Reality (XR) media experiences become a commodity, there is a unique opportunity to deploy XR for environmental awareness. Interaction challenges of Augmented Reality (AR) still exist, focused on limited gesture and head tracking affordances. AR technologies should also be seamlessly integrated with sensor data, analytics and ultimately status prediction, to be visualized in an AR experience, rather than merely superimposing visuals onto the real world. This paper presents an innovative, work-in-progress, multimodal AR experience integrating interactive narration, gestures, hands recognition and voice commands, while a citizen is wearing a head-worn AR display, promoting environmental awareness, health, and wellness. By combining AR technologies and a sensor network, GoNature AR provides citizen awareness of real-time, multimodal, air and noise pollution data.","PeriodicalId":120872,"journal":{"name":"Proceedings of the 2023 ACM International Conference on Interactive Media Experiences","volume":"22 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128992434","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Developments in object-based media and IP-based delivery offer an opportunity to create superior audience experiences through personalisation. Towards the aim of making personalised experiences regularly available across the breadth of audio-visual media, we conducted a study to understand how personalised experiences are being created. This consisted of interviews with producers of six representative case studies, followed by a thematic analysis. We describe the workflows and report on the producers’ experiences and obstacles faced. We found that the metadata models, enabling personalisation, were developed independently for each experience, restricting interoperability of personalisation affordances provided to users. Furthermore, the available tools were not effectively integrated into preferred workflows, substantially increasing role responsibilities and production time. To ameliorate these issues, we propose the development of a unifying metadata framework and novel production tools. These tools should be integrated into existing workflows; improve efficiency using AI; and enable producers to serve more diverse audiences.
{"title":"Producing Personalised Object-Based Audio-Visual Experiences: an Ethnographic Study","authors":"Craig Cieciura, M. Glancy, P. Jackson","doi":"10.1145/3573381.3596156","DOIUrl":"https://doi.org/10.1145/3573381.3596156","url":null,"abstract":"Developments in object-based media and IP-based delivery offer an opportunity to create superior audience experiences through personalisation. Towards the aim of making personalised experiences regularly available across the breadth of audio-visual media, we conducted a study to understand how personalised experiences are being created. This consisted of interviews with producers of six representative case studies, followed by a thematic analysis. We describe the workflows and report on the producers’ experiences and obstacles faced. We found that the metadata models, enabling personalisation, were developed independently for each experience, restricting interoperability of personalisation affordances provided to users. Furthermore, the available tools were not effectively integrated into preferred workflows, substantially increasing role responsibilities and production time. To ameliorate these issues, we propose the development of a unifying metadata framework and novel production tools. These tools should be integrated into existing workflows; improve efficiency using AI; and enable producers to serve more diverse audiences.","PeriodicalId":120872,"journal":{"name":"Proceedings of the 2023 ACM International Conference on Interactive Media Experiences","volume":"71 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125271190","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Different brain diseases are associated with distinct patterns of affected brain regions. Neuroscientists aim to understand the variability of these patterns across different brain diseases. We used a user-centered design approach to develop support for neuroscientists in exploring and comparing which regions are affected by which diseases based on neuroscience literature. Our study involved six neuroscientists and nine visualization experts who evaluated the usability and explainability of our visualization tool for comparing disease patterns in brain regions. Our results show that our tool can help neuroscientists explore and visualize relationships between brain regions and diseases, and compare patterns of affected regions in an immersive AR environment.
{"title":"DatAR: Comparing Relationships between Brain Regions and Diseases: An Immersive AR Tool for Neuroscientists","authors":"Ghazaleh Tanhaei, L. Hardman, Wolfgang Hürst","doi":"10.1145/3573381.3597222","DOIUrl":"https://doi.org/10.1145/3573381.3597222","url":null,"abstract":"Different brain diseases are associated with distinct patterns of affected brain regions. Neuroscientists aim to understand the variability of these patterns across different brain diseases. We used a user-centered design approach to develop support for neuroscientists in exploring and comparing which regions are affected by which diseases based on neuroscience literature. Our study involved six neuroscientists and nine visualization experts who evaluated the usability and explainability of our visualization tool for comparing disease patterns in brain regions. Our results show that our tool can help neuroscientists explore and visualize relationships between brain regions and diseases, and compare patterns of affected regions in an immersive AR environment.","PeriodicalId":120872,"journal":{"name":"Proceedings of the 2023 ACM International Conference on Interactive Media Experiences","volume":"28 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127656045","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The 1980s saw full-motion video (FMV) titles inaugurate design patterns for a new genre of interactive digital narrative (IDN). Recently, this genre made a resurgence in popularity. Despite the intervening years, FMV’s design conventions remain tightly coupled to the affordances of laserdisc technology. This paper employs IDN affordances and aesthetics as a lens to examine modern FMV games, namely the recent works of Wales Interactive. Also, this paper leverages research on the emerging conventions of Timeline–an authoring platform for tightly parallel, parameterized stories–to address the challenges of FMV design.
{"title":"Navigating Full-Motion Video: Emerging Design Patterns for Parameterized Replay Stories","authors":"Pedro Silva","doi":"10.1145/3573381.3596455","DOIUrl":"https://doi.org/10.1145/3573381.3596455","url":null,"abstract":"The 1980s saw full-motion video (FMV) titles inaugurate design patterns for a new genre of interactive digital narrative (IDN). Recently, this genre made a resurgence in popularity. Despite the intervening years, FMV’s design conventions remain tightly coupled to the affordances of laserdisc technology. This paper employs IDN affordances and aesthetics as a lens to examine modern FMV games, namely the recent works of Wales Interactive. Also, this paper leverages research on the emerging conventions of Timeline–an authoring platform for tightly parallel, parameterized stories–to address the challenges of FMV design.","PeriodicalId":120872,"journal":{"name":"Proceedings of the 2023 ACM International Conference on Interactive Media Experiences","volume":"8 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125184269","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
An existing body of Human-Computer Interaction (HCI) work has had a focus on cultural heritage settings such as historic sites and museums where digital technologies are used to augment user experience through a variety of methods such as interactive displays, museum guides and mixed/virtual reality experiences. However, very little of this work has been used to explore some of the more contemporary and pressing issues surrounding museums and heritage today. This project, Nothing Beside Remains explores the nature of narratives and stories in the context of museums and heritage through the creation of tangible interactive digital artefacts. Specifically, it looks at ‘Contested Histories’ - a broad spectrum of issues relating to heritage sites and museum artefacts that in recent years have become more urgent to address. These can include issues such as heritage loss (e.g., through decay, destruction or theft either through nature or human conflict), unsustainable preservation and conservation practice (the continuing accumulation of historic artefacts, lack of financial and specialist resources in maintaining sites and artefacts) and repatriation (the disputed claims by nations in reclaiming artefacts that were looted often due to colonial activity). Design (and additionally Speculative Design) as an approach to these issues allows individuals or groups to imagine radically different futures or pose questions relating to phenomena through the convergence of creativity and critical theory – therefore this work aims to extend how this kind of critical enquiry can pose questions to museum and heritage sectors as well as other stakeholders, surrounding contentious narratives and propose possible futures. It involves the creation of physical digital artefacts and tests their effectiveness in creating debate and discussion surrounding these issues. In addition, it will look at the agency of museum and heritage organisations to address these problems and explore experimental design work as a vehicle for actionable change.
{"title":"Nothing Beside Remains: How might we use digital technologies and speculative design to explore the contested and hidden histories of heritage and museum artefacts?","authors":"Saad Ahmed Maqbool, D. Maxwell","doi":"10.1145/3573381.3596505","DOIUrl":"https://doi.org/10.1145/3573381.3596505","url":null,"abstract":"An existing body of Human-Computer Interaction (HCI) work has had a focus on cultural heritage settings such as historic sites and museums where digital technologies are used to augment user experience through a variety of methods such as interactive displays, museum guides and mixed/virtual reality experiences. However, very little of this work has been used to explore some of the more contemporary and pressing issues surrounding museums and heritage today. This project, Nothing Beside Remains explores the nature of narratives and stories in the context of museums and heritage through the creation of tangible interactive digital artefacts. Specifically, it looks at ‘Contested Histories’ - a broad spectrum of issues relating to heritage sites and museum artefacts that in recent years have become more urgent to address. These can include issues such as heritage loss (e.g., through decay, destruction or theft either through nature or human conflict), unsustainable preservation and conservation practice (the continuing accumulation of historic artefacts, lack of financial and specialist resources in maintaining sites and artefacts) and repatriation (the disputed claims by nations in reclaiming artefacts that were looted often due to colonial activity). Design (and additionally Speculative Design) as an approach to these issues allows individuals or groups to imagine radically different futures or pose questions relating to phenomena through the convergence of creativity and critical theory – therefore this work aims to extend how this kind of critical enquiry can pose questions to museum and heritage sectors as well as other stakeholders, surrounding contentious narratives and propose possible futures. It involves the creation of physical digital artefacts and tests their effectiveness in creating debate and discussion surrounding these issues. In addition, it will look at the agency of museum and heritage organisations to address these problems and explore experimental design work as a vehicle for actionable change.","PeriodicalId":120872,"journal":{"name":"Proceedings of the 2023 ACM International Conference on Interactive Media Experiences","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130088221","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
In the context of Television (TV), intelligent voice assistants are still mainly based on reactive behaviours, only responding to users' requests. However, the addition of proactivity in TV assistants could have the potential to reduce the user-assistant interaction effort and bring an additional "human" layer, leading to a more organic adoption of these systems. Within this framework, this paper aims to contribute to the understanding of the relevance of a set of proactive scenarios for TV and get clues on how to design these behaviours to be efficient and avoid intrusion feelings. To this end, focus groups were conducted to discuss the conceptualised scenarios. The results showed that most of the idealized scenarios were relevant for the participants and that they will be positive if some characteristics can be pre-configured regarding the type of communication and intrusion of proactive suggestions.
{"title":"Proactivity in the TV Context: Understanding the Relevance and Characteristics of Proactive Behaviours in Voice Assistants","authors":"Tiffany Marques, J. Abreu, Rita Santos","doi":"10.1145/3573381.3597219","DOIUrl":"https://doi.org/10.1145/3573381.3597219","url":null,"abstract":"In the context of Television (TV), intelligent voice assistants are still mainly based on reactive behaviours, only responding to users' requests. However, the addition of proactivity in TV assistants could have the potential to reduce the user-assistant interaction effort and bring an additional \"human\" layer, leading to a more organic adoption of these systems. Within this framework, this paper aims to contribute to the understanding of the relevance of a set of proactive scenarios for TV and get clues on how to design these behaviours to be efficient and avoid intrusion feelings. To this end, focus groups were conducted to discuss the conceptualised scenarios. The results showed that most of the idealized scenarios were relevant for the participants and that they will be positive if some characteristics can be pre-configured regarding the type of communication and intrusion of proactive suggestions.","PeriodicalId":120872,"journal":{"name":"Proceedings of the 2023 ACM International Conference on Interactive Media Experiences","volume":"129 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127093748","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
F. Rocca, Madison Dave, Valérie Duvivier, Agnès Van Daele, M. Demeuse, Antoine Derobertmasure, M. Mancas, B. Gosselin
Human audience analysis is crucial in numerous domains to understand people’s behavior based on their knowledge and environment. In this paper we focus on simulation-based training which has become a popular teaching approach that requires a trainer able to manage a lot of data at the same time. We present tools that are currently being developed to help trainers from two different fields: training for practical teaching gestures and training in civil defense. In this sense, three technological blocks are built to collect and analyze data about trainers’ and learners’ gestures, gaze, speech and movements. The paper also discusses the future work planned for this project, including the integration of the framework into the Noldus system and its use in civil security training. Overall, the article highlights the potential of technology to improve simulation-based training and provides a roadmap for future development.
{"title":"Designing an Assistance Tool for Analyzing and Modeling Trainer Activity in Professional Training Through Simulation","authors":"F. Rocca, Madison Dave, Valérie Duvivier, Agnès Van Daele, M. Demeuse, Antoine Derobertmasure, M. Mancas, B. Gosselin","doi":"10.1145/3573381.3596475","DOIUrl":"https://doi.org/10.1145/3573381.3596475","url":null,"abstract":"Human audience analysis is crucial in numerous domains to understand people’s behavior based on their knowledge and environment. In this paper we focus on simulation-based training which has become a popular teaching approach that requires a trainer able to manage a lot of data at the same time. We present tools that are currently being developed to help trainers from two different fields: training for practical teaching gestures and training in civil defense. In this sense, three technological blocks are built to collect and analyze data about trainers’ and learners’ gestures, gaze, speech and movements. The paper also discusses the future work planned for this project, including the integration of the framework into the Noldus system and its use in civil security training. Overall, the article highlights the potential of technology to improve simulation-based training and provides a roadmap for future development.","PeriodicalId":120872,"journal":{"name":"Proceedings of the 2023 ACM International Conference on Interactive Media Experiences","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128941781","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Point clouds are acknowledged as an essential data structure to represent 3D objects in many use cases, notably immersive experience settings such as Virtual, Augmented or Mixed Reality. This work is the first part of a research project on immersive Quality Assessment of point clouds in different lighting. In this report, I focus mainly on the physically-based rendering of such data in Unity 3D, and the impact of point cloud compression when considering various lighting conditions on the objects. These first observations and results will serve in the implementation of a 6DoF immersive experiment setting for subjective quality assessment.
{"title":"Physically-based Lighting of 3D Point Clouds for Quality Assessment","authors":"Amar Tious, Toinon Vigier, V. Ricordel","doi":"10.1145/3573381.3597231","DOIUrl":"https://doi.org/10.1145/3573381.3597231","url":null,"abstract":"Point clouds are acknowledged as an essential data structure to represent 3D objects in many use cases, notably immersive experience settings such as Virtual, Augmented or Mixed Reality. This work is the first part of a research project on immersive Quality Assessment of point clouds in different lighting. In this report, I focus mainly on the physically-based rendering of such data in Unity 3D, and the impact of point cloud compression when considering various lighting conditions on the objects. These first observations and results will serve in the implementation of a 6DoF immersive experiment setting for subjective quality assessment.","PeriodicalId":120872,"journal":{"name":"Proceedings of the 2023 ACM International Conference on Interactive Media Experiences","volume":"26 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129134665","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Wala Elsharif, James She, Preslav Nakov, Simon Wong
With the current and continuous advancements in the field of text-to-image modeling, it has become critical to design prompts that make the best of these model capabilities and guides them to generate the most desirable images, and thus the field of prompt engineering has emerged. Here, we study a method to use prompt engineering to enhance text-to-image model representation of the Arabic culture. This work proposes a simple, novel approach for prompt engineering that uses the domain knowledge of a state-of-the-art language model, GPT, to perform the task of prompt augmentation, where a simple, initial prompt is used to generate multiple, more detailed prompts related to the Arabic culture from multiple categories through a GPT model through a process known as in-context learning. The augmented prompts are then used to generate images enhanced for the Arabic culture. We perform multiple experiments with a number of participants to evaluate the performance of the proposed method, which shows promising results, specially for generating prompts that are more inclusive of the different Arabic countries and with a wider variety in terms of image subjects, where we find that our proposed method generates image with more variety 85 % of the time and are more inclusive of the Arabic countries more than 72.66 % of the time, compared to the direct approach.
{"title":"Enhancing Arabic Content Generation with Prompt Augmentation Using Integrated GPT and Text-to-Image Models","authors":"Wala Elsharif, James She, Preslav Nakov, Simon Wong","doi":"10.1145/3573381.3596466","DOIUrl":"https://doi.org/10.1145/3573381.3596466","url":null,"abstract":"With the current and continuous advancements in the field of text-to-image modeling, it has become critical to design prompts that make the best of these model capabilities and guides them to generate the most desirable images, and thus the field of prompt engineering has emerged. Here, we study a method to use prompt engineering to enhance text-to-image model representation of the Arabic culture. This work proposes a simple, novel approach for prompt engineering that uses the domain knowledge of a state-of-the-art language model, GPT, to perform the task of prompt augmentation, where a simple, initial prompt is used to generate multiple, more detailed prompts related to the Arabic culture from multiple categories through a GPT model through a process known as in-context learning. The augmented prompts are then used to generate images enhanced for the Arabic culture. We perform multiple experiments with a number of participants to evaluate the performance of the proposed method, which shows promising results, specially for generating prompts that are more inclusive of the different Arabic countries and with a wider variety in terms of image subjects, where we find that our proposed method generates image with more variety 85 % of the time and are more inclusive of the Arabic countries more than 72.66 % of the time, compared to the direct approach.","PeriodicalId":120872,"journal":{"name":"Proceedings of the 2023 ACM International Conference on Interactive Media Experiences","volume":"35 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128603920","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：