When a ChatBot Meets the Scenic Beauty of Poetry: Integrating ChatBots with Narrative Components to Explore the Beauty of Poetry
/GuanZe Liao - Professor, Institute of Learning Sciences and Technologies/
/Peiyi Zhou-Graduate student, Institute of Learning Sciences and Technologies/
Through the tempering of time, poetry leaves behind everlasting emotions and sentiments. Poetry not only carries the wisdom of Chinese literature, but also demonstrates a perpetual format for art manifestation. It allows people’s minds to break free from the restrictions of the real world, and extends the breadth and diversify of people’s thinking through imagination. In literary composition and creation, poets crystallize beautiful and romantic scenes through imagination, and create poems and verses that are appreciated and recited by readers. In artistic creation, and through imagination, artists transform their inner ideas into real works.
Through the scientific methods proposed in the project "Meanings of Words/Phrases and Associative Images — Visualization Design Methods for Poetry Scenarios" led by Professor Liao Guan-Ze under the Ministry of Science and Technology, the texts and images of poetry can be deconstructed and then reconstructed. The first phase of the project completed a definition of vocabulary connotative meanings, a collection of associative vocabulary, and construction of poetic imagery for ten Tang Shi poems and ten Song Ci poems. From these, scene/image composition elements and vocabulary correspondence were depicted, and the associations inspired by the texts and the image content features were interpreted; from this, a prototype of the Visualization Design Methods for Poetry Scenarios was compiled. The construction of scenic images was based on the word/phrase stimulation model in semantics. Through associative vocabulary collected from the cloud, content from the semantic network constructed based on poetry texts and image association was created; this allowed the process of converting texts into scenic images to be grasped, and the characteristics of poetic images to be discovered.
In the second and third phases of the project, a digital cloud platform for the collection of associative vocabulary was established, and the completed poetic image designs were compiled into a poetic text corpus. The cloud platform was used to collect the associative vocabulary of the poetry texts and to analyze the local features of images and vocabulary correspondence, thereby constructing the semantic network of scenario images and exploring the semantic network differences between poetry texts and images. Thus, the relationship between the perception of poetry texts/poetry scenarios and scenario images could be interpreted, the relationship between the association inspired by texts and the features of image content could be defined, and a system of visualization design methods for poetry scenarios could be constructed.
The fourth phase of the project was to develop a LINE Chatbot, which was realized via a learning tool collaboratively developed by a team formed by students of the Institute of Learning Sciences and Technologies. With graduate student Wu Wen-hsiu participating in the backend development for system programming, and Chou Pei-i taking part in the creation of the robot and the “narrative component” editing, a ChatBot named “The Beauty of Poetic Images: My Poetic Bot” was created. A chat space was provided through the aesthetic experiences produced by the synesthesia between poetry texts and images. The edited narrative components were exported through the backend programming, then the next chat was triggered through the question & answer options, and users were guided to learn more about the poetry through the questions and corresponding poetic images. This tool is not restricted by time and space; it can interact with users at any time, motivating learning through conversation based on narrative components, thus unveiling the composer’s ideas and the underlying image atmosphere.
The “narrative components” were created based on guiding questions, with connections extending to the domain of poetic literature, so that users are guided to associate the questions with poetic images. Moreover, the narrative components provide comparison of the poetic images and the original poetry texts, so that users can observe the details of pictures and verses, and comprehend the beauty and scenic atmosphere brought out by the poetic images. In the study, two representative Tang Shi poems and Song Ci poems were selected for a study of the relationship between poetic images and aesthetic perception, by reference to the textual meaning of the poems and the associative images produced when perspectives change between close and distant images.
After the Beauty of Poetry Scenarios: My Poetic Bot platform was combined with the narrative components, in-person workshops were organized to explain applications of the ChatBot, in which subjects were invited to use the platform and fill out questionnaires. The “Satisfaction Survey for Learning Poetic Images through the Beauty of Poetry Scenes: My Poetic Bot” was used to analyze users’ satisfaction in learning poetry and aesthetics. By using a LINE Chatbot as a learning tool and incorporating the narrative components for learning poetic images and aesthetics, the relationship between the images of poetic texts and aesthetic perceptions of their images can be explored, which makes it easier for users to perceive the beauty conveyed in the poetic images.
