人工智慧在語言學習的應用
Hey, for whoever is reading this right now, you can call me Jason. I’m currently in my third year of studying in the English Department at SCU. My time here has had its ups and downs, but things are going well overall. I've been reading and speaking English for a while now, and at the same time, Artificial Intelligence is rising very fast these days, tools like ChatGpt, Suno, etc are capable are doing countless jobs at a speed that we could never have imagined in the past. So here's the question, how can we combine this very new and convenient technology into the learning of language, what benefits it can bring us, and how do we properly implement it without violating academic ethics, what are the concerns? Throughout this course, we get to use generative AI to basically produce music and generate stories, audio, and videos relating to our thoughts, after creating multiple projects with these AI tools on our own, we are able to further understand the role of Artificial Intelligence in language learning.
This is an imitation of an actual cover letter as if I was trying to get the job offer in the research assistant position for the AI project advertised by the National Science Council. I first included my own experiences regarding programming experience in PHP (scripting language) that I learned in SCU and a passion for computer science. To further establish exactly why should they hire me, I talked about the skills and abilities I learned in my high school and college years, and how my time in the hot music and musical clubs really enriched my understanding of how to cooperate and resolve conflicts while working as a group. I later mentioned the special experience of giving a speech, actually two speeches in two of the ICEE weeks events in the SCU, what I learned from them, and what could I do better about them. Then I talked about my experience through coursework in digital multimedia, website development, and eLearning platforms, which makes me more than well-equipped and definitely a qualified candidate for the job.
Although I wouldn't say I have a solid background in music, my experience as a vocalist in my high school's pop music club and my performances in the university's musical theater club have given me some familiarity with the basics of music creation. I understand how complex and challenging it can be to compose a song, especially when it comes to brainstorming ideas, which can be quite time-consuming. In this regard, Suno is a great tool for beginners to learn music creation. It allows anyone to easily design their own song with simple ideas, making it highly accessible and an excellent platform for newcomers. Due to time constraints, I had no choice but to rely on ChatGPT to generate the lyrics for my song. While the result was fairly decent overall and transformed into something quite magical when set to music, I could still sense its limitations. If I didn't provide specific prompts, the AI would solely base its creation on the keywords mentioned in my description. From my perspective, its ability to synthesize and create more nuanced content still needs improvement. That said, I believe that with the rapid advancements in artificial intelligence today, generative AI will one day surpass the limitations of human creativity.
The Rock-Paper-Scissors module, which includes pre-made images for AI training, was prepared by the instructor in advance. This part is similar to the original concept, so we just followed the instructor's guidance to complete it. For the color recognition part, we were directed to another website—though I forgot its name—to train the AI's ability to recognize colors. For example, to teach the robot to recognize the color red, we needed to collect various images of red objects from the internet to train it. Although the training was completed on the website, exporting the module took a very long time, likely due to network issues. As a result, we ended up downloading the pre-prepared module provided by the instructor. Nevertheless, it was an interesting experience.
As shown in the Rock-Paper-Scissors video, although the AI could usually identify the gestures I made or the colors I showed it, it wasn’t always successful. This might be due to limitations in the database’s ability to recognize all inputs accurately. For the Rock-Paper-Scissors module, I initially planned to modify the program so Kebi (the robot) could display what it chose, as the original program only showed the win or lose result without indicating Kebi's choice. I tried to make some changes to the program, but it became evident that my design still had some shortcomings. However, this was a valuable learning experience.
Throughout these lessons, I’ve learned not only how to design a robot’s conversation logic to enable natural interaction with humans but also when to adjust the robot’s responses to better suit specific needs or contexts. Occasionally, I was able to incorporate creativity into the robot's dialogues, making the interaction more enjoyable and seamless. Overall, this course has enhanced my problem-solving skills and creativity, and it has made me excited about the potential of integrating robots and AI into future learning and applications.
This is a rather challenging project in my opinion, since my time is extremely limited due to my other courses and a rather personal illness, so I didn't had a lot of time to work on this one. This video is actually the very combination of what we learned throughout this semester and how can we properly apply it, ElevenLabs for the AI voice, Suno for the background music, ChatGpt for the storyboard and plot, Hailuo for the overall animation and movements of the pictures generated by dreamina. Together, these tools enabled me to overcome the time constraints and technical challenges, allowing me to focus on the creative aspects and the integration of everything we’ve learned. This project represents not just the application of semester-long lessons but also a testament to the potential of AI in facilitating and enhancing creative projects.