以人工智慧（AI）輔助自學程式編碼：以Google Colaboratory整理Google Scholar資料為例

Artificial Intelligence-assisted Self-learning of Coding: An Example of Using Google Colaboratory for Organization of data from Google Scholar

目的
隨著人工智慧（Artificial Intelligence, AI）技術的發展，自學程式編碼（Coding）變得高效且便捷。本研究探討如何利用AI工具和Google Colaboratory（Colab）平台輔助自學Python程式編碼，並以整理Google Scholar搜尋到的資料為例做說明。
方法
以Microsoft Copilot、DeepSeek、ChatGPT及Google Gemini等AI模型交替使用，輔助Python編碼，於Google雲端硬碟/Colab平台執行程式。使用電腦型號為ASUS筆記型LAPTOP- TSDMU3MP，配備Intel Core i7-1165G7 2.80GHz處理器、運行Windows 11。
結果
本研究探討AI輔助程式編碼應用的可行性，發現AI可作為開發者的得力助手。透過Colab雲端運算能力，完全沒有基礎的初學者亦能快速掌握Python程式應用技巧。以整理Google Scholar學術資料為例，發現給予AI提示詞，即可自動生成Python編碼，有些會有使用步驟、功能說明、程式碼說明、範例運行及注意事項，且會除錯及優化程式碼等，於Colab平臺，對於沒有Python程式語言基礎的初學者可輕易上手。再者透過不同的輸入條件，及於不同AI，可產生不同的結果以資比較，可達到Python程式語言的自我學習。
結論
以AI輔助，可大大降低進入程式編碼的門檻，於Colab平臺可輕易上手執行Python。因為AI會有註解及程式碼解釋，因此也可達到初學程式語言者的自學參考。

Purposes
Rapid advances in artificial intelligence (AI) techniques have made self-learning coding highly efficient and convenient. This study investigated the use of AI tools and the Google Colaboratory (Colab) platform to assist self-learning of Python coding, with the application of Colab to the organization of data retrieved from Google Scholar as an illustrative example.
Methods
AI models such as Microsoft Copilot, DeepSeek, ChatGPT, and Google Gemini were used sequentially to assist with Python coding, and codes were executed in Google Drive/ Colab. All processes were performed on an ASUS laptop (TSDMU3MP) with an Intel Core i7-1165G7 2.80 GHz processor and the Windows 11 operating system.
Results
We investigated the feasibility of using AI to assist with coding and discovered that AI could serve as a powerful aid for developers. By leveraging the cloud computing capabilities of Colab, even beginners with no prior experience can quickly learn Python programming techniques. During the application of self-learned Python coding to the organization of academic data from Google Scholar, the provision of prompts to AI models led to the automatic generation of Python codes. Certain models provided additional information such as user instructions, function explanations, code annotations, example executions, and precautions. AI could also debug and optimize code. This enabled beginners with no prior experience to effortlessly gain familiarity with Python programming on the Colab platform. Using different input conditions in different AI models produced different results, which could be compared to achieving self-learning of Python programming.
Conclusions
With AI assistance, the barrier to learning coding can be significantly reduced, and familiarity with Python code execution can be easily gained on the Colab platform. The annotations and explanations provided by AI can also serve as references for self-learning by programming beginners.