Why Mathematical Modelers Must Embrace LLMs and Forget Outdated Skills

In the world of mathematical modeling, models serve as windows into complex phenomena, tools for forecasting, and essential references for decision‑making. As technology advances and data proliferates, yesterday’s models quickly become obsolete, and many once‑vital skills lose relevance, making lifelong learning crucial.

Henry Ford, the founder of Ford Motor Company, famously said that anyone can learn anything if they are willing to put in the necessary effort, inspiring modelers to adopt new ideas and master emerging tools and techniques.

Alvin Toffler, a renowned futurist, predicted the information age in his seminal work "The Third Wave" and emphasized that the future belongs to those who can learn new skills and continuously relearn, a strong signal for modelers to keep up with evolving data‑science and technology landscapes.

Knowledge and Skill Iteration in Mathematical Modeling

Mathematical modeling is a dynamic process that requires continual iteration and updating. New data and theories may render existing models inaccurate, necessitating reevaluation or complete replacement—for example, financial market models must adapt to changing market behavior by incorporating new algorithms and theories.

Similarly, skills must evolve. In the era of data science and machine learning, programming, data‑processing, and algorithmic understanding become essential, and the emergence of large language models (LLMs) such as ChatGPT further transforms the field.

Impact of Large Language Models (LLMs)

LLMs like ChatGPT affect mathematical modeling in several ways:

They can quickly process and analyze massive text datasets, handling data cleaning and preliminary analysis, thereby greatly improving efficiency.

Through natural‑language interaction, they help modelers swiftly grasp model requirements and offer construction suggestions.

LLMs assist in interpreting model results, providing more intuitive explanations that aid non‑experts in understanding complex outputs.

They serve as powerful educational tools, enabling modelers to learn new algorithms, programming languages, or modeling techniques.

Why Discard Certain Knowledge and Skills?

Technological progress, exemplified by LLMs, simplifies many previously complex processes, making some specialized skills (e.g., certain programming or statistical analysis tasks) less necessary as practitioners rely on advanced tools.

How to Achieve Continuous Learning?

Integrate new tools: Modelers should learn to use LLMs and embed them into existing workflows.

Commit to ongoing learning and adaptation: As LLMs evolve rapidly, practitioners must continuously update their understanding and usage methods.

Foster interdisciplinary collaboration: Working with data scientists, software engineers, and machine‑learning experts reveals effective ways to leverage LLMs.

Practice and reflect: While applying LLMs, modelers should evaluate outcomes, adjust models, and optimize results.

Lifetime learning is not merely a choice but a necessity for mathematical modelers. By recognizing that “much knowledge will eventually be discarded and many skills become unnecessary,” modelers can better harness tools like LLMs to boost efficiency and quality, ensuring their modeling expertise remains current and capable of tackling complex problems.

— Author: Wang Haihua