The purpose of this page is to share basic knowledge with those new to CL research and explain what to consider when conducting research. When writing the content, I used Gemini's image creation and text editing features, as well as Notion's writing features.
If you'd like to contribute to this page, please contact me via: [email protected].
1. Data Collection and Recruitment Optimization (Maximizing Advertising ROI)
1.3 Advertising Goal Optimization
1.4 Ethical/Bias Assessment of Datasets
1.5 Standardization of Human Evaluation
1.7 Multidisciplinary Expert Consulting
2. LLM Analysis and Code Execution Tips (Cost and Reliability)
2.1 API Cost Resolution Strategy
2.2 Utilizing Pre-Testing Tools
3. Reproducibility & System Rigor (Visual & Statistical)
3.3 Detailed records of models and hyperparameters
3.4 Similarity / Distance Metrics
3.5 Matrix Correlation & Significance Testing
1.1 Localization
Consider the cultural characteristics of each target market.
Quickly grasp what content is trending.
Design costs and experimental plans to suit the characteristics of the market.
Example: In Korean culture, people view typology tests as fun games.
In contrast, English-speaking countries tended to view the typology test as an opportunity for serious self-reflection.
Aside from variables inherent to the experiment that must be strictly controlled, when distributing content, consider making subtle adjustments to the wording to align with market characteristics.
To reduce participant dropout rates and increase engagement in high-intensity cognitive experiments, elements such as personalized type classification tests utilizing 'fun metrics' are effective.
The Vibe coding platform makes it easy to get started, even if you don't have extensive development knowledge. Recommended: Lovable.dev / AI Studio (Google)