Can we trust our model?
Robustness refers to an AI system's ability to maintain stable performance despite noise, adversarial attacks, or shifts in data distribution.
Fairness and bias mitigation are also core elements, ensuring AI systems produce equitable outcomes without discrimination against specific groups. Explainable AI technologies enhance transparency in the model's decision-making process, thereby building user trust. 

Foundation models are general-purpose AI models pre-trained on large-scale datasets. Representative models such as GPT, BERT, and CLIP belong to this category, demonstrating outstanding performance in natural language processing, computer vision, audio, and multimodal tasks.
These models can be fine-tuned for specific domains or tasks through transfer learning.

AI for X refers to the research and development domain of applying artificial intelligence technology specialized for specific domains or industries.
For instance, AI for Climate contributes to climate change prediction, energy efficiency optimization, and carbon emission reduction.
Furthermore, through Agentic AI, our ultimate goal is to build autonomous advanced AI systems that set their own goals, formulate plans, execute tasks, and adapt to changing environments with minimal human intervention.