I am deeply interested in exploring and exploiting the uncertainty of deep neural networks. Uncertainty in deep neural networks is a valuable concept that can be used to make models more reliable, robust, and useful in various applications; by providing insights into the model's level of confidence in its predictions. 

• Relevant domain: Active Learning, Out-of-Distribution Detection, Data Augmentation, etc.

Artificial Intelligence has a wide range of applications across various industries and domains; including social, economic, and military fields. Its ability to mimic human intelligence and automate tasks makes it a powerful tool for solving complex problems and improving efficiency. The field continues to evolve, and the potential of AI technologies will grow as it evolves.

• Relevant domain: Object Detection, Semantic Segmentation, Target Tracking, Autonomous Driving, etc.

Theoretical analysis of AI involves delving deep into the mathematical and algorithmic foundations of artificial intelligence. Through this analysis, we can understand the limits, biases, and uncertainties inherent in AI algorithms, as well as ways to improve their robustness and fairness. Theoretical insights guide the development of more advanced AI models.

• Relevant domain: Bayesian approach, PAC-Bayes, etc.