At the CUHK REliable Computing laboratory (CURE Lab.), our core research methodology is rooted in creating innovative solutions that directly tackle the limitations of state-of-the-art computing technologies. Currently, we are passionately committed to the exploration and advancement of artificial intelligence across a variety of cutting-edge disciplines. First, we seek to build revolutionary generative AI models, focusing on text-to-image generation at the moment. Simultaneously, we are working to enhance AI's capacity for logical reasoning with innovative circuit learning techniques, striving to create models that can dramatically improve performance for hard computational problems such as SAT/SMT. We are also working on time series analysis, aiming to improve AI's ability to understand and interact with the continuous world. Last but not least, we strive to make AI not just more powerful, but also safer and more trustworthy for everyone. Therefore, we continue to work on the safety and security aspects of AI, developing robust AI models that can withstand adversarial attacks and operate reliably even under uncertain or volatile conditions.