Introspectus AI is a generative AI-based system designed to enhance self-reflection and support positive behavior change. By leveraging multimodal information from users' daily life recordings, it provides personalized and detailed feedback, aiming to deepen self-awareness and facilitate positive behavioral adjustments. This study explores the short-term and long-term impacts of interacting with Introspectus AI, focusing on its potential to enhance reflective practices and improve the acceptance of generative AI tools. Following the user experience was defined through an initial round of workshops with four experts. The resulting system was evaluated through a long-term study involving 64 participants. The results demonstrate that AI-supported interventions significantly improved engagement in self-reflection, the need for reflection, and insight, while also increasing user acceptance of generative AI over time. These findings underscore the potential of generative AI as a practical tool for self-improvement, offering insights into its broader applicability in promoting well-being and personal growth.

Mixed Integer Programming (MIP) is one of the most widely used modeling techniques for combinatorial optimization problems. In many applications, a similar MIP model is solved on a regular basis, maintaining remarkable similarities in model structures and solution appearances but differing in formulation coefficients. This offers the opportunity for machine learning methods to explore the correlations between model structures and the resulting solution values. To address this issue, we propose to represent a MIP instance using a tripartite graph, based on which a Graph Convolutional Network (GCN) is constructed to predict solution values for binary variables. The predicted solutions are used to generate a local branching type cut which can be either treated as a global (invalid) inequality in the formulation resulting in a heuristic approach to solve the MIP, or as a root branching rule resulting in an exact approach. Computational evaluations on 8 distinct types of MIP problems show that the proposed framework improves the primal solution finding performance significantly on a state-of-the-art open-source MIP solver.