Automated map generalization has been a major area of research for decades but has still not reached maturity. Besides the needs for more adaptive algorithms, a fundamental question remains: How can we transfer human generalization knowledge into a computational system more effectively? Previous efforts do not seem capable to fully overcome the “knowledge acquisition bottleneck.” As new theories and technologies emerged in artificial intelligence (particularly deep learning), computers are now able to tackle human-level tasks with superior performance, showing great potential in automated generalization. Meanwhile, crowdsourced geographic information and social sensing is growing at an increasing speed, and the needs for visualizing and analyzing massive geo-referenced data at various scales are numerous. It is therefore necessary to adapt map generalization to these fields. This highlights the potential of applying map generalization in the visual, interactive, and exploratory analysis of abstract (e.g., hierarchical relations) and physical (e.g., movement trajectories) data. This topical collection brings six contributions reporting recent progress and trends in automated generalization in various aspects mentioned above, with which we hope to trigger further discussion and research in our field with new ideas and methodologies.