Visual Generation in the New Era: An Evolution from Atomic Mapping to Agentic World Modeling

TL;DR

This paper advocates shifting visual generation research from appearance-focused synthesis to models that incorporate structure, dynamics, and causal understanding, proposing a new taxonomy and evaluation framework.

Contribution

It introduces a five-level taxonomy for visual generation, analyzes key technical drivers, and proposes a comprehensive evaluation approach emphasizing structural and causal capabilities.

Findings

Current evaluations overemphasize perceptual quality.

Structural, temporal, and causal failures are often overlooked.

The proposed roadmap guides future development of intelligent visual generators.

Abstract

Recent visual generation models have made major progress in photorealism, typography, instruction following, and interactive editing, yet they still struggle with spatial reasoning, persistent state, long-horizon consistency, and causal understanding. We argue that the field should move beyond appearance synthesis toward intelligent visual generation: plausible visuals grounded in structure, dynamics, domain knowledge, and causal relations. To frame this shift, we introduce a five-level taxonomy: Atomic Generation, Conditional Generation, In-Context Generation, Agentic Generation, and World-Modeling Generation, progressing from passive renderers to interactive, agentic, world-aware generators. We analyze key technical drivers, including flow matching, unified understanding-and-generation models, improved visual representations, post-training, reward modeling, data curation, synthetic data distillation, and sampling acceleration. We further show that current evaluations often overestimate progress by emphasizing perceptual quality while missing structural, temporal, and causal failures. By combining benchmark review, in-the-wild stress tests, and expert-constrained case studies, this roadmap offers a capability-centered lens for understanding, evaluating, and advancing the next generation of intelligent visual generation systems.