semantic coherence
Semantic coherence is a measure of the logical and meaningful flow of ideas within generated content, ensuring that the output makes sense in its context, which is particularly important in natural language processing tasks.
- AdmTree: Compressing Lengthy Context with Adaptive Semantic Trees
- Color Conditional Generation with Sliced Wasserstein Guidance
- Diffusion Feature Field for Text-based 3D Editing with Gaussian Splatting
- GD$^2$: Robust Graph Learning under Label Noise via Dual-View Prediction Discrepancy
- GraphMaster: Automated Graph Synthesis via LLM Agents in Data-Limited Environments
- Motion4D: Learning 3D-Consistent Motion and Semantics for 4D Scene Understanding
- NeuroPath: Neurobiology-Inspired Path Tracking and Reflection for Semantically Coherent Retrieval
- Rendering-Aware Reinforcement Learning for Vector Graphics Generation
- SECA: Semantically Equivalent and Coherent Attacks for Eliciting LLM Hallucinations
- Training-Free Safe Text Embedding Guidance for Text-to-Image Diffusion Models