iterative refinement
A technique in machine learning where an initial output is progressively improved through repeated iterations or adjustments. This is common in tasks such as image generation or natural language processing.
- Brain-like Variational Inference
- Compositional Neural Network Verification via Assume-Guarantee Reasoning
- Conditional Diffusion Anomaly Modeling on Graphs
- Continuous Diffusion Model for Language Modeling
- DeblurDiff: Real-Word Image Deblurring with Generative Diffusion Models
- DeepASA: An Object-Oriented Multi-Purpose Network for Auditory Scene Analysis
- Diffusion-Driven Progressive Target Manipulation for Source-Free Domain Adaptation
- ECO: Evolving Core Knowledge for Efficient Transfer
- GraphMaster: Automated Graph Synthesis via LLM Agents in Data-Limited Environments
- Human-assisted Robotic Policy Refinement via Action Preference Optimization
- HyPINO: Multi-Physics Neural Operators via HyperPINNs and the Method of Manufactured Solutions
- MLE-STAR: Machine Learning Engineering Agent via Search and Targeted Refinement
- Multi-Agent Debate for LLM Judges with Adaptive Stability Detection
- NEP: Autoregressive Image Editing via Next Editing Token Prediction
- Remasking Discrete Diffusion Models with Inference-Time Scaling
- SuperGPQA: Scaling LLM Evaluation across 285 Graduate Disciplines