[NeurIPS 2025] Domain-RAG: Retrieval-Guided Compositional Image Generation for Cross-Domain Few-Shot Object Detection
-
Updated
Feb 2, 2026 - Python
[NeurIPS 2025] Domain-RAG: Retrieval-Guided Compositional Image Generation for Cross-Domain Few-Shot Object Detection
Mcity Data Engine
Build and deploy AI agents for automated labeling, quality assurance, and workflow automation in the Encord ecosystem.
An India-native, self-improving data engine for autonomous driving. Auto-labeling with a calibrated confidence gate, active learning, HD maps, per-object dynamics, and a governed closed loop.
A unified hierarchical engine for managing parameters, state, workflows, and knowledge with structure, meaning, and AI-powered reasoning.
Auto-labeling and VLM scene understanding pipeline for driving data — open-vocabulary detection (Grounding DINO), SAM segmentation, traffic-rule-aware scene tagging (Qwen2.5-VL), pseudo-label mAP evaluation and long-tail mining. Designed to fit in 6 GB VRAM.
Automated visual data harvesting and VLM grounding engine for indoor door detection, video frame filtering, pseudo-labeling, and leakage-free YOLO dataset synthesis.
End-to-end pipeline ingesting product docs from GitHub, refining through bronze/silver/gold on AWS, and serving to an AI agent
To associate your repository with the data-engine topic, visit your repo's landing page and select "manage topics."