SYNTRA: self-improving data synthesis scales embedding models across 15 domains
_reachsumit · x · 2026-10-14
A new arXiv paper introduces SYNTRA, a retrieval data synthesis method that needs only an unlabelled corpus and a few query-document pairs. Its self-improvement loop iteratively refines reusable synthesis instructions using feedback on trial outputs.
- Validated on 24 retrieval datasets across 15 domains; improves strong existing embedding models with and without labelled sets
- Enhances retrieval on all six evaluated tasks, beating human calibration on four
- MS MARCO retrieval keeps improving as training data scales to 32x
- Code and data released
Related event: SYNTRA Enables Label-Free Cross-Domain Expansion of Embedding Models(2 posts)→
More from Research
- FrontierML benchmark tests whether agents can spend money to scale model performance — JayAlammar · 2026-10-14
- AIPOCH's Open-Science ships portable .science packages that share full research context, not just outputs — ahuja_priyank · 2026-10-14
- Video-Arena: Game-Based Video Benchmark Shows Native Models Score Just 13.2 F1 — Kuan Zhang · 2026-10-14
- LSCBench: Vibe Coding Agents Write Functional Code, But Only 25% Is Also Secure — Qwen · 2026-10-14
- Omni-LLMs Know More Than They Say: NAVER's OmniCatalyst Adds 7.29% via Self-Generated Text — naver-ai · 2026-10-14
- Repeated-Data Pretraining Study: Smaller Batches Act as Regularizers and Beat More GPUs — Sewon Min · 2026-10-14