Amazon's SMART Self-Evolving Multi-Agent System Tops All 15 Subtitle Arena Directions, Cuts Penalty 6.9%
amazon · hf · 2026-10-01
Amazon's SMART is a self-evolving multi-agent system for long-form subtitle translation: it builds series-level persistent memory, uses a dynamic router and Mixture-of-Agents layer with terminology and constraint tools, and a judge-refiner loop that updates prompts and routing policies without retraining LLMs. It ships Subtitle Arena (14 genres, 2-198 episodes per series, 15 locales) and the SubMQM framework. SMART scores best MQM in all 15 directions, cutting average penalty 6.9% vs. the strongest competing system, and tops MuSC across all four language pairs with a 4.50/5 human eval.
More from coding & agent
- Multiple AI agents per person is going normal: OpenClaw now runs on a $10/mo VPS — steipete · 2026-10-01
- Early user: GPT-6 Sol feels slower and dumber than 5.6 in Codex — burkov · 2026-10-01
- No Evals, Building Blind: Contextual Embedding Models Must Rank Disambiguating Chunks — antoine_chaffin · 2026-10-01
- Jev Sentinel open-sources per-action agent monitor that scored 53,870 HF payloads, flagging 98.3% — schwentker · 2026-10-01
- Priors: An Onchain Credit Bureau for AI Agents Emerges — econoar · 2026-10-01
- Redditor runs unattended DeepSeek loops for days: 237M tokens for just $3.48 — dogfoodarchitect · 2026-10-01