PEARL Debuts First User-History Personalized Image Generation Benchmark, +15% Over Baselines
Bo Ni · hf · 2026-10-02
What
A paper introduces the first unified benchmark for personalized image generation driven by a user's accumulated history (reviews, posts, images, captions, metadata), plus a method called PEARL.
Benchmark
- Grounded in real e-commerce and social media settings, with two tasks:
- Personalized Scene Generation: place a given object in a scene reflecting the user's preferences and lifestyle.
- Personalized Creative Generation: create a new image on a topic faithful to the user's aesthetic.
- Multi-axis evaluation: target fidelity, visual quality, user distinguishability, semantic alignment with history, task utility.
Method
PEARL couples a multimodal reasoner with a frozen image generator in an interleaved reason-reflect loop optimized with differential data reward, averaging +15% across personalization metrics over strong baselines.
More from Multimodal
- HuST Lab's Multimodal Flow: Fully Continuous Unified Language-Vision Generation — hustvl · 2026-10-02
- Tencent Hunyuan's Tex-Zero Trains 3D Texture Generation Without Any 3D Assets — Tencent-Hunyuan · 2026-10-02
- One prompt gets Claude Fable 5.5 to render a AAA-grade 3D Rube Goldberg machine — imjustnewatai · 2026-10-02
- Seedance 2.5 Video Goes Viral for Ultra-Realistic Early-2000s DV Camcorder Look — SimplyAnnisa · 2026-10-02
- ComfyUI Clips + sync.so Lip-Sync: Edit Before or After Processing? — Positive_Society1876 · 2026-10-02
- Dolly or Zoom? Prompting AI Video for Camera-Only Motion on a Static Car — Classic-Jellyfish-26 · 2026-10-02