Alibaba-NLP CORE: boosting compositional reasoning in MLLM embeddings via reranker distillation
Alibaba-NLP · hf · 2026-09-04
Alibaba-NLP introduces CORE, an approach to improve compositional reasoning in MLLM embeddings.
- Method: distills compositional ranking judgments from a cross-attentive reranker into an embedding model
- Training uses synthesized multi-level candidates and a Rank-KL objective
- Result: improves compositional retrieval without degrading standard retrieval performance
More from Research
- FlashRender: Few-step camera-controlled generative rendering via MeanFlow distillation — everex · 2026-09-04
- LatentStream: progressive latent memory evolution for streaming video understanding — Hongyu Qu · 2026-09-04
- Temporal Context Routing aligns script timing in joint audio-video generation — Yichen Liu · 2026-09-04
- Puffin-World scales unified multimodal model with native 3D world states — Kang Liao · 2026-09-04
- Agent reliability degrades with trajectory length, but no benchmark isolates it, dev finds — rio_ARC · 2026-09-04
- The Bayes Bandit: A Mathematical Take on Curiosity in Reinforcement Learning — CatAstro_Piyush · 2026-09-04