Search papers, labs, and topics across Lattice.
This paper introduces MERGED, a distillation framework that efficiently transfers structured reasoning from large vision-language models (VLMs) to a compact 7B-parameter student model, eliminating the need for costly human annotation. By leveraging multiple teacher models to provide both labels and reasoning, MERGED achieves a significant 13.79% improvement in PR-AUC over human-labeled data and outperforms a larger baseline model at a fraction of the cost. The framework allows for rapid adaptation to new relationship definitions in entity resolution, reducing the time required for updates from months to days.
MERGED achieves a 13.79% boost in PR-AUC while being 6x cheaper than larger models, revolutionizing entity resolution in dynamic business environments.
In product entity resolution, relationship definitions constantly evolve with business needs, yet adapting to each change traditionally requires slow, costly human annotation that is often noisy and carries no reasoning. Large vision-language models (VLMs) prompted zero-shot can adapt to a new definition immediately and supply the reasoning that human labels lack, but their cost and latency are prohibitive at production scale. We present MERGED, a distillation framework that transfers not just labels but structured reasoning from large teacher VLMs into a compact 7B-parameter student, requiring no human annotation. Multiple teachers label each product pair and articulate the reasoning behind their decision: agreement pairs supply supervised fine-tuning, while disagreements are resolved by a meta-judge into preference pairs for Direct Preference Optimization. Evaluated against human-labeled ground truth on a multilingual e-commerce dataset, the resulting student improves PR-AUC by 13.79% over the same backbone trained on human labels and surpasses the larger Qwen2.5-32B-VL baseline by 6.32% at 6x lower cost, while also yielding tighter label-reasoning alignment (over 10% above Qwen2.5-32B-VL). Moreover, re-applying MERGED from an existing checkpoint adapts to a new relationship definition with only 10K samples, improving PR-AUC by 6.97% over zero-shot and outperforming from-scratch training. MERGED enables rapid adaptation to evolving relationship definitions, supporting a new one in days rather than months, at a cost and latency suitable for large-scale industrial deployment.
