Milvus tests hyperbolic embeddings in image search rankings
The geometry matters less than the query type: hyperbolic scoring helps descriptive captions but hurts broad category queries where rare taxonomy terms dominate the gap. Teams running image search should benchmark Lorentz against cosine on their own query logs instead of assuming hyperbolic is uniformly better. Milvus supports hyperbolic ranking through a standard inner product, so no new infrastructure is required to test it.
Milvus benchmarked Hyper3-CLIP V1 against OpenAI CLIP and compared Lorentz versus cosine scoring on the same embeddings. On 50,000 ImageNet validation images, Lorentz lifted V1's classification accuracy from 50.14% to 51.06%. Across 132 broad category searches like "bird", Lorentz trailed cosine by 2.2 average precision points, while CLIP led by 6.4 points on those searches. V1 performed better on descriptive captions, such as separating "a dog on a sofa" from "a sofa on a dog", and led CLIP in a smaller three-concept COCO test. Milvus notes hyperbolic ranking can run with a standard inner product.
We tested how hyperbolic embeddings change image search rankings. First, we compared Hyper3-CLIP V1 with OpenAI CLIP. Then we scored the same V1 embeddings with Lorentz (full hyperbolic geometry) and cosine (direction only) to isolate the effect of the geometry. On 50,000 ImageNet validation images, Lorentz raised V1’s classification accuracy from 𝟱𝟬.𝟭𝟰% 𝘁𝗼 𝟱𝟭.𝟬𝟲%. But across 132 broad category searches such as “bird,” it trailed cosine by 𝟮.𝟮 𝗮𝘃𝗲𝗿𝗮𝗴𝗲 𝗽𝗿𝗲𝗰𝗶𝘀𝗶𝗼𝗻 𝗽𝗼𝗶𝗻𝘁𝘀. Its remaining classification errors did not land closer to the right category either. Against CLIP, V1 did better at finding the right image for a descriptive caption and at distinguishing “a dog on a sofa” from “a sofa on a dog.” CLIP led by 6.4 average precision points on our ImageNet broad category searches, with rare taxonomy terms accounting for much of the gap. In a smaller COCO test covering three broad concepts, V1 led instead. Our takeaway: Hyper3-CLIP V1 is worth testing when users search for images with descriptive captions. For broad category search, evaluate the words your users actually type and compare Lorentz with cosine on your own data. Milvus can run hyperbolic ranking with a standard inner product milvus.io/blog/what-hype… xperiment: https://t.co/UWlkSaC3Tw 💬 0 🔄 0 ❤️ 0 👀 190 ⚡
