Table 2 Image retrieval results
From: Task-agnostic representation learning of multimodal twitter data for downstream applications
Method | Recall @ 1 (%) | Recall @ 5 (%) | Recall @ 10 (%) | Median rank |
|---|---|---|---|---|
VSE++ | 0.02 | 0.1 | 0.18 | 2500 |
VSE++ with hashtags as text | 0.02 | 0.16 | 0.34 | 2329 |
Proposed method | 0.04 | 0.24 | 0.36 | 2213 |