AI Paper Atlas

Research map / Vision-language and multimodal / Cross-modal learning

Deep Global-sense Hard-negative Discriminative Generation Hashing for Cross-modal Retrieval

Kun Cheng, Qibing Qin, Wenfeng Zhang, Lei Huang, Jie Nie

ICLR 2026 · Vision-language and multimodal › Cross-modal learning · approach: Cross-modal hashing

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add the three generation losses (interpolation similarity, semantic preservation, parameter diversification) together (ablation)DGHDGH · MIRFLICKR-25K, avg over 16-128 bits I2TmAP (%): 81.64 -> 86.43Improved
add the three generation losses together (ablation)DGHDGH · MIRFLICKR-25K, avg over 16-128 bits T2ImAP (%): 80.04 -> 84.52Improved
use all three losses instead of the best two-loss combination (84.44 at 16 bits) (ablation)DGHDGH · MIRFLICKR-25K I2T 16 bitsmAP (%): 84.44 -> 84.66Improved
use DGHDGH instead of DDBH, I2T, 128 bitsDGHDGH · NUS-WIDE I2T 128 bitsmAP@all (%): 72.29 -> 73.76Improved
use DGHDGH instead of DDBH, I2T, 128 bitsDGHDGH · MS COCO I2T 128 bitsmAP@all (%): 78.24 -> 79.19Improved
use DGHDGH instead of DDBH, T2I, 16 bitsDGHDGH · MS COCO T2I 16 bitsmAP@all (%): 71.67 -> 71.16Hurt
use DGHDGH instead of DDBH under noise 0.8, 16 bitsDGHDGH · MIRFLICKR-25K I2T 16 bits noise 0.2mAP (%): 83.79 -> 81.95Hurt
use DGHDGH instead of DDBH under label noise 0.8, 128 bitsDGHDGH · MIRFLICKR-25K I2T 128 bits noise 0.8mAP (%): 80.29 -> 84.16Improved
swap backbone from CLIP ViT-B/32 to CLIP ViT-L/14DGHDGH · MIRFLICKR-25K I2T 16 bitsmAP (%): 84.66 -> 84.82No clear effect
swap backbone from CLIP ViT-B/32 to SigLIP2 Large-16/256DGHDGH · MIRFLICKR-25K I2T 16 bitsmAP (%): 84.66 -> 84.50No clear effect
swap backbone from CLIP ViT-B/32 to BLIP I-C-BaseDGHDGH · MIRFLICKR-25K I2T 16 bitsmAP (%): 84.66 -> 82.96Hurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy