Research map / Label-efficient and robust learning / Noisy labels
CogniTrust: Cognitive Memory-Driven Verifiable Supervision for Robust Hashing
Yiyang Gu, Bohan Wu, Yifang Qin, Jiaru Tang, Rong-Cheng Tu, Zhiping Xiao, Taian Guo, Junyu Luo, Wei Ju, Xiao Luo, Dacheng Tao, Ming Zhang
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 changed | Where | Measured | Effect |
|---|---|---|---|
| add episodic verifiability score (ablation) | CogniTrust · Flickr25K 64 bits pairflip | MAP: 72.72 -> 74.62 | Improved |
| add semantic verifiability score (ablation) | CogniTrust · Flickr25K 64 bits pairflip | MAP: 71.89 -> 74.62 | Improved |
| add combined EVS and SVS (ablation) | CogniTrust · CIFAR-10 32 bits pairflip | MAP: 69.55 -> 70.91 | Improved |
| add diffusion-based mixup (ablation) | CogniTrust · CIFAR-10 32 bits pairflip | MAP: 68.03 -> 70.91 | Improved |
| add diffusion-based mixup (ablation) | CogniTrust · Flickr25K 64 bits pairflip | MAP: 70.53 -> 74.62 | Improved |
| add episodic score in Flickr25K symmetric 32 bits (ablation) | CogniTrust · Flickr25K 32 bits symmetric | MAP: 71.39 -> 71.59 | No clear effect |
| set mixup alpha to 0.4 rather than 0.1 or 10.0 (ablation) | CogniTrust · Flickr25K/NUS-WIDE 64 bits | MAP | Improved |
| use 5 structural factors rather than 3 or 20 (ablation) | CogniTrust · Flickr25K/NUS-WIDE 64 bits | MAP | Improved |
| use CogniTrust instead of STAR (strongest baseline) | VGG-16 hashing · NUS-WIDE 128 bits pairflip noise 60% | MAP: 58.17 -> 65.27 | Improved |
| use CogniTrust instead of STAR (strongest baseline) | VGG-16 hashing · MS COCO 128 bits symmetric noise 60% | MAP: 42.79 -> 46.92 | Improved |
Closest papers by meaning
- Robust Self-Paced Hashing for Cross-Modal Retrieval with Noisy Labels (AAAI 2025 · similarity 0.86)
- On the Recall Scaling Laws in Mamba: A Theoretical and Mechanistic Study via Hashing (NeurIPS 2026 · similarity 0.85)
- Exploiting Descriptive Completeness Prior for Cross Modal Hashing with Incomplete Labels (NeurIPS 2024 · similarity 0.85)
- Dual Self-Paced Cross-Modal Hashing (AAAI 2024 · similarity 0.84)
- Label-Agnostic Forgetting: A Supervision-Free Unlearning in Deep Models (ICLR 2024 · similarity 0.84)
- An Analysis of Concept Bottleneck Models: Measuring, Understanding, and Mitigating the Impact of Noisy Annotations (NeurIPS 2025 · similarity 0.84)
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.
