Research map / Graph learning / Knowledge and temporal graphs
GraphBit: Bitwise Interaction Mining via Deep Reinforcement Learning
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 bitwise interaction mining to GraphBit (ablation) | GraphBit · CIFAR-10, 16 bits | mAP (%): 27.95 -> 32.15 | Helped |
| add bitwise interaction mining to GraphBit (ablation) | GraphBit · CIFAR-10, 32 bits | mAP (%): 32.77 -> 36.74 | Helped |
| add bitwise interaction mining to GraphBit (ablation) | GraphBit · CIFAR-10, 64 bits | mAP (%): 36.16 -> 39.90 | Helped |
| add bitwise interaction mining to GraphBit (ablation) | GraphBit · Brown, average across six train/test combinations | 95% error rate (ERR): 32.83 -> 29.75 | Helped |
| add bitwise interaction mining to GraphBit (ablation) | GraphBit · HPatches, verification | mAP (%): 62.32 -> 65.19 | Helped |
| add bitwise interaction mining to GraphBit (ablation) | GraphBit · HPatches, matching | mAP (%): 13.42 -> 14.22 | Helped |
| add bitwise interaction mining to GraphBit (ablation) | GraphBit · HPatches, retrieval | mAP (%): 21.45 -> 25.19 | Helped |
| add J2 mutual-information objective to J1 (ablation) | GraphBit with independent bits · CIFAR-10, 16 bits | mAP (%): 21.45 -> 27.95 | Helped |
| add J2 mutual-information objective to J1 (ablation) | GraphBit with interacted bits · CIFAR-10, 16 bits | mAP (%): 26.20 -> 30.43 | Helped |
| add J3 to J1 + J2 (ablation) | GraphBit with interacted bits · CIFAR-10, 16 bits | mAP (%): 30.43 -> 32.15 | Helped |
| replace DBD-MQ with GraphBit | unsupervised binary descriptors · CIFAR-10, 16 bits | mAP (%): 21.53 -> 32.15 | Helped |
| replace DBD-MQ with GraphBit | unsupervised binary descriptors · CIFAR-10, 32 bits | mAP (%): 26.50 -> 36.74 | Helped |
Closest papers by meaning
- Symmetric Replay Training: Enhancing Sample Efficiency in Deep Reinforcement Learning for Combinatorial Optimization (ICML 2024 · similarity 0.83)
- A Graph is Worth 1-bit Spikes: When Graph Contrastive Learning Meets Spiking Neural Networks (ICLR 2024 · similarity 0.82)
- Learning Latent Structures in Network Games via Data-Dependent Gated-Prior Graph Variational Autoencoders (ICML 2024 · similarity 0.82)
- SiT: Symmetry-invariant Transformers for Generalisation in Reinforcement Learning (ICML 2024 · similarity 0.82)
- DiMa: Understanding the Hardness of Online Matching Problems via Diffusion Models (ICML 2025 · similarity 0.82)
- HGAP: Boosting Permutation Invariant and Permutation Equivariant in Multi-Agent Reinforcement Learning via Graph Attention Network (ICML 2024 · similarity 0.82)
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.