Research map / Time series / Long-term forecasting
Dynamic Multi-Context Attention Networks for Citation Forecasting of Scientific Publications
Taoran Ji, Nathan Self, Kaiqun Fu, Zhiqian Chen, Naren Ramakrishnan, Chang-Tien Lu
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 |
|---|---|---|---|
| Use DMA-Nets instead of PC-RNN for citation-time forecasting | citation forecasting · USPTO, 30% observations | MAE: 132.2665 -> 131.0907 | Improved |
| Use DMA-Nets instead of DotSeq2Seq for citation-time forecasting | citation forecasting · MAG, 30% observations | MAE: 47.0133 -> 42.2267 | Improved |
| Use DMA-Nets instead of PC-RNN for category prediction | citation forecasting · USPTO, 80% observations | ACC: 0.8121 -> 0.8570 | Improved |
| Use DMA-Nets instead of CYAN-RNN for category prediction | citation forecasting · USPTO, 30% observations | ACC: 0.8431 -> 0.8444 | Improved |
| Use DMA-Nets instead of PC-RNN for category prediction | citation forecasting · MAG, 80% observations | ACC: 0.6720 -> 0.6647 | Hurt |
| Use DMA-Nets instead of CYAN-RNN for category prediction | citation forecasting · MAG, 30% observations | ACC: 0.6390 -> 0.6425 | Improved |
| Add the GMTA layer (ablation) | DMA-Nets · USPTO, 80% observations | MAE: 58.8086 -> 56.8676 | Improved |
| Add the GMTA layer (ablation) | DMA-Nets · MAG, 80% observations | MAE: 17.4309 -> 17.2212 | Improved |
| Add the LTA layer (ablation) | DMA-Nets · USPTO, 80% observations | MAE: 57.9564 -> 56.8676 | Improved |
| Add the LTA layer (ablation) | DMA-Nets · MAG, 80% observations | MAE: 17.5243 -> 17.2212 | Improved |
| Reduce hidden dimension dh from the base setting (ablation) | DMA-Nets · USPTO, 80% observations | MAE: 56.8676 -> 58.3457 | Hurt |
| Reduce hidden dimension dh from the base setting (ablation) | DMA-Nets · USPTO, 80% observations | ACC: 0.8570 -> 0.8516 | Hurt |
Closest papers by meaning
- HLM-Cite: Hybrid Language Model Workflow for Text-based Scientific Citation Prediction (NeurIPS 2024 · similarity 0.86)
- Context is Key: A Benchmark for Forecasting with Essential Textual Information (ICML 2025 · similarity 0.82)
- Introducing Spectral Attention for Long-Range Dependency in Time Series Forecasting (NeurIPS 2024 · similarity 0.82)
- LLM Processes: Numerical Predictive Distributions Conditioned on Natural Language (NeurIPS 2024 · similarity 0.82)
- TStruct: Learning Shared Temporal Structures for Long-Term Time-Series Forecasting (NeurIPS 2026 · similarity 0.82)
- GPT4MTS: Prompt-based Large Language Model for Multimodal Time-series Forecasting (AAAI 2024 · similarity 0.81)
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.
