Research map / Vision-language and multimodal / Human and embodied modeling
TokenUnify: Scaling Up Autoregressive Pretraining for Neuron Segmentation
Yinda Chen, Haoyuan Shi, Xiaoyu Liu, Te Shi, Ruobing Zhang, Dong Liu, Zhiwei Xiong, Feng Wu
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 next-token prediction to random-token pretraining (ablation) | TokenUnify · Wafer4 (Waterz) | VOI: 1.2680 -> 1.1300 | Improved |
| add next-all token prediction to random-token pretraining (ablation) | TokenUnify · Wafer4 (Waterz) | VOI: 1.2680 -> 1.1907 | Improved |
| add next-all token prediction to random + next-token pretraining (ablation) | TokenUnify · Wafer4 (Waterz) | VOI: 1.1300 -> 0.9951 | Improved |
| add next-all token prediction to random + next-token pretraining (ablation) | TokenUnify · Wafer4 (Waterz) | ARAND: 0.0692 -> 0.0509 | Improved |
| use next-token prediction alone instead of random-token prediction alone (ablation) | TokenUnify · Wafer4 (Waterz) | VOI: 1.2680 -> 4.0418 | Hurt |
| add decoder to fine-tuning of Mamba + encoder (ablation) | TokenUnify · Wafer4 (Waterz) | VOI: 1.1065 -> 0.9951 | Improved |
| fine-tune Mamba blocks only rather than encoder only (ablation) | TokenUnify · Wafer4 (Waterz) | VOI: 1.5556 -> 1.1362 | Improved |
| fine-tune Mamba blocks only rather than decoder only (ablation) | TokenUnify · Wafer4 (Waterz) | VOI: 1.5295 -> 1.1362 | Improved |
| use TokenUnify pretraining instead of no pretraining | EMmamba · Wafer4 (Waterz) | VOI: 1.7839 -> 0.9951 | Improved |
| use TokenUnify pretraining instead of MAE pretraining | EMmamba · Wafer4 (Waterz) | VOI: 1.3248 -> 0.9951 | Improved |
| use TokenUnify pretraining instead of MAE pretraining | EMmamba · Wafer4 (LMC) | VOI: 2.0709 -> 1.6521 | Improved |
| use TokenUnify pretraining instead of MAE pretraining | EMmamba · AC3/4 (Waterz) | VOI: 1.1131 -> 0.9918 | Improved |
Closest papers by meaning
- Cross-Dimension Affinity Distillation for 3D EM Neuron Segmentation (CVPR 2024 · similarity 0.89)
- Learning Multimodal Volumetric Features for Large-Scale Neuron Tracing (AAAI 2024 · similarity 0.88)
- $\epsilon$-Seg: Sparsely Supervised Semantic Segmentation of Microscopy Data (NeurIPS 2025 · similarity 0.87)
- FISBe: A Real-World Benchmark Dataset for Instance Segmentation of Long-Range Thin Filamentous Structures (CVPR 2024 · similarity 0.85)
- Efficient Connectivity-Preserving Instance Segmentation with Supervoxel-Based Loss Function (AAAI 2025 · similarity 0.85)
- ActSort: An active-learning accelerated cell sorting algorithm for large-scale calcium imaging datasets (NeurIPS 2024 · similarity 0.85)
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.
