Research map / Label-efficient and robust learning / Recognition and long-tail
Connectionist Temporal Classification with Maximum Entropy Regularization
Hu Liu, Sheng Jin, Changshui 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 maximum conditional entropy regularization (EnCTC) (ablation) | CTC (CRNN) · Synth5K | sequence accuracy: 38.1 -> 45.5 | Improved |
| add equal spacing path pruning (EsCTC) (ablation) | CTC (CRNN) · Synth5K | sequence accuracy: 38.1 -> 46.3 | Improved |
| combine entropy regularization and equal spacing pruning (ablation) | EnCTC · Synth5K | sequence accuracy: 45.5 -> 47.2 | Improved |
| entropy over feasible paths (EnCTC) instead of per-time-step confidence penalty | CTC + CP · Synth5K | sequence accuracy: 44.4 -> 45.5 | Improved |
| EnCTC vs label smoothing | CTC + LS · Synth5K | sequence accuracy: 42.9 -> 45.5 | Improved |
| add EnCTC to CRNN | CRNN · IC13 | sequence accuracy: 86.7 -> 90.0 | Improved |
| add EsCTC to CRNN | CRNN · IC03 | sequence accuracy: 89.4 -> 92.6 | Improved |
| add EnEsCTC to CRNN | CRNN · SVT | sequence accuracy: 80.8 -> 80.6 | No clear effect |
| add EnCTC to CRNN | CRNN · IIIT5K | sequence accuracy: 78.2 -> 82.6 | Improved |
| add equal spacing pruning to EnCTC (ablation) | EnCTC · IC13 | sequence accuracy: 90.0 -> 90.6 | Mixed |
Closest papers by meaning
- CR-CTC: Consistency regularization on CTC for improved speech recognition (ICLR 2025 · similarity 0.90)
- Align With Purpose: Optimize Desired Properties in CTC Models with a General Plug-and-Play Framework (ICLR 2024 · similarity 0.89)
- Self-Distillation Regularized Connectionist Temporal Classification Loss for Text Recognition: A Simple Yet Effective Approach (AAAI 2024 · similarity 0.87)
- Uncertainty-Aware Self-Training for CTC-Based Automatic Speech Recognition (AAAI 2025 · similarity 0.81)
- Incremental Sequence Classification with Temporal Consistency (NeurIPS 2025 · similarity 0.80)
- Structural Entropy Guided Probabilistic Coding (AAAI 2025 · similarity 0.80)
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.
