Research map / Learning theory and algorithms / Combinatorial and search optimization
A Credit Assignment Compiler for Joint Prediction
Kai-Wei Chang, He He, Stephane Ross, Hal Daume III, John Langford
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 the L2S dependency parser instead of the strongest reported system | L2S dependency parser · CoNLL-X language benchmarks, PTB, and CTB | UAS | Mixed |
| Add the parser's morphological features and Brown clusters where used, rather than using the feature-limited L2S variant (ablation) | L2S dependency parser · CoNLL-X, PTB, and CTB | UAS | Improved |
| Use memoization during training (ablation) | LOLS and Searn · NER and POS | training runtime reduction: about 25% on NER; about 70% on POS | Improved |
| Stop rollouts early using forced path collapse (ablation) | LOLS and Searn · POS and NER | training speed: factor of as much as 9 | Improved |
| Combine Collapse@2 with memoization (ablation) | LOLS and Searn · NER | prediction performance | Hurt |
| Use BILOU instead of BIO encoding (ablation) | NER sequence labeling · CoNLL 2003 NER | F-measure | Hurt |
Closest papers by meaning
- Hindsight Network Credit Assignment: Efficient Credit Assignment in Networks of Discrete Stochastic Units (AAAI 2022 · similarity 0.85)
- Generalization Guarantees for Learning Score-Based Branch-and-Cut Policies in Integer Programming (NeurIPS 2025 · similarity 0.82)
- Reciprocal Learning (NeurIPS 2024 · similarity 0.82)
- Retrospective In-Context Learning for Temporal Credit Assignment with Large Language Models (NeurIPS 2025 · similarity 0.82)
- Contrastive Predict-and-Search for Mixed Integer Linear Programs (ICML 2024 · similarity 0.82)
- Apollo-MILP: An Alternating Prediction-Correction Neural Solving Framework for Mixed-Integer Linear Programming (ICLR 2025 · 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.
