Skip to Main Content
Table 1 

A road map of the various elements that affect MT optimization.

Which Loss Functions? Which Optimization Algorithm? 
Error (§3.1) Minimum Error Rate Training (§5.1) 
Softmax (§3.2) Gradient-based Methods (§5.2, §6.5) 
Risk (§3.3) Margin-based Methods (§5.3) 
Margin, Perceptron (§3.4) Linear Regression (§5.4) 
Ranking (§3.5) Perceptron (§6.2) 
Minimum Squared Error (§3.6) MIRA (§6.3) 
 AROW (§6.4) 
 
Which Evaluation Measure? Which Hypotheses to Target? 
Corpus-level, Sentence Level (§2.5) k-best vs. Lattice vs. Forest (§2.4) 
BLEU and Approximations (§2.5.1, §2.5.2) Merged k-bests (§5) 
Other Measures (§8.3) Forced Decoding (§2.4), Oracles (§4) 
 
Other Topics: 
Large Data Sets (§7), Non-linear Models (§8.1), 
Domain Adaptation (§8.2), Search and Optimization (§8.4) 
Which Loss Functions? Which Optimization Algorithm? 
Error (§3.1) Minimum Error Rate Training (§5.1) 
Softmax (§3.2) Gradient-based Methods (§5.2, §6.5) 
Risk (§3.3) Margin-based Methods (§5.3) 
Margin, Perceptron (§3.4) Linear Regression (§5.4) 
Ranking (§3.5) Perceptron (§6.2) 
Minimum Squared Error (§3.6) MIRA (§6.3) 
 AROW (§6.4) 
 
Which Evaluation Measure? Which Hypotheses to Target? 
Corpus-level, Sentence Level (§2.5) k-best vs. Lattice vs. Forest (§2.4) 
BLEU and Approximations (§2.5.1, §2.5.2) Merged k-bests (§5) 
Other Measures (§8.3) Forced Decoding (§2.4), Oracles (§4) 
 
Other Topics: 
Large Data Sets (§7), Non-linear Models (§8.1), 
Domain Adaptation (§8.2), Search and Optimization (§8.4) 
Close Modal

or Create an Account

Close Modal
Close Modal