Why Translation Candidates Receive Scores
Question: In a machine translation system that uses heuristic search, why does the system compute a score, such as Score_C(E), for a candidate output?
Sample answer: The score gives each possible translation a numerical quality estimate so the system can compare options and choose the most promising one.
Key points:
- Assigns a score to each candidate translation.
- Helps rank alternatives by how well they fit the source sentence or how likely they are to be correct.
Rubric: The answer should say that the score is used to evaluate, compare, rank, or estimate the quality/probability of candidate translations.
0
1
Tags
Data Science
Foundations of Large Language Models Course
Computing Sciences
D2L
Dive into Deep Learning @ D2L
Machine Learning
Deep Learning
Supervised Learning
Machine Learning Strategy
Machine Learning Yearning @ DeepLearning.AI
Related
Diagnosing a decoding error
Can Exhaustive Search Handle Very Large Route Spaces?
A large candidate space often requires _____ Search.
Match the score comparison to the likely cause.
Order the steps for diagnosing a bad recommendation in a score-based selection system.
Why Compare Scores During Translation Error Diagnosis?
Finding a Fault in a Translation Score
Why Translation Candidates Receive Scores
Where Did the Ranking Error Come From?
Search Procedure Failure Condition