Google

A common task in natural language processing is to compute the semantic similarity between two given sentences. This involves analyzing two pieces of text, such as 'Here is a sentence' and 'Here is another sentence', to determine the degree to which they convey the same meaning.

Example of a Semantic Similarity Task

A system is designed to measure the degree to which two sentences convey the same meaning. Which of the following pairs of sentences would the system rate as having the highest degree of shared meaning?

A system designed to measure the degree of shared meaning between two sentences on a scale of 0 (no similarity) to 1 (perfectly equivalent) analyzes the following pair: 

Sentence A: 'The delivery driver was late because of heavy traffic.'
Sentence B: 'The package arrived behind schedule due to road congestion.'

The system outputs a similarity score of 0.95. In your own words, critique this score. Is it an appropriate measure of the shared meaning between the two sentences? Justify your reasoning by identifying specific similarities and differences in the text.

Evaluating a Semantic Similarity Score

Based on the principle of analyzing sentences for shared meaning, which department would the automated system most likely route the customer's question to? Justify your choice by explaining why that department's description is the most semantically similar to the customer's query.

Learn Before

Related