問22
頻出度★★☆
AIの活用領域には音声認識,画像認識,自然言語処理などがある。音声認識と自然言語処理の両方が利用されているシステムの事例として,最も適切なものはどれか。
令和2年度 ITパスポート試験 問22
テクノロジ系基礎理論 › 情報に関する理論
This question's explanation and your answers so far.
解説
問 22
Incorrectア
Detecting the knocking sound is merely a process of picking up a specific sound and does not deal with language, so it is not natural language processing; and recognizing the face is image recognition. Essentially it uses neither speech recognition nor natural language processing.
- イ
正解 Speech recognition is the technology that converts a person’s utterance from sound into text or semantic information, while natural language processing is the technology that interprets the meaning of sentences written in language and performs tasks such as answering or summarizing. In choice イ, questions asked by a person, such as about weather or traffic information, are first received as sound (speech recognition), and then the content of the question is interpreted to generate an appropriate answer (natural language processing). It therefore uses both speech recognition and natural language processing, making it the most appropriate example, typical of a smart speaker or voice assistant.
Incorrectウ
Detecting the figures of people is image recognition, and detecting voices merely captures the presence or absence of sound; what it actually does is count the number of people passing by. Because it does not interpret the content of the utterances as language, natural language processing is not used.
Incorrectエ
Reading the English text with a camera is image recognition (character recognition), and the part that summarizes and translates it into Japanese does fall under natural language processing; however, because it does not take audio as input, speech recognition is not used.
ここがポイント
- Distinguish the different roles of each technology: speech recognition = converting spoken audio into text or information; natural language processing = interpreting the meaning of language and performing tasks such as answering or summarizing.
- The approach of reading, from an example of an AI system, which recognition and processing technologies (speech recognition / image recognition / natural language processing) are combined.
用語