AI 용어집
인공지능 완전 사전
Voicebot
Automated conversational agent designed to interact with users through voice commands, using speech recognition and synthesis to simulate a natural human conversation.
Automatic Speech Recognition (ASR)
Technology that converts spoken speech into written text, constituting the essential first step of voice processing by a voicebot to understand user queries.
Text-to-Speech (TTS)
Process of converting text into artificial speech, allowing a voicebot to communicate oral responses to the user with natural intonation and rhythm.
Natural Language Understanding (NLU)
Subset of NLP specialized in understanding the intent and meaning of user utterances, enabling the voicebot to correctly interpret voice requests.
Wake Word Detection
Ability of a voicebot to remain on standby and activate only when it hears a specific keyword, thus optimizing resource usage and privacy.
Background Noise
Irrelevant environmental sounds that ASR systems must filter for accurate speech recognition, a major challenge in real-world voicebot deployments.
Speaker Verification
Biometric process that confirms a user's identity based on the unique characteristics of their voice, enhancing the security of sensitive voice transactions.
Dialogue Management System
Central component of a voicebot that maintains conversation context, manages turn-taking, and determines the system's next action based on the dialogue state.
Proactive Response
Ability of a voicebot to anticipate user needs and initiate actions or suggestions without waiting for an explicit command, improving the user experience.
CTI Integration
Interface between the voicebot and the enterprise telephony system, enabling actions such as call transfer to a human agent or real-time customer data consultation.
Acoustic Model
Component of the ASR that maps audio signals to phonetic units, trained on vast speech corpora to recognize pronunciation variations and accents.
Audio Front-End
Set of processing applied to the raw audio signal before speech recognition, including noise reduction, equalization, and dereverberation to improve input quality.
Response Repository
Structured database containing pre-recorded or dynamically generated voice responses by the voicebot, organized by intent and conversational context.
Vocal Sentiment Analysis
Technique that evaluates user emotions (joy, anger, frustration) from the prosodic characteristics of their voice, allowing the voicebot to adapt its behavior.
Barge-in
Feature allowing a user to interrupt the voicebot's speech to ask a question or give an instruction, making the interaction more natural and efficient.
Endpointing
Process of detecting the end of a spoken utterance, crucial for the voicebot to know when the user has finished their sentence and it can begin processing.
Voice User Interface (VUI)
Interaction paradigm centered on voice, where the interface is entirely designed for oral commands and responses, without visual or tactile components.
Speech-to-Intent
Approach that directly maps speech to user intent, bypassing explicit text transcription for faster responses in voicebots.