Glossary of Terms
Here is a list of key concepts and terms that you will encounter when using Parloa:
A
Agent
A human who is connected to the caller and speaks a different language. An agent assists callers who can't communicate with the bot.
ASR (Automatic Speech Recognition)
A technology that turns spoken words into written text. It helps the bot understand what the caller says by converting their speech to text.
C
Custom Lifetime
A setting that lets you decide how long data is stored before it’s automatically deleted.
D
DTMF (Dual-tone Multi-frequency Signaling)
A method used in used in phones to dial numbers and interact with automated systems by pressing buttons. Each button press generates a unique sound made up of two tones, which the system can recognize.
Duckling
Duckling is a natural language processing (NLP) library that enables the chatbot to interpret the caller's utterances, even if they do not use the same words or phrasing that the chatbot is programmed to recognize. Duckling is able to understand a wide range of natural language constructs, such as dates, times, locations, and quantities.
E
EOS (End Of Speech)
EOS is a signal that indicates that the caller has stopped speaking. Parloa uses EOS to determine when the chatbot should stop listening to the caller's input and start processing it.
G
GDPR (General Data Protection Regulation)
A regulation in EU law on data protection and privacy in the European Union and the European Economic Area. It also addresses the transfer of personal data outside these areas.
H
HMAC (Hash-Based Message Authentication Code)
A type of cryptographic security feature that uses a secret key and a hash function to ensure data integrity and authenticity (such as SHA256).
I
Intent
The purpose what the user wants to achieve. It represents the underlying "what" behind a message. For example, the intent "BookFlight" represents the user's desire to book a flight.
L
LLMs (Large Language Models)
Advanced AI models that can understand and generate human language by analyzing large amounts of text.
M
MLB (Multi-Language Bot)
MLB is a feature that enables users to create Parloa bots in multiple languages. This allows users to create bots that can communicate with users in their native language, regardless of where they are located in the world.
N
NLP (Natural Language Processing)
NLP encompasses automated strategies for dissecting human language into its basic components, which can then be processed and understood by machine learning models.
NLU (Natural Language Understanding)
NLU refers to a model that deduces semantic meaning from texts by applying a set of rules it has 'learned.' First, we employ Natural Language Processing (NLP) techniques to dissect and prepare the text. These processed texts are then utilized to train a model; this model, the NLU, is subsequently tasked with interpreting entirely new texts and determining their underlying significance.
P
R
Response
The action the bot takes after receiving the caller’s input, such as replying with text, playing audio, or making an API call.
S
Session Data Lifecycle
The process of creating, using, and eventually deleting or archiving data related to a session.
sessionid (Session Identifier)
A unique number assigned to each session to maintain the context of interactions.
SHA256 Hashing
A cryptographic hashing method to convert data into a fixed-size string that can't be easily reversed.
SIP (Session Initiation Protocol)
A protocol used to start and manage real-time communication sessions, such as phone calls.
SIP Trunk
A digital method to initiating and receiving calls, along with other forms of digital communication, via an internet connection. In the context of SIP trunking, the 'trunk' signifies virtual phone lines, enabling calls over the internet to any traditional telephone number.
State
The current state of the conversation. It can be used to track things like the caller's progress through the conversation and the information that has been collected from the caller so far.
T
Training Data
A set of labeled examples used to teach the bot’s language model, including both correct and incorrect responses.
U
V
Last updated