References - Chatbot - The Chatbot Explained

Technical Overview: Custom Chatbot Engine

Our custom-built chatbot leverages a rich dictionary, natural language parsing, dynamic memory management, and specialized external API integrations to create fluid, human-like conversations.

Core Components & Features

1. Rich 142,000+ Word Dictionary

At the heart of the chatbot is a comprehensive dictionary containing over 142,694 words spanning nouns, verbs, adjectives, auxiliaries, and pronouns. Each entry includes detailed linguistic attributes:

• Word Type: Noun, verb, adjective, etc.

• Gender: Masculine / Feminine

• Number: Singular / Plural

• Tense: Present, Past, Future

2. Sentence Parsing & Processing

When a user submits input, the system breaks it down into structured data:

• Tokenization: Sentences are split by periods, then further divided into words by spaces.

• Data Lookup: Each word is matched against the dictionary to extract its grammatical properties. Adjectives are mapped against a synonym table to expand vocabulary.

• Perspective Shift: Dynamic pronouns and verbs are inverted automatically (e.g., transforming "I am" to "You are" and vice versa).

Memory & Response Generation

Memory Querying

The system stores user keyword objects, syntax patterns, and context in its conversation database:

• Flexible Matching: User verbs are evaluated across all tenses, and nouns are expanded using synonyms.

• Logical Association: Keywords use OR conditions within word types and AND conditions across word types to identify the best contextual response.

• Proportional Memory: Short inputs match short-memory responses, while detailed inputs trigger long-memory responses.

Dynamic Sentence Generation

If no existing memory match is found, the engine builds a new sentence from scratch using structured grammar rules:

• Words are organized dynamically based on gender, number, tense, and sentence structure patterns.

• The system validates the syntax before returning a natural-sounding response to the user.

Advanced Context & Memory Systems

• Topic Tracking & Anti-Repetition: When a subject is discussed, its keywords are cached in a temporary session to avoid repeating the same response. After a few turns, the topic counter resets.

• Short-Term Memory & Questions: When the chatbot asks a question, it captures and reformulates the user’s next answer into short-term memory. If the user changes topics completely, the short-term memory array automatically flushes.

• Quality Filtering: If the user’s input is negative immediately after a new memory row is created, the system deletes that row as a safety measure to prevent saving invalid or undesirable responses.

Specialized Features & Integrations

Google Natural Language & Wikipedia Integration

When enabled, unmatched proper nouns (such as company or brand names) are identified via the Google Natural Language API and looked up on Wikipedia.

• Knowledge Retrieval: Direct queries like "What is..." or "Do you know..." prompt the bot to answer with the fetched Wikipedia summary.

• Autonomous Learning (Cron Task): Every minute, a background cron job uses these stored descriptions to help the bot build independent associations and expand its memory (requires a minimum of 30 stored memory rows to make meaningful links).

Unit Conversions & Mathematics

The bot features an internal unit conversion matrix to solve basic arithmetic (addition, subtraction, multiplication, division).

• Supported Units: Liters, milliliters, ounces, centimeters, meters, millimeters, kilometers, feet, inches, square units, and cubic units.

• It automatically converts mixed units to a matching base unit before calculating or converting between distinct units.

Interactive Animated Avatar

The front-end face is powered by a custom Python application that renders geometric grid shapes (0s and 1s) to express emotions:

• Includes built-in emotional facial expressions mapped to specific user keywords.

• Plays animations whenever a keyword is saved to short-term memory, repeating until the current subject is flushed.

Source : learningbot

Price: $ 00.00 CAD to access learningbot extensions, forum support, chatbox support and update support