Chatbot Context via Device Sensor Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Interactive chatbots lack the ability to provide contextually relevant and emotionally resonant responses, relying primarily on pre-configured statements or natural language input, which limits their engagement and understanding of user interactions.
Innovation Solution
Implementing a large language model (LLM) that processes sensor data, including non-acoustic inputs from devices like ambient light and accelerometer sensors, to generate contextually relevant and emotionally expressive responses, allowing the chatbot to adapt its output and virtual character's behavior in real-time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If interactive chatbots rely on pre-configured statements and natural language input only, then the system complexity remains low, but the contextual relevance and emotional resonance of responses deteriorate
Solution Approach 1:
The patent combines multiple data sources including sensor data, user input, and contextual information into a unified processing framework. The system merges structured sensor readings with unstructured natural language input, allowing the chatbot to generate responses that are both contextually relevant and emotionally resonant while managing complexity through integrated processing.
Solution Approach 2:
The chatbot system is designed to handle multiple types of input simultaneously - acoustic sensor data, non-acoustic sensor data, and natural language input - through a single unified processing architecture. This multi-functional capability allows the system to adapt to various interaction scenarios without requiring separate processing pipelines for each input type.
2Adaptability or versatility
If the chatbot processes only natural language input, then the processing speed is fast, but the emotional resonance and contextual understanding deteriorate
Solution Approach 1:
The system performs preliminary processing of sensor data and contextual information before the main conversation occurs. By pre-processing and structuring sensor inputs in advance, the chatbot can quickly integrate this information with natural language input during actual interactions, reducing processing time while maintaining emotional resonance and contextual understanding.
Solution Approach 2:
The patent replaces traditional rule-based response generation with a large language model that can process and integrate multiple input types simultaneously. This substitution allows the system to handle complex sensor data and natural language input in parallel, maintaining fast processing speeds while improving emotional resonance through the LLM's ability to understand nuanced contexts.
3Productivity
If the chatbot provides detailed contextually relevant responses, then user engagement improves, but the conversation length and time required to conclude interactions increases
Solution Approach 1:
The system incorporates feedback loops where the chatbot monitors user responses and adjusts the level of detail in subsequent messages. By analyzing user engagement signals and conversation flow, the chatbot can provide sufficiently detailed responses that maintain user engagement while avoiding unnecessary elaboration that would extend conversation duration.
Solution Approach 2:
The chatbot employs a strategy of providing just enough contextual information to maintain engagement without over-explaining. By calibrating the level of detail to match user needs and conversation context, the system achieves high user engagement while keeping interactions concise and efficient.
Data Source
AI summary
Implementations relate to processing, utilizing a large language model (“LLM”), input that is based on sensor data, from sensor(s) of a client device, to generate LLM output—and causing output, that is based on the generated LLM output, to be rendered by an interactive chatbot. The input that is based on sensor data and that is processed by the LLM in generating the LLM output can be, or can include, non-acoustic input based on non-acoustic sensor data. For example, an instance of LLM output can be generated based on processing of non-acoustic input using the LLM and without any processing of acoustic input (that is based on acoustic sensor data) using the LLM. As another example, an instance of LLM output can be generated based on processing, using the LLM, both non-acoustic input that is based on non-acoustic data and acoustic input that is based on acoustic sensor data.