Learning Long Short-Term Intention within Human Daily Behaviors

TL;DR

Experimental results indicate that the proposed long short-term intention model can assist robots in comprehending human behavioral patterns over both long-term and short-term durations, which helps determine the consistency between long-term and short-term intentions of humans.

Abstract

In the domain of autonomous household robots, it is of utmost importance for robots to understand human behaviors and provide appropriate services. This requires the robots to possess the capability to analyze complex human behaviors and predict the true intentions of humans. Traditionally, humans are perceived as flawless, with their decisions acting as the standards that robots should strive to align with. However, this raises a pertinent question: What if humans make mistakes? In this research, we present a unique task, termed "long short-term intention prediction". This task requires robots can predict the long-term intention of humans, which aligns with human values, and the short term intention of humans, which reflects the immediate action intention. Meanwhile, the robots need to detect the potential non-consistency between the short-term and long-term intentions, and provide necessary warnings and suggestions. To facilitate this task, we propose a long short-term intention model to represent the complex intention states, and build a dataset to train this intention model. Then we propose a two-stage method to integrate the intention model for robots: i) predicting human intentions of both value-based long-term intentions and action-based short-term intentions; and 2) analyzing the consistency between the long-term and short-term intentions. Experimental results indicate that the proposed long short-term intention model can assist robots in comprehending human behavioral patterns over both long-term and short-term durations, which helps determine the consistency between long-term and short-term intentions of humans.

Figures (3)

• Figure 1: Comparison between intention model-based help and other forms. The top block denotes "no help" scenarios, in which the people could forget important issues. The middle block denotes "direct help" scenarios, in which the people may receive inappropriate help and fail to finish the more important thing. The bottom block denotes the "help via intention model" scenario, in which the robot can infer the long-term intention and warn the short-term mistake, thus providing appropriate service for people.
• Figure 2: The long short-term intention model. The model involves the modeling for value-based long-term intention and the action-based short-term intention. This model allows the robot to decompose the complex human intention into long-term and short-term intentions, thus handling the potential human mistakes when short-term behaviors violate the long-term behavioral patterns.
• Figure 3: System pipeline. The temporal observation data is serialized before being fed into the neural networks. Intentions are singled out for feature encoding and decoding. The output of the system includes the prediction of actions, durations, short-term intentions, and long-term intentions.