Deep hybrid models: infer and plan in a dynamic world

2024-02-01

AI Summary

Planning for complex tasks involving multiple entities and hierarchical relational structures in dynamic environments remains challenging. Method: This paper proposes a novel planning framework grounded in active inference, unifying perception and action within a joint Bayesian inference process. We introduce the first deep hybrid active inference model, integrating body configuration representations, scalable hierarchical body schemas, and multi-timescale intention trajectory inference—combined with variational Bayesian inference, hierarchical generative models, and continuous-discrete hybrid state representations. Crucially, we incorporate dynamic tool-use modeling to enable online inference of tool–goal coupling relationships. Contribution/Results: Evaluated on a composite task—mobile robotic tool grasping and dynamic target tracking—the framework significantly outperforms conventional optimal control baselines. Results demonstrate its effectiveness and robustness in cross-timescale planning and perception–action co-inference, validating the theoretical and practical advantages of active inference for embodied intelligence in non-stationary settings.

Technology Category

• Humans and AI: Human-Aware Planning and Behavior Prediction
• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Planning, Routing, and Scheduling: Activity and Plan Recognition

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

To determine an optimal plan for complex tasks, one often deals with dynamic and hierarchical relationships between several entities. Traditionally, such problems are tackled with optimal control, which relies on the optimization of cost functions; instead, a recent biologically-motivated proposal casts planning and control as an inference process. Active inference assumes that action and perception are two complementary aspects of life whereby the role of the former is to fulfill the predictions inferred by the latter. Here, we present an active inference approach that exploits discrete and continuous processing, based on three features: the representation of potential body configurations in relation to the objects of interest; the use of hierarchical relationships that enable the agent to easily interpret and flexibly expand its body schema for tool use; the definition of potential trajectories related to the agent's intentions, used to infer and plan with dynamic elements at different temporal scales. We evaluate this deep hybrid model on a habitual task: reaching a moving object after having picked a moving tool. We show that the model can tackle the presented task under different conditions. This study extends past work on planning as inference and advances an alternative direction to optimal control.

Problem

• Optimizing complex tasks with dynamic hierarchical relationships
• Replacing optimal control with biologically-inspired inference planning
• Integrating discrete and continuous processing for flexible action perception

Innovation

• Combines discrete and continuous processing for planning
• Uses hierarchical relationships for flexible body schema
• Infers trajectories at multiple temporal scales dynamically