Modulation of temporal decision-making in a deep reinforcement learning agent under the dual-task paradigm

2025-11-03

AI Summary

This study investigates the mechanisms by which artificial intelligence agents simulate human temporal perception biases—specifically, time overestimation—under dual-task paradigms involving concurrent time production and numerical comparison. Method: Leveraging a simplified Overcooked environment, we develop single- and dual-task deep reinforcement learning frameworks, employing LSTM networks to model temporal dynamics. Agents are trained to produce target durations while simultaneously performing digit magnitude judgments. Contribution/Results: Under dual-task conditions, agents consistently overproduce time across diverse target durations—a robust, human-like bias absent in single-task baselines. Crucially, no explicit neural representation of an internal clock emerges in the agent’s recurrent states, suggesting that this temporal distortion arises not from dedicated timing circuitry but as an emergent property of strategic task interference. To our knowledge, this is the first demonstration and mechanistic analysis of dual-task temporal interference in deep RL agents, offering novel computational evidence and methodological pathways for modeling embodied time perception.

Technology Category

• Multiagent Systems: Agent-Based Simulation and Emergent Behavior
• Humans and AI: Human-Aware Planning and Behavior Prediction
• Cognitive Modeling & Cognitive Systems: Simulating Human Behavior

Application Category

• Economics, Online Markets and Human Computation: Research challenges in human and human-AI computation
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

This study explores the interference in temporal processing within a dual-task paradigm from an artificial intelligence (AI) perspective. In this context, the dual-task setup is implemented as a simplified version of the Overcooked environment with two variations, single task (T) and dual task (T+N). Both variations involve an embedded time production task, but the dual task (T+N) additionally involves a concurrent number comparison task. Two deep reinforcement learning (DRL) agents were separately trained for each of these tasks. These agents exhibited emergent behavior consistent with human timing research. Specifically, the dual task (T+N) agent exhibited significant overproduction of time relative to its single task (T) counterpart. This result was consistent across four target durations. Preliminary analysis of neural dynamics in the agents' LSTM layers did not reveal any clear evidence of a dedicated or intrinsic timer. Hence, further investigation is needed to better understand the underlying time-keeping mechanisms of the agents and to provide insights into the observed behavioral patterns. This study is a small step towards exploring parallels between emergent DRL behavior and behavior observed in biological systems in order to facilitate a better understanding of both.

Problem

• Investigating temporal decision-making interference in dual-task AI paradigms
• Analyzing time production differences between single-task and dual-task agents
• Exploring neural mechanisms underlying timing behaviors in deep reinforcement learning

Innovation

• Deep reinforcement learning agents trained in dual-task paradigm
• LSTM neural dynamics analyzed for time-keeping mechanisms
• Overcooked environment used to study temporal decision-making interference