Time-Scaling Is What Agents Need Now
Zhi Liu, Guangzhi Wang
TL;DR
The paper addresses how to overcome fixed-parameter limitations in deep reasoning by introducing Time-Scaling, an architectural principle that extends reasoning along extended temporal pathways. It fuses cognitive psychology concepts—Problem Space Theory, problem-solving strategies, inner speech, and metacognition—with AI foundations on temporal and structural credit assignment to argue for temporally extended reasoning in large models. It reviews time-aware reasoning methods (CoT, ToT, DeepSeek-R1) and highlights their limitations, advocating interactive, metacognitive, multi-turn reasoning to dynamically select and orchestrate strategies, potentially paving the way toward advanced cognitive agents. The proposed framework aims to enhance problem-solving depth and adaptability without proportional increases in static model parameters, offering practical implications for building more capable agents and informing the trajectory toward artificial general intelligence.
Abstract
Early artificial intelligence paradigms exhibited separated cognitive functions: Neural Networks focused on "perception-representation," Reinforcement Learning on "decision-making-behavior," and Symbolic AI on "knowledge-reasoning." With Transformer-based large models and world models, these paradigms are converging into cognitive agents with closed-loop "perception-decision-action" capabilities. Humans solve complex problems under limited cognitive resources through temporalized sequential reasoning. Language relies on problem space search for deep semantic reasoning. While early large language models (LLMs) could generate fluent text, they lacked robust semantic reasoning capabilities. Prompting techniques like Chain-of-Thought (CoT) and Tree-of-Thought (ToT) extended reasoning paths by making intermediate steps explicit. Recent models like DeepSeek-R1 enhanced performance through explicit reasoning trajectories. However, these methods have limitations in search completeness and efficiency. This highlights the need for "Time-Scaling"--the systematic extension and optimization of an agent's ability to unfold reasoning over time. Time-Scaling refers to architectural design utilizing extended temporal pathways, enabling deeper problem space exploration, dynamic strategy adjustment, and enhanced metacognitive control, paralleling human sequential reasoning under cognitive constraints. It represents a critical frontier for enhancing deep reasoning and problem-solving without proportional increases in static model parameters. Advancing intelligent agent capabilities requires placing Time-Scaling principles at the forefront, positioning explicit temporal reasoning management as foundational.