Toward Ultra-Long-Horizon Agentic Science: Cognitive Accumulation for Machine Learning Engineering

TL;DR

This paper introduces ML-Master 2.0, an autonomous agent designed for ultra-long-horizon machine learning engineering, employing hierarchical cognitive caching to sustain strategic coherence over days or weeks, surpassing previous short-horizon models.

Contribution

It proposes Hierarchical Cognitive Caching, a novel multi-tiered architecture that enables long-term experience accumulation and strategic planning in autonomous AI agents.

Findings

Achieved a 56.44% medal rate on OpenAI's MLE-Bench within 24 hours.

Demonstrated effective long-term knowledge consolidation beyond static context limits.

Showed that ultra-long-horizon autonomy enables scalable scientific exploration.

Abstract

The advancement of artificial intelligence toward agentic science is currently bottlenecked by the challenge of ultra-long-horizon autonomy, the ability to sustain strategic coherence and iterative correction over experimental cycles spanning days or weeks. While Large Language Models (LLMs) have demonstrated prowess in short-horizon reasoning, they are easily overwhelmed by execution details in the high-dimensional, delayed-feedback environments of real-world research, failing to consolidate sparse feedback into coherent long-term guidance. Here, we present ML-Master 2.0, an autonomous agent that masters ultra-long-horizon machine learning engineering (MLE) which is a representative microcosm of scientific discovery. By reframing context management as a process of cognitive accumulation, our approach introduces Hierarchical Cognitive Caching (HCC), a multi-tiered architecture inspired by computer systems that enables the structural differentiation of experience over time. By dynamically distilling transient execution traces into stable knowledge and cross-task wisdom, HCC allows agents to decouple immediate execution from long-term experimental strategy, effectively overcoming the scaling limits of static context windows. In evaluations on OpenAI's MLE-Bench under 24-hour budgets, ML-Master 2.0 achieves a state-of-the-art medal rate of 56.44%. Our findings demonstrate that ultra-long-horizon autonomy provides a scalable blueprint for AI capable of autonomous exploration beyond human-precedent complexities.