Spatial Atlas: Compute-Grounded Reasoning for Spatial-Aware Research Agent Benchmarks

TL;DR

Spatial Atlas employs compute-grounded reasoning to enhance spatial-aware research agents by combining deterministic spatial computations with large language models, improving accuracy and interpretability.

Contribution

It introduces a novel CGR paradigm implemented as an agent-to-agent server that integrates deterministic spatial reasoning with LLMs for complex benchmarks.

Findings

Achieves competitive accuracy on spatial and ML benchmarks.

Maintains interpretability through structured spatial representations.

Reduces hallucinations by deterministic spatial computations.

Abstract

We introduce compute-grounded reasoning (CGR), a design paradigm for spatial-aware research agents in which every answerable sub-problem is resolved by deterministic computation before a language model is asked to generate. Spatial Atlas instantiates CGR as a single Agent-to-Agent (A2A) server that handles two challenging benchmarks: FieldWorkArena, a multimodal spatial question-answering benchmark spanning factory, warehouse, and retail environments, and MLE-Bench, a suite of 75 Kaggle machine learning competitions requiring end-to-end ML engineering. A structured spatial scene graph engine extracts entities and relations from vision descriptions, computes distances and safety violations deterministically, then feeds computed facts to large language models, thereby avoiding hallucinated spatial reasoning. Entropy-guided action selection maximizes information gain per step and routes queries across a three-tier frontier model stack (OpenAI + Anthropic). A self-healing ML pipeline with strategy-aware code generation, a score-driven iterative refinement loop, and a prompt-based leak audit registry round out the system. We evaluate across both benchmarks and show that CGR yields competitive accuracy while maintaining interpretability through structured intermediate representations and deterministic spatial computations.