AGEL-Comp: A Neuro-Symbolic Framework for Compositional Generalization in Interactive Agents

arXiv 2026 41.3 method

TLDR

AGEL-Comp is a neuro-symbolic agent architecture using a causal program graph world model and inductive logic programming for compositional generalization in interactive environments.

Reasoning

The paper presents a novel integration of symbolic world models with LLM reasoning, addressing compositional generalization with a clear evaluation in a simulation environment. However, it lacks real-world validation and may face scalability challenges with complex domains.

Read-first score

Read-first score 41.3, weighted from topical fit, citation, graph, method, reproducibility, and recency signals. Original total remains 26.

Recency 6%

100

Uses a gentle age decay so recent papers surface without erasing older foundations. 2026

Citation impact 18%

58.6

Uses OpenAlex-shaped citation metadata as a bibliometric attention signal, separate from paper quality. citation_normalized_percentile=0.58569544

Methodology quality 18%

Screens visible abstract and analysis fields for experiment, dataset, baseline, metric, and limitation evidence. markers=evaluation

Topical relevance 29%

37.1

Uses existing LLM keyword relevance scores normalized to 0-100. world model,world simulator,generative world model,interactive world model,video world model,world dynamics prediction,model-based reinforcement learning world model

Reproducibility 18%

Screens links and visible text for paper, code, dataset, artifact, and repository signals. pdf=True; code=False; dataset=False; markers=none

Citation velocity 12%

Citation velocity estimates citations per publication-year to reduce old-paper bias. velocity=0.00

Field roles

FrontierBridge

Rank sensitivity

Stability: volatile; rank range: 211.

Keyword Scores

world model

interactive world model

world dynamics prediction

model-based reinforcement learning world model

world simulator

generative world model

video world model

Deep Analysis

Innovations

Dynamic Causal Program Graph (CPG) as a world model representing procedural and causal knowledge as a directed hypergraph
Inductive Logic Programming (ILP) engine that synthesizes new Horn clauses from experiential feedback
Hybrid reasoning core where an LLM proposes candidate sub-goals verified for logical consistency by a Neural Theorem Prover (NTP)

Methodology

AGEL-Comp integrates a dynamic Causal Program Graph (CPG) as a world model, an Inductive Logic Programming (ILP) engine for synthesizing Horn clauses from feedback, and a hybrid reasoning core combining an LLM for sub-goal proposal with a Neural Theorem Prover (NTP) for logical verification. These components operate in a deduction-abduction learning cycle, enabling the agent to deduce plans and abductively expand its symbolic world model while a neural adaptation phase aligns reasoning with new knowledge. The framework is evaluated within the Retro Quest simulation environment using a protocol designed to probe compositional generalization.

Key Results

The AGEL-Comp model demonstrates better performance than pure LLM-based models in compositional generalization scenarios within the Retro Quest simulation environment.

Resources

Paper