AgentGL: Towards Agentic Graph Learning with LLMs via Reinforcement Learning
Abstract
AgentGL is a reinforcement learning-driven framework that enables large language models to navigate and reason over complex relational data by integrating graph-native tools and curriculum learning strategies.
Large Language Models (LLMs) increasingly rely on agentic capabilities-iterative retrieval, tool use, and decision-making-to overcome the limits of static, parametric knowledge. Yet existing agentic frameworks treat external information as unstructured text and fail to leverage the topological dependencies inherent in real-world data. To bridge this gap, we introduce Agentic Graph Learning (AGL), a paradigm that reframes graph learning as an interleaved process of topology-aware navigation and LLM-based inference. Specifically, we propose AgentGL, the first reinforcement learning (RL)-driven framework for AGL. AgentGL equips an LLM agent with graph-native tools for multi-scale exploration, regulates tool usage via search-constrained thinking to balance accuracy and efficiency, and employs a graph-conditioned curriculum RL strategy to stabilize long-horizon policy learning without step-wise supervision. Across diverse Text-Attributed Graph (TAG) benchmarks and multiple LLM backbones, AgentGL substantially outperforms strong GraphLLMs and GraphRAG baselines, achieving absolute improvements of up to 17.5% in node classification and 28.4% in link prediction. These results demonstrate that AGL is a promising frontier for enabling LLMs to autonomously navigate and reason over complex relational environments. The code is publicly available at https://github.com/sunyuanfu/AgentGL.
Community
[🏖️ ACL 2026 Main Conference] What if your LLM could stop getting lost in graphs and start cruising through them like it just turned on Google Maps for agentic reasoning? Meet AgentGL — our new paper on Agentic Graph Learning with LLMs via Reinforcement Learning!
🔥 In this work, we introduce Agentic Graph Learning (AGL), a new paradigm that reframes graph learning as an interleaved process of topology-aware navigation and LLM-based reasoning. Building on this idea, we propose AgentGL, the first RL-driven framework for AGL.
🔍 AgentGL equips LLMs with graph-native search tools
🧠 Encourages search-constrained thinking to avoid redundant exploration
📈 Leverages graph-conditioned curriculum RL to stabilize long-horizon policy learning
✅ Across diverse text-attributed graph benchmarks and multiple LLM backbones, AgentGL consistently outperforms strong GraphLLM and GraphRAG baselines in both in-domain and zero-shot transfer settings, with gains of up to 17.5% on node classification and 28.4% on link prediction.
🌐 We hope this work can inspire a new direction for LLMs: not just reading static context during inference time, but actively exploring, navigating, and reasoning over complex relational worlds.
This is an automated message from the Librarian Bot. I found the following papers similar to this paper.
The following papers were recommended by the Semantic Scholar API
- GraphWalker: Agentic Knowledge Graph Question Answering via Synthetic Trajectory Curriculum (2026)
- GraphScout: Empowering Large Language Models with Intrinsic Exploration Ability for Agentic Graph Reasoning (2026)
- KG-Hopper: Empowering Compact Open LLMs with Knowledge Graph Reasoning via Reinforcement Learning (2026)
- SynPlanResearch-R1: Encouraging Tool Exploration for Deep Research with Synthetic Plans (2026)
- In-Context Reinforcement Learning for Tool Use in Large Language Models (2026)
- Beyond Explicit Edges: Robust Reasoning over Noisy and Sparse Knowledge Graphs (2026)
- Search-P1: Path-Centric Reward Shaping for Stable and Efficient Agentic RAG Training (2026)
Please give a thumbs up to this comment if you found it helpful!
If you want recommendations for any Paper on Hugging Face checkout this Space
You can directly ask Librarian Bot for paper recommendations by tagging it in a comment: @librarian-bot recommend
Get this paper in your agent:
hf papers read 2604.05846 Don't have the latest CLI?
curl -LsSf https://hf.co/cli/install.sh | bash Models citing this paper 0
No model linking this paper
Datasets citing this paper 0
No dataset linking this paper
Spaces citing this paper 0
No Space linking this paper
Collections including this paper 0
No Collection including this paper