---
title: "SALE-MLP: Structure Aware Latent Embeddings for GNN to Graph-free MLP Distillation"
authors:
  - "Harsh Pal"
  - "Sarthak Malik"
  - "Rajat Patel"
  - "Aakarsh Malhotra"
doi: "10.24963/ijcai.2025/668"
url: "https://ijcai.org/proceedings/2025/668"
pdf: "https://ijcai.org/proceedings/2025/0668.pdf"
venue: "Proceedings of the Thirty-Fourth International Joint Conference on Artificial Intelligence"
year: 2025
pages: "6003-6011"
track: "Main Track"
keywords:
  - "Machine Learning: ML: Deep learning architectures"
  - "Machine Learning: ML: Optimization"
  - "Machine Learning: ML: Representation learning"
  - "Machine Learning: ML: Sequence and graph learning"
---

# SALE-MLP: Structure Aware Latent Embeddings for GNN to Graph-free MLP Distillation

Harsh Pal, Sarthak Malik, Rajat Patel, Aakarsh Malhotra

*Proceedings of the Thirty-Fourth International Joint Conference on Artificial Intelligence*, pages 6003-6011. DOI: [10.24963/ijcai.2025/668](https://doi.org/10.24963/ijcai.2025/668)

## Abstract

Graph Neural Networks (GNNs), with their ability to effectively handle non-Euclidean data structures, have demonstrated state-of-the-art performance in learning node and graph-level representations. However, GNNs face significant computational overhead due to their message-passing mechanisms, making them impractical for real-time large-scale applications. Recently, Graph-to-MLP (G2M) knowledge distillation has emerged as a promising solution, utilizing MLPs to reduce inference latency. However, existing methods often lack structural awareness (SA), limiting their ability to capture essential graph-specific information. Moreover, some methods require access to large-scale graphs, undermining their scalability. To address these issues, we propose SALE-MLP (Structure-Aware Latent Embeddings for GNN-to-Graph-Free MLP Distillation), a novel graph-free and structure-aware approach that leverages unsupervised structural losses to align the MLP feature space with the underlying graph structure. SALE-MLP does not rely on precomputed GNN embeddings nor require graph during inference, making it efficient for real-world applications. Extensive experiments demonstrate that SALE-MLP outperforms existing G2M methods across tasks and datasets, achieving 3–4% improvement in node classification for inductive settings while maintaining strong transductive performance.