Pith. sign in

REVIEW 21 cited by

Harnessing Explanations: LLM-to-LM Interpreter for Enhanced Text-Attributed Graph Representation Learning

Not yet reviewed by Pith; the record is open.

Re-run · record.json Download PDF Read on arXiv ↗

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2305.19523 v5 pith:VEKQZK3D submitted 2023-05-31 cs.LG

Xiaoxin He , Xavier Bresson , Thomas Laurent , Adam Perold , Yann LeCun , Bryan Hooi This is my paper
classification cs.LG
keywords featuresexplanationsgraphlearningllmsmethodtextualattributes
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

Representation learning on text-attributed graphs (TAGs) has become a critical research problem in recent years. A typical example of a TAG is a paper citation graph, where the text of each paper serves as node attributes. Initial graph neural network (GNN) pipelines handled these text attributes by transforming them into shallow or hand-crafted features, such as skip-gram or bag-of-words features. Recent efforts have focused on enhancing these pipelines with language models (LMs), which typically demand intricate designs and substantial computational resources. With the advent of powerful large language models (LLMs) such as GPT or Llama2, which demonstrate an ability to reason and to utilize general knowledge, there is a growing need for techniques which combine the textual modelling abilities of LLMs with the structural learning capabilities of GNNs. Hence, in this work, we focus on leveraging LLMs to capture textual information as features, which can be used to boost GNN performance on downstream tasks. A key innovation is our use of explanations as features: we prompt an LLM to perform zero-shot classification, request textual explanations for its decision-making process, and design an LLM-to-LM interpreter to translate these explanations into informative features for downstream GNNs. Our experiments demonstrate that our method achieves state-of-the-art results on well-established TAG datasets, including Cora, PubMed, ogbn-arxiv, as well as our newly introduced dataset, tape-arxiv23. Furthermore, our method significantly speeds up training, achieving a 2.88 times improvement over the closest baseline on ogbn-arxiv. Lastly, we believe the versatility of the proposed method extends beyond TAGs and holds the potential to enhance other tasks involving graph-text data. Our codes and datasets are available at: https://github.com/XiaoxinHe/TAPE.

Discussion (0). Sign in to comment.

Forward citations

Cited by 21 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Agentic Graph Token Reasoning

    cs.LG 2026-08 conditional novelty 7.0 of 10

    Agentic Graph Token (AGT) reasoning lets an LLM iteratively select graph views, encode them as tokens, and reason step-by-step, beating static graph-token and text-agent baselines on seven graph domains.

  2. LLM Features Can Hurt GNNs: Concatenation Interference on Homophilous Graph Benchmarks

    cs.LG 2026-06 accept novelty 7.0 of 10

    Pure concatenation of LLM features degrades GNN accuracy on homophilous datasets, with Delta_sig metric predicting when the drop occurs better than homophily.

  3. BLEG: LLM Functions as Powerful fMRI Graph-Enhancer for Brain Network Analysis

    cs.LG 2026-04 unverdicted novelty 7.0 of 10

    BLEG enhances GNNs for fMRI brain network analysis by prompting LLMs for text augmentation, using cost-effective instruction tuning, and applying alignment losses during joint training.

  4. Beyond One-Size-Fits-All: Adaptive Subgraph Denoising for Zero-Shot Graph Learning with Large Language Models

    cs.LG 2026-03 unverdicted novelty 7.0 of 10

    GraphSSR introduces an adaptive SSR pipeline with SSR-SFT data synthesis and SSR-RL (Authenticity-Reinforced and Denoising-Reinforced stages) to overcome one-size-fits-all subgraph noise in zero-shot LLM graph reasoning.

  5. When LLM Agents Meet Graph Optimization: An Automated Data Quality Improvement Approach

    cs.LG 2025-10 unverdicted novelty 7.0 of 10

    LAGA is a unified multi-agent LLM framework that automates comprehensive quality optimization for text-attributed graphs by running detection, planning, action, and evaluation agents in a closed loop.

  6. One Model, Many Graphs: Learning over Attributed Graphs across Heterogeneous Modalities with Vision-Language Models

    cs.LG 2026-07 conditional novelty 6.0 of 10

    OMG-VLM is a single VLM-based model that handles text-, image-, and multi-attributed graphs through structure-aware adapters, reporting gains on several node/link prediction benchmarks.

  7. Beyond the Golden Teacher: Enhancing Graph Learning through LLM-GNN Co-teaching

    cs.LG 2026-06 unverdicted novelty 6.0 of 10

    Bidirectional LLM-GNN co-teaching with round-based pseudo-label preference optimization outperforms golden-teacher baselines on few-shot TAG benchmarks by 3-8% absolute gains.

  8. Where LLM Annotators Fail: Label-Free Learning on Graphs with LLMs

    cs.LG 2026-05 unverdicted novelty 6.0 of 10

    CANE estimates cluster-specific reliability of noisy LLM pseudo-labels on graphs without ground truth to improve label-free node classification.

  9. S2Aligner: Pair-Efficient and Transferable Pre-Training for Sparse Text-Attributed Graphs

    cs.LG 2026-05 unverdicted novelty 6.0 of 10

    S2Aligner decouples semantic and structural components in LLM-as-Aligner pre-training for sparse TAGs and uses structure-oriented reconstruction plus domain risk balancing to improve transferability and reduce general...

  10. Linking Extreme Discourse to Structural Polarization in Signed Interaction Networks

    cs.SI 2026-05 unverdicted novelty 6.0 of 10

    A pipeline derives continuous signed edges from LLM stance scores on text and links discourse signals such as toxicity and extreme claims to changes in structural polarization measured by spectral and frustration scor...

  11. PRISM: Iterative Cross-Modal Posterior Refinement for Dynamic Text-Attributed Graphs

    cs.LG 2026-05 unverdicted novelty 6.0 of 10

    PRISM iteratively transforms semantic priors into behavior-conditioned posteriors via cross-modal refinement to improve representation learning on dynamic text-attributed graphs.

  12. LoReC: Rethinking Large Language Models for Graph Data Analysis

    cs.LG 2026-04 unverdicted novelty 6.0 of 10

    LoReC enhances LLMs for graph tasks via attention redistribution, graph re-injection into FFN, and logit rectification, yielding improvements over GraphLLM and GNN baselines on diverse datasets.

  13. DuConTE: Dual-Granularity Text Encoder with Topology-Constrained Attention for Text-attributed Graphs

    cs.CL 2026-04 unverdicted novelty 6.0 of 10

    DuConTE is a dual-granularity text encoder that incorporates graph topology into language model attention for improved node representations in text-attributed graphs.

  14. Graph-Based Alternatives to LLMs for Human Simulation

    cs.CL 2025-11 conditional novelty 6.0 of 10

    GEMS formulates close-ended human-behavior simulation as link prediction on a heterogeneous graph and matches or exceeds LLM performance with three orders of magnitude fewer parameters across three datasets and three ...

  15. UNIT: Unleash Large Language Models Potential for Graph Continual Learning

    cs.AI 2026-07 conditional novelty 5.5 of 10

    UNIT reaches SOTA graph continual learning by one-shot LLM instruction tuning plus uncertainty-aware semantic and structural class anchors that jointly set classifier weights.

  16. OpenRTAG: A Comprehensive Benchmark for Robust Text-Attributed Graph Learning under Data Quality Degradation

    cs.AI 2026-07 conditional novelty 5.0 of 10

    OpenRTAG is a benchmark that organizes text-attributed-graph data-quality issues into a 3x3 taxonomy (text/structure/label by sparsity/noise/imbalance) and evaluates model robustness across nine datasets and three tasks.

  17. S2Aligner: Pair-Efficient and Transferable Pre-Training for Sparse Text-Attributed Graphs

    cs.LG 2026-05 unverdicted novelty 5.0 of 10

    S2Aligner decouples semantic and structural components in LLM-based graph-text alignment, adds structure-oriented reconstruction with consistency control, and uses sparsity-aware risk balancing to reduce cross-domain ...

  18. GSTBench: A Benchmark Study on the Transferability of Graph Self-Supervised Learning

    cs.LG 2025-08 conditional novelty 5.0 of 10

    GSTBench finds that masked feature reconstruction (GraphMAE) is the only one of five graph self-supervised pretraining objectives that consistently transfers across eight datasets, while contrastive methods often perf...

  19. CS-Agent: LLM-based Community Search via Dual-agent Collaboration

    cs.SI 2025-08 conditional novelty 5.0 of 10

    CS-Agent, a Solver-Validator two-agent dialogue with a Decider selector, improves LLM community search on synthetic graphs, and GraphCS is a new benchmark for measuring it.

  20. G2LoRA: Gradient Orthogonal Low-Rank Adaptation Framework for Graph Continual Learning on Text-Attributed Graphs

    cs.LG 2026-06 unverdicted novelty 4.0 of 10

    G2LoRA proposes category-aware gradient projection and magnitude modulation within a unified graph-text alignment objective to mitigate interference and promote transfer in continual learning on text-attributed graphs.

  21. LightRAG: Simple and Fast Retrieval-Augmented Generation

    cs.IR 2024-10 unverdicted novelty 4.0 of 10

    LightRAG builds graph structures into RAG indexing and retrieval with dual-level search and incremental updates to improve accuracy and speed.

Pith tools