Pith. sign in

REVIEW 3 major objections 5 minor 1 cited by

GraphAgent: Agentic Graph Language Assistant

T0 review · 3 major / 5 minor · reviewed 2026-08-11 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read GraphAgent claims that an 8B open-source model, using three collaborating agents that build a semantic knowledge graph from text, can exceed prior graph language models by 28% on average and larger LLMs by 31.9% on graph-aware prediction…

desk verdict Solid integration paper for graph-LLM agents, but the zero-shot evaluation protocol is unspecified in exactly the place the headline claim depends on. read the letter →

arxiv 2412.17029 v1 pith:XWC3PBKS submitted 2024-12-22 cs.AI

Yuhao Yang , Jiabin Tang , Lianghao Xia , Xingchen Zou , Yuxuan Liang , Chao Huang This is my paper · ORCID
classification cs.AI
keywords graphlanguagemodelssemanticknowledgegraphsagenticpipelinezero-shotnodeclassificationtokenizationLLM-as-judgetextgenerationheterogeneous
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

GraphAgent is an automated pipeline that turns text into a semantic knowledge graph and uses that graph, plus any user-supplied graph, to answer questions and generate text. The paper argues that this lets a small open 8B model achieve stronger zero-shot node classification than the previous best graph language model, by over 28% on average, and beat much larger language models on prediction tasks that require implicit semantic relationships. In generation, it reports lower perplexity than GPT-4o-mini, Gemini-1.5-Flash, and Llama3-70B, with GPT-4o as judge preferring its related-work text in most comparisons. A sympathetic reader would care because the claim is that automatic relational structure extraction can compensate for model size, making capable graph analysis available in an open model.

What carries the argument

The load-bearing component is the Semantic Knowledge Graph (SKG), produced by an iterative two-phase extraction where an LLM first identifies high-level scaffold entities, then augments each with a detailed description, and repeats so that fine-grained nodes hang off coarse ones; edges connect each derived node to the parent whose description generated it. The SKG is encoded with a pre-trained text encoder and a heterogeneous GNN into graph tokens, projected into the language model's embedding space through a learnable linear layer, and interleaved with language tokens. Training rests on graph-instruction alignment (intra-type and inter-type matching between graph tokens and text) followed by curriculum-ordered multi-task fine-tuning over predictive and generative agent tasks.

What would settle it

Audit the training and evaluation splits: enumerate the exact ICLR-Peer Reviews samples (Table 1 lists 4,155 training samples) and Arxiv-Papers samples used for GraphAgent-Zero-Shot, and check them against the 500-sample evaluation sets; any overlap would invalidate the zero-shot claim. Separately, human rating of 200 generated related-work passages would check whether the lower perplexity and GPT-4o preferences match human judgment.

Watch

Extended reading notes

Core claim

The central discovery is that a semantic knowledge graph constructed automatically from plain text can serve as a compact, tokenizable representation that lets a graph-language model reason about both explicit links and implicit semantic ties. On node classification with explicit graphs, the paper reports an average improvement of more than 28% over the previous state-of-the-art graph language model across settings, with specific gains of 12.8% to 48.5% depending on shot count and metric. On implicit graph prediction, the 8B GraphAgent outperforms Llama3-70B and Qwen2-72B by a 31.9% average across all metrics on Arxiv-Papers and ICLR-Peer Reviews. For generation, GraphAgent achieves lower perplexity than several much larger or closed models and GPT-4o judges its related-work text better than Llama3-8B in 67% of cases and better than Llama3-70B in 58%.

Load-bearing premise

The headline zero-shot numbers stand on the premise that GraphAgent-Zero-Shot was trained only on data that is fully disjoint from the Arxiv-Papers and ICLR-Peer Reviews evaluation samples, so those results measure true transfer rather than memorized overlap.

Editorial extensions

If this is right

  • If the results hold, graph-aware prediction no longer requires a task-specific GNN or a very large proprietary model: an 8B open model with an automatically built SKG can carry zero-shot node classification and implicit semantic prediction.
  • Graph-enhanced generation benefits more from structured knowledge injection than from retrieval prompting: GraphAgent beats vanilla SFT and GraphRAG baselines of the same size on perplexity and judged quality.
  • A user who can upload a graph file or paste text, then ask in natural language, can get predictions with reasoning and generated text, because task planning and graph construction are automated inside the pipeline.
  • The ablation results imply that semantic knowledge graphs contribute most to predictive gains, while graph-instruction alignment is the key to generation quality, and curriculum training helps both.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The paper does not test whether the quality of the SKG is the bottleneck; a natural extension is to vary the number of extraction iterations or entity granularity and measure how prediction and generation performance respond, which would tell where the gains saturate.
  • Because the zero-shot claim depends on training/evaluation split disjointness that the paper does not explicitly document, an audit of the sample overlap is the first check a reader should perform before relying on the Table 3 numbers.
  • The task planning and graph generation agents currently rely on GPT-3.5-Turbo, so the full pipeline is not entirely open; replacing those two agents with an open model would test how much of the benefit is due to the architecture versus the proprietary planner.
  • If the pipeline transfers to other text-heavy domains, a likely consequence is that scientific literature, legal documents, and clinical notes could get graph-aware assistants built on small models plus automated knowledge graphs, without hand-curated ontologies.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. The paper proposes GraphAgent, an agentic pipeline that combines a Graph Generator Agent (which constructs semantic knowledge graphs from text), a Task Planning Agent (which parses user queries and grounds explicit or implicit graphs into graph tokens), and a Graph Action Agent (a Llama3-8B-based graph language model trained with graph-instruction alignment and curriculum multi-task finetuning). The system is evaluated on predictive tasks (node classification on IMDB, ACM, Arxiv-Papers, ICLR-Peer Reviews) and generative tasks (related-work generation and GovReport summarization). The headline claims are an average improvement of over 28% over HiGPT on IMDB-to-ACM zero-shot transfer, consistent outperformance of larger open-source LLMs on implicit-semantic prediction, and lower perplexity than several commercial and open-source LLMs on graph-enhanced generation, with additional LLM-as-judge preference results.

Significance. If the empirical results hold, the paper makes a useful contribution by showing that a relatively small open LLM, when embedded in a multi-agent pipeline with automatic semantic knowledge graph construction and graph tokenization, can match or exceed much larger models on graph-related predictive and generative tasks. The ablations (Figure 7) provide evidence that each of the three main components contributes, and the release of code at https://github.com/HKUDS/GraphAgent supports reproducibility. The main scientific value lies in the system-level integration and the corresponding evidence that the design choices matter; however, the zero-shot generalization claim, which is central to the paper's framing, rests on an under-specified evaluation protocol that must be clarified before the results can be accepted at face value.

major comments (3)
  1. [Section 3.3, Table 3, Section 3.1.2] The zero-shot configuration of GraphAgent-Zero-Shot is not specified. Unlike Section 3.2, which explicitly states that the model is trained on IMDB and evaluated on ACM, Section 3.3 never states what data GraphAgent-Zero-Shot was trained on. Table 1 lists 5,175 training samples for Arxiv-Papers and 3,141 for ICLR-Peer Reviews, and Section 3.1.2 explicitly says the ICLR-Peer Reviews dataset 'is used for both training and testing purposes.' If GraphAgent-Zero-Shot was fine-tuned on either of those training splits, or if the 500 evaluation samples overlap with the training data, then the Table 3 numbers for this variant are in-distribution supervised results, not zero-shot transfer, and the comparison against non-fine-tuned LLMs would be misleading. The authors must clearly report the training data for each variant (Task Expert, General, Zero-Shot) and, if the zero-shot model was trained on other data, re-run the evaluation on a truly disjoint split to substantiate the zero-shot claim.
  2. [Section 3.4, Tables 4 and 5, Figure 5] The generation-quality comparison relies primarily on perplexity computed with Llama3-70b and Qwen2-72b. The paper acknowledges same-family bias only by shading those cells lightly grey, but the central claims of lower perplexity than GPT-4o-mini, Gemini-1.5-Flash, and other evaluators of different families are still based on a metric that is known to favor models that are close to the evaluating LM. The authors should either add a third evaluator from a different model family, supplement with human judgments (even a small sample), or report significance/confidence intervals for the PPL differences; as written, a large part of RQ3's support rests on a single, potentially biased metric.
  3. [Section 3.1.2, Table 1] The dataset statistics for the related-work generation task are inconsistent between the table and the text. Table 1 reports 4,155 training samples for 'Related Work Generation' but no evaluation count, while the text states 'approximately 5,000 topic-content pairs' and Section 3.4 mentions '200 samples from the text generation test set' for the judge-based evaluation. Please reconcile these numbers and provide an exact train/test split for the generative datasets so that the experimental results are reproducible and the evaluation scale is clear.
minor comments (5)
  1. [Section 2.4.2, Eq. (17)] Typo: 'anthor' should be 'another'.
  2. [References] The reference list contains duplicate entries, e.g., Wang et al. 2019a and 2019b, and Velickovic et al. 2018 appears twice in different formats; these should be consolidated.
  3. [Appendix A.1.2 and Section 3.6] Table 8 is referenced as both the curriculum training table and a qualitative comparison table; the numbering is confusing. Renumber the appendix tables to avoid ambiguity.
  4. [Section 3.3, Table 3] The claim of '31.9% improvement across all metrics on both datasets' is not supported by the full table: GPT-4o-mini achieves a higher Mi-F1 on ICLR-Peer Reviews (0.692 vs 0.686 for GraphAgent-Task Expert), and several commercial models have comparable or higher AUC on one dataset. The claim should either be scoped to open-source LLMs or supported with a table-level significance test.
  5. [Section 3.1.1, Implementation Details] The system uses GPT-3.5-Turbo for the task planning and graph generation agents, so the open-source release cannot reproduce the full pipeline without relying on a proprietary API. The GitHub repository should clearly document which components can run entirely with open-weight models and which require external API access.

Circularity Check

0 steps flagged · score 2.0 of 10

No demonstrated circularity; the central predictive and generative results are empirical against external baselines. Score reflects minor reuse of the authors' own HiGPT/GraphGPT components and a missing zero-shot training-protocol statement.

full rationale

GraphAgent's main claims are supported by experiments against external baselines, not by equations that reduce a predicted quantity to a fitted input. Table 2 uses an explicit train-on-IMDB / evaluate-on-ACM zero-shot protocol; Table 3 compares against open and closed LLMs, fine-tuned LLMs, and GraphRAG; Table 4 uses external PPL evaluators (Llama3-70b, Qwen2-72b) and GPT-4o-as-judge. The architecture borrows the graph-token and alignment recipe from the authors' own HiGPT and GraphGPT works (Tang et al. 2024a,b), but this is component reuse rather than a load-bearing self-citation: HiGPT itself appears as a baseline in Table 2, and the headline improvements are measured against that baseline and against non-self models. No derivation step sets up a definition of X in terms of Y and then claims to predict Y from X; no fitted parameter is renamed as a prediction; no uniqueness theorem is imported from prior work. One verification gap is flagged but is not a demonstrated circularity: Section 3.1.2 says of ICLR-Peer Reviews, 'This dataset is used for both training and testing purposes,' while Section 3.3 presents GraphAgent-Zero-Shot results without stating that variant's training data, in contrast to the explicit IMDB-to-ACM protocol in Section 3.2. If GraphAgent-Zero-Shot were trained on the Arxiv-Papers or ICLR training splits, the 'zero-shot' label would be inapplicable and the comparison to non-fine-tuned LLMs would be misleading; however, the paper does not exhibit that training configuration, so this is an unverified evaluation-protocol assumption rather than a demonstrated equivalence. Score 2 reflects minor self-citation and the reporting gap, not a circular derivation.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The central claim takes as given that automatically extracted knowledge graphs are task-relevant, that same-family perplexity and GPT-4o-as-judge are valid quality measures, and that the zero-shot evaluation is genuinely zero-shot. No new physical or theoretical entities are introduced; the semantic knowledge graph is a data structure built from input text. The hand-set hyperparameters (hop count, curriculum ratios, few-shot prompt content) are untested degrees of freedom.

free parameters (3)
  • Iterative semantic knowledge graph hop count k = 2
    Section 3.1.4 sets the iterative steps to 2 without ablation across k. The knowledge graph depth directly affects the graph tokens fed to the model and hence all downstream results.
  • Curriculum mixing ratios = Epoch 1: 10% alignment, 70% predictive, 20% generative; evolves to 0%/40%/60%
    Appendix Table 8 defines hand-chosen data proportions per epoch; no sensitivity analysis is reported, though the ablation shows curriculum matters.
  • Few-shot example count in system prompts = Not specified
    All three agents use GPT-3.5-Turbo with few-shot examples in prompts; the number and content of these examples is not provided, so the entire pipeline depends on unstated prompt content.
assumptions (4)
  • domain assumption The semantic knowledge graph generated by GPT-3.5 from text captures task-relevant semantic dependencies.
    Used throughout Section 2.2; if the knowledge graph is noisy or off-task, the performance gains attributed to it in ablations would not transfer.
  • domain assumption Perplexity computed by Llama3-70b and Qwen2-72b is a valid measure of text generation quality.
    Section 3.1.4 declares perplexity the primary generative metric. Same-family bias is acknowledged in Figures 4 and 5 but not corrected, and perplexity is known to reward repetitive or short text.
  • domain assumption GPT-4o as judge approximates human judgment of generated text.
    Section 3.4 and Appendix Table 4 use GPT-4o as a judge; no human agreement study is reported.
  • domain assumption The zero-shot evaluation protocol is genuinely zero-shot.
    Table 1 lists ICLR-Peer Reviews with 4,155 training samples; the paper never states which data GraphAgent-Zero-Shot was trained on, so the transfer claim rests on an unverified premise.

how reviews work

0 comments
Cite this review

Pith. "Pith review of GraphAgent: Agentic Graph Language Assistant." pith.science (2026). https://pith.science/paper/XWC3PBKS

@misc{pith2026241217029,
  author       = {Pith},
  title        = {Pith review of: GraphAgent: Agentic Graph Language Assistant},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/XWC3PBKS}},
  note         = {Machine review of arXiv:2412.17029}
}
read the original abstract

Real-world data is represented in both structured (e.g., graph connections) and unstructured (e.g., textual, visual information) formats, encompassing complex relationships that include explicit links (such as social connections and user behaviors) and implicit interdependencies among semantic entities, often illustrated through knowledge graphs. In this work, we propose GraphAgent, an automated agent pipeline that addresses both explicit graph dependencies and implicit graph-enhanced semantic inter-dependencies, aligning with practical data scenarios for predictive tasks (e.g., node classification) and generative tasks (e.g., text generation). GraphAgent comprises three key components: (i) a Graph Generator Agent that builds knowledge graphs to reflect complex semantic dependencies; (ii) a Task Planning Agent that interprets diverse user queries and formulates corresponding tasks through agentic self-planning; and (iii) a Task Execution Agent that efficiently executes planned tasks while automating tool matching and invocation in response to user queries. These agents collaborate seamlessly, integrating language models with graph language models to uncover intricate relational information and data semantic dependencies. Through extensive experiments on various graph-related predictive and text generative tasks on diverse datasets, we demonstrate the effectiveness of our GraphAgent across various settings. We have made our proposed GraphAgent open-source at: https://github.com/HKUDS/GraphAgent.

Figures

Figures reproduced from arXiv: 2412.17029 by the authors.

Figure 1
Figure 1. GraphAgent processes both structured and unstructured data, adapting seamlessly to vari￾ous downstream tasks across diverse scenarios. Inspired by the recent success of large language models (LLMs), researchers are striving to en￾hance the generalization capabilities of graph learning models by enabling LLMs to compre￾hend graph structural information. Notable ex￾amples include GraphGPT (Tang et al., 2024a) and LLaG… view at source ↗
Figure 2
Figure 2. The overall framework of the proposed GraphAgent. [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Performance comparison with state-of￾the-art LLMs on complex graph prediction tasks involving implicit semantic relationships. Re￾sults marked with * indicate statistical significance (p<0.01) compared to the second-best performer. Method Model Size Arxiv-Papers ICLR-Peer Reviews Mi-F1 Ma-F1 AUC Mi-F1 Ma-F1 AUC Open-sourced LLMs Llama3-8b 8B 0.514 0.289 0.527 0.402 0.394 0.502 Mistral-Nemo 12B 0.510 0.292 0.615 0.27… view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: Performances on ACL-EMNLP related works content generation. Light grey denotes that the score is computed with the same-family model. Method Model Size PPL-Llama3-70b PPL-Qwen2-72b Mean Max Mean Max Open-sourced LLMs Llama3-8b 8B 7.016 13.061 7.491 12.787 Mistral-Nemo …
Figure 5
Figure 5. Figure 5: GovReport summarization performance. Evaluation scores are presented with same-family model comparisons highlighted in light grey. Method Model Size PPL-Llama3-70b PPL-Qwen2-72b Mean Max Mean Max Llama3-8b 8B 9.476 25.355 7.564 17.443 Mistral-Nemo 12B 9.333 28.537 7.19…
Figure 6
Figure 6. Figure 6: Comparative evaluation results: GPT￾4o as judge assessing our proposed GraphAgent framework against state-of-the-art open-source LLMs. We evaluated GraphAgent against Llama3-8b and Llama3-70b on two distinct graph-enhanced text generation tasks, with results presented …
Figure 7
Figure 7. Figure 7: Ablation study comparing GraphAgent with its variants on both graph-related prediction and graph-enhanced text generation tasks. To evaluate each component in GraphAgent, we conducted an ablation study with the following variants: • (-) SKG: Removes the graph gener￾ati…
Figure 8
Figure 8. Figure 8: Curriculum Learning Strategy: Training Epochs and Data Mixing Ratios. Alignment Data Predictive Data Generative Data Epoch 1 10% 70% 20% Epoch 2 5% 60% 35% Epoch 3 0% 50% 50% Afterwards 0% 40% 60% We employ a curriculum learning strategy to effectively train our graph …

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

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

  1. Graphs Meet AI Agents: Taxonomy, Progress, and Future Opportunities

    cs.AI 2025-06 conditional novelty 5.0 of 10

    A survey that groups graph-empowered AI agent research into planning, execution, memory, and multi-agent coordination, plus agents-for-graphs and applications.

Reference graph

Works this paper leans on

57 extracted references · 41 canonical work pages · cited by 1 Pith paper

  1. [1]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION format.date year duplicate empty "emp...

  2. [2]

    Curriculum learning

    Yoshua Bengio, J \'e r \^o me Louradour, Ronan Collobert, and Jason Weston. Curriculum learning. In ICML, pp.\ 41--48, 2009

    work page 2009

  3. [3]

    Edgi: Equivariant diffusion for planning with embodied agents

    Johann Brehmer, Joey Bose, Pim De Haan, and Taco S Cohen. Edgi: Equivariant diffusion for planning with embodied agents. NeurIPS, 36, 2024

    work page 2024

  4. [4]

    Web-scale academic name disambiguation: the whoiswho benchmark, leaderboard, and toolkit

    Bo Chen, Jing Zhang, Fanjin Zhang, Tianyi Han, Yuqing Cheng, Xiaoyan Li, Yuxiao Dong, and Jie Tang. Web-scale academic name disambiguation: the whoiswho benchmark, leaderboard, and toolkit. In KDD, pp.\ 3817--3828, 2023

    work page 2023

  5. [5]

    Simple and deep graph convolutional networks

    Ming Chen, Zhewei Wei, Zengfeng Huang, Bolin Ding, and Yaliang Li. Simple and deep graph convolutional networks. In ICML, pp.\ 1725--1735. PMLR, 2020

    work page 2020

  6. [6]

    Llaga: Large language and graph assistant

    Runjin Chen, Tong Zhao, Ajay Jaiswal, Neil Shah, and Zhangyang Wang. Llaga: Large language and graph assistant. ICML, 2024 a

    work page 2024

  7. [7]

    Llaga: Large language and graph assistant

    Runjin Chen, Tong Zhao, AJAY KUMAR JAISWAL, Neil Shah, and Zhangyang Wang. Llaga: Large language and graph assistant. In ICML, 2024 b

    work page 2024

  8. [8]

    Towards robust graph neural networks for noisy graphs with sparse labels

    Enyan Dai, Wei Jin, Hui Liu, and Suhang Wang. Towards robust graph neural networks for noisy graphs with sparse labels. In WSDM, pp.\ 181--191, 2022

    work page 2022

Show all 57 references
  1. [9]

    Benchmarking graph neural networks

    Vijay Prakash Dwivedi, Chaitanya K Joshi, Anh Tuan Luu, Thomas Laurent, Yoshua Bengio, and Xavier Bresson. Benchmarking graph neural networks. JMLR, 24 0 (43): 0 1--48, 2023

    2023

  2. [10]

    Fast graph representation learning with pytorch geometric

    Matthias Fey and Jan Eric Lenssen. Fast graph representation learning with pytorch geometric. arXiv preprint arXiv:1903.02428, 2019

    1903 arXiv

  3. [11]

    Relational deep learning: Graph representation learning on relational databases

    Matthias Fey, Weihua Hu, Kexin Huang, Jan Eric Lenssen, Rishabh Ranjan, Joshua Robinson, Rex Ying, Jiaxuan You, and Jure Leskovec. Relational deep learning: Graph representation learning on relational databases. arXiv preprint arXiv:2312.04615, 2023

    2023 arXiv

  4. [12]

    Magnn: Metapath aggregated graph neural network for heterogeneous graph embedding

    Xinyu Fu, Jiani Zhang, Ziqiao Meng, and Irwin King. Magnn: Metapath aggregated graph neural network for heterogeneous graph embedding. In Proceedings of the web conference 2020, pp.\ 2331--2341, 2020

    2020

  5. [13]

    Graph representation learning

    William L Hamilton. Graph representation learning. Morgan & Claypool Publishers, 2020

    2020

  6. [14]

    Hamilton, Zhitao Ying, and Jure Leskovec

    William L. Hamilton, Zhitao Ying, and Jure Leskovec. Inductive representation learning on large graphs. In NeurIPS , pp.\ 1024--1034, 2017

    2017

  7. [15]

    Gat-mf: Graph attention mean field for very large scale multi-agent reinforcement learning

    Qianyue Hao, Wenzhen Huang, Tao Feng, Jian Yuan, and Yong Li. Gat-mf: Graph attention mean field for very large scale multi-agent reinforcement learning. In KDD, pp.\ 685--697, 2023

    2023

  8. [16]

    Harnessing explanations: Llm-to-lm interpreter for enhanced text-attributed graph representation learning

    Xiaoxin He, Xavier Bresson, Thomas Laurent, Adam Perold, Yann LeCun, and Bryan Hooi. Harnessing explanations: Llm-to-lm interpreter for enhanced text-attributed graph representation learning. arXiv preprint arXiv:2305.19523, 2023

    2023 arXiv

  9. [17]

    Cogagent: A visual language model for gui agents

    Wenyi Hong, Weihan Wang, Qingsong Lv, Jiazheng Xu, Wenmeng Yu, Junhui Ji, Yan Wang, Zihan Wang, Yuxiao Dong, Ming Ding, et al. Cogagent: A visual language model for gui agents. In CVPR, pp.\ 14281--14290, 2024

    2024

  10. [18]

    Heterogeneous graph transformer

    Ziniu Hu, Yuxiao Dong, Kuansan Wang, and Yizhou Sun. Heterogeneous graph transformer. In WWW , pp.\ 2704--2710. ACM / IW3C2 , 2020

    2020

  11. [19]

    Uncertainty quantification over graph with conformalized graph neural networks

    Kexin Huang, Ying Jin, Emmanuel Candes, and Jure Leskovec. Uncertainty quantification over graph with conformalized graph neural networks. NeurIPS, 36, 2024 a

    2024

  12. [20]

    Grounded decoding: Guiding text generation with grounded models for embodied agents

    Wenlong Huang, Fei Xia, Dhruv Shah, Danny Driess, Andy Zeng, Yao Lu, Pete Florence, Igor Mordatch, Sergey Levine, Karol Hausman, et al. Grounded decoding: Guiding text generation with grounded models for embodied agents. NeurIPS, 36, 2024 b

    2024

  13. [21]

    Swe-bench: Can language models resolve real-world github issues? In ICLR, 2024

    Carlos E Jimenez, John Yang, Alexander Wettig, Shunyu Yao, Kexin Pei, Ofir Press, and Karthik R Narasimhan. Swe-bench: Can language models resolve real-world github issues? In ICLR, 2024

    2024

  14. [22]

    Graph structure learning for robust graph neural networks

    Wei Jin, Yao Ma, Xiaorui Liu, Xianfeng Tang, Suhang Wang, and Jiliang Tang. Graph structure learning for robust graph neural networks. In KDD, pp.\ 66--74, 2020

    2020

  15. [23]

    Semi-supervised classification with graph convolutional networks

    Thomas N Kipf and Max Welling. Semi-supervised classification with graph convolutional networks. In ICLR, 2017

    2017

  16. [24]

    Visualwebarena: Evaluating multimodal agents on realistic visual web tasks

    Jing Yu Koh, Robert Lo, Lawrence Jang, Vikram Duvvur, Ming Chong Lim, Po-Yu Huang, Graham Neubig, Shuyan Zhou, Ruslan Salakhutdinov, and Daniel Fried. Visualwebarena: Evaluating multimodal agents on realistic visual web tasks. ACL, 2024

    2024

  17. [25]

    Text is all you need: Learning language representations for sequential recommendation

    Jiacheng Li, Ming Wang, Jin Li, Jinmiao Fu, Xin Shen, Jingbo Shang, and Julian McAuley. Text is all you need: Learning language representations for sequential recommendation. In KDD, pp.\ 1258--1267, 2023

    2023

  18. [26]

    Zerog: Investigating cross-dataset zero-shot transferability in graphs

    Yuhan Li, Peisong Wang, Zhixun Li, Jeffrey Xu Yu, and Jia Li. Zerog: Investigating cross-dataset zero-shot transferability in graphs. In KDD, pp.\ 1725--1735, 2024

    2024

  19. [27]

    Rouge: A package for automatic evaluation of summaries

    Chin-Yew Lin. Rouge: A package for automatic evaluation of summaries. In Text summarization branches out, pp.\ 74--81, 2004

    2004

  20. [28]

    Visual instruction tuning

    Haotian Liu, Chunyuan Li, Qingyang Wu, and Yong Jae Lee. Visual instruction tuning. NeurIPS, 36, 2024

    2024

  21. [29]

    Revisiting graph contrastive learning from the perspective of graph spectrum

    Nian Liu, Xiao Wang, Deyu Bo, Chuan Shi, and Jian Pei. Revisiting graph contrastive learning from the perspective of graph spectrum. NeurIPS, 35: 0 2972--2983, 2022

    2022

  22. [31]

    Llmscore: Unveiling the power of large language models in text-to-image synthesis evaluation

    Yujie Lu, Xianjun Yang, Xiujun Li, Xin Eric Wang, and William Yang Wang. Llmscore: Unveiling the power of large language models in text-to-image synthesis evaluation. NeurIPS, 36, 2024

    2024

  23. [32]

    Graph foundation models

    Haitao Mao, Zhikai Chen, Wenzhuo Tang, Jianan Zhao, Yao Ma, Tong Zhao, Neil Shah, Michael Galkin, and Jiliang Tang. Graph foundation models. ICML, 2024

    2024

  24. [33]

    Bleu: a method for automatic evaluation of machine translation

    Kishore Papineni, Salim Roukos, Todd Ward, and Wei-Jing Zhu. Bleu: a method for automatic evaluation of machine translation. In ACL, pp.\ 311--318, 2002

    2002

  25. [34]

    Reflexion: Language agents with verbal reinforcement learning

    Noah Shinn, Federico Cassano, Ashwin Gopinath, Karthik Narasimhan, and Shunyu Yao. Reflexion: Language agents with verbal reinforcement learning. NeurIPS, 36, 2023

    2023

  26. [35]

    A review-aware graph contrastive learning framework for recommendation

    Jie Shuai, Kun Zhang, Le Wu, Peijie Sun, Richang Hong, Meng Wang, and Yong Li. A review-aware graph contrastive learning framework for recommendation. In SIGIR, pp.\ 1283--1293, 2022

    2022

  27. [36]

    Graphgpt: Graph instruction tuning for large language models

    Jiabin Tang, Yuhao Yang, Wei Wei, Lei Shi, Lixin Su, Suqi Cheng, Dawei Yin, and Chao Huang. Graphgpt: Graph instruction tuning for large language models. In SIGIR, pp.\ 491--500, 2024 a

    2024

  28. [37]

    Higpt: Heterogeneous graph language model

    Jiabin Tang, Yuhao Yang, Wei Wei, Lei Shi, Long Xia, Dawei Yin, and Chao Huang. Higpt: Heterogeneous graph language model. In KDD, 2024 b

    2024

  29. [38]

    Graph attention networks

    Petar Veli c kovi \'c , Guillem Cucurull, Arantxa Casanova, Adriana Romero, Pietro Lio, and Yoshua Bengio. Graph attention networks. In ICLR, 2018

    2018

  30. [39]

    Graph attention networks

    Petar Velickovic, Guillem Cucurull, Arantxa Casanova, Adriana Romero, et al. Graph attention networks. In ICLR (Poster) . OpenReview.net, 2018

    2018

  31. [40]

    Heterogeneous graph attention network

    Xiao Wang, Houye Ji, Chuan Shi, Bai Wang, Yanfang Ye, Peng Cui, and Philip S Yu. Heterogeneous graph attention network. In WWW, pp.\ 2022--2032, 2019 a

    2022

  32. [41]

    Heterogeneous graph attention network

    Xiao Wang, Houye Ji, Chuan Shi, Bai Wang, Yanfang Ye, et al. Heterogeneous graph attention network. In WWW , pp.\ 2022--2032. ACM , 2019 b

    2022

  33. [42]

    A survey on heterogeneous graph embedding: methods, techniques, applications and sources

    Xiao Wang, Deyu Bo, Chuan Shi, Shaohua Fan, Yanfang Ye, and S Yu Philip. A survey on heterogeneous graph embedding: methods, techniques, applications and sources. TBD, 9 0 (2): 0 415--436, 2022

    2022

  34. [43]

    Chain-of-thought prompting elicits reasoning in large language models

    Jason Wei, Xuezhi Wang, Dale Schuurmans, Maarten Bosma, Fei Xia, Ed Chi, Quoc V Le, Denny Zhou, et al. Chain-of-thought prompting elicits reasoning in large language models. NeurIPS, 35: 0 24824--24837, 2022

    2022

  35. [44]

    Graph convolutional kernel machine versus graph convolutional networks

    Zhihao Wu, Zhao Zhang, and Jicong Fan. Graph convolutional kernel machine versus graph convolutional networks. NeurIPS, 36, 2024

    2024

  36. [45]

    A comprehensive survey on graph neural networks

    Zonghan Wu, Shirui Pan, Fengwen Chen, Guodong Long, Chengqi Zhang, and S Yu Philip. A comprehensive survey on graph neural networks. TPAMI, 32 0 (1): 0 4--24, 2020

    2020

  37. [46]

    Anygraph: Graph foundation model in the wild

    Lianghao Xia and Chao Huang. Anygraph: Graph foundation model in the wild. arXiv preprint arXiv:2408.10700, 2024

    2024 arXiv

  38. [47]

    Openagents: An open platform for language agents in the wild

    Tianbao Xie, Fan Zhou, Zhoujun Cheng, Peng Shi, Luoxuan Weng, Yitao Liu, Toh Jing Hua, Junning Zhao, Qian Liu, Che Liu, et al. Openagents: An open platform for language agents in the wild. In COLM, 2023

    2023

  39. [48]

    Curriculum learning for natural language understanding

    Benfeng Xu, Licheng Zhang, Zhendong Mao, Quan Wang, Hongtao Xie, and Yongdong Zhang. Curriculum learning for natural language understanding. In ACL, pp.\ 6095--6104, 2020

    2020

  40. [49]

    Qwen2 technical report

    An Yang, Baosong Yang, Binyuan Hui, Bo Zheng, Bowen Yu, Chang Zhou, Chengpeng Li, Chengyuan Li, Dayiheng Liu, Fei Huang, et al. Qwen2 technical report. arXiv preprint arXiv:2407.10671, 2024

    2024 arXiv

  41. [50]

    Understanding negative sampling in graph representation learning

    Zhen Yang, Ming Ding, Chang Zhou, Hongxia Yang, Jingren Zhou, and Jie Tang. Understanding negative sampling in graph representation learning. In KDD, pp.\ 1666--1676, 2020

    2020

  42. [51]

    React: Synergizing reasoning and acting in language models

    Shunyu Yao, Jeffrey Zhao, Dian Yu, Nan Du, Izhak Shafran, Karthik R Narasimhan, and Yuan Cao. React: Synergizing reasoning and acting in language models. In ICLR, 2023

    2023

  43. [52]

    Chuxu Zhang, Dongjin Song, Chao Huang, Ananthram Swami, and Nitesh V. Chawla. Heterogeneous graph neural network. In KDD , pp.\ 793--803. ACM , 2019

    2019

  44. [53]

    Graph attention multi-layer perceptron

    Wentao Zhang, Ziqi Yin, Zeang Sheng, Yang Li, Wen Ouyang, Xiaosen Li, Yangyu Tao, Zhi Yang, and Bin Cui. Graph attention multi-layer perceptron. In KDD, pp.\ 4560--4570, 2022

    2022

  45. [54]

    Judging llm-as-a-judge with mt-bench and chatbot arena

    Lianmin Zheng, Wei-Lin Chiang, Ying Sheng, Siyuan Zhuang, Zhanghao Wu, Yonghao Zhuang, Zi Lin, Zhuohan Li, Dacheng Li, Eric Xing, et al. Judging llm-as-a-judge with mt-bench and chatbot arena. NeurIPS, 36, 2024

    2024

  46. [55]

    Knowledge-augmented graph machine learning for drug discovery: From precision to interpretability

    Zhiqiang Zhong and Davide Mottin. Knowledge-augmented graph machine learning for drug discovery: From precision to interpretability. In KDD, pp.\ 5841--5842, 2023

    2023

  47. [56]

    @esa (Ref

    \@ifxundefined[1] #1\@undefined \@firstoftwo \@secondoftwo \@ifnum[1] #1 \@firstoftwo \@secondoftwo \@ifx[1] #1 \@firstoftwo \@secondoftwo [2] @ #1 \@temptokena #2 #1 @ \@temptokena \@ifclassloaded agu2001 natbib The agu2001 class already includes natbib coding, so you should ...

  48. [57]

    \@lbibitem[] @bibitem@first@sw\@secondoftwo \@lbibitem[#1]#2 \@extra@b@citeb \@ifundefined br@#2\@extra@b@citeb \@namedef br@#2 \@nameuse br@#2\@extra@b@citeb \@ifundefined b@#2\@extra@b@citeb @num @parse #2 @tmp #1 NAT@b@open@#2 NAT@b@shut@#2 \@ifnum @merge>\@ne @bibitem@firs...

  49. [58]

    ̮t O : ,EC[v t cf, Ked /w(]GJ>ѵ z z ƹAH\ 4 ߀k S<_|d 1=8 El5GK;Θ9 Iijk_Fuar=h* es

    @open @close @open @close and [1] URL: #1 \@ifundefined chapter * \@mkboth \@ifxundefined @sectionbib * \@mkboth * \@mkboth\@gobbletwo \@ifclassloaded amsart * \@ifclassloaded amsbook * \@ifxundefined @heading @heading NAT@ctr thebibliography [1] @ \@biblabel @NAT@ctr \@bibset...

    2024 arXiv

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.