Pith. sign in

REVIEW 3 major objections 5 minor 64 references

Enhancing the Patent Matching Capability of Large Language Models via the Memory Graph

T0 review · 3 major / 5 minor · reviewed 2026-08-16 · deepseek-v4-flash

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

Pith's one-line read MemGraph claims that prompting a large language model to traverse its memory graph for entities and ontologies improves patent matching by 17.68% over baseline LLMs and 10.85% over vanilla RAG on the PatentMatch dataset.

desk verdict The method works in practice, but the paper's 'memory graph' explanation is probably not the real story: the gain likely comes from a pre-view effect in the ontology-generation prompt, and that is only a fixable confound, not a fatal flaw. read the letter →

arxiv 2504.14845 v1 pith:3BDTTUT2 submitted 2025-04-21 cs.IR

Qiushi Xiong , Zhipeng Xu , Zhenghao Liu , Mengjia Wang , Zulong Chen , Yue Sun , Yu Gu , Xiaohua Li
show 1 more author
Ge Yu
This is my paper · ORCID
classification cs.IR
keywords patentmatchinglargelanguagemodelsmemorygraphretrieval-augmentedgenerationentityextractionontologyclassificationMatchdataset
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

This paper claims that patent matching can be substantially improved by prompting a large language model to mine its own parametric memory instead of retraining or relying on external knowledge. MemGraph first asks the LLM to list key technical entities in the query patent, then to assign a three-level ontology classification to the query and each candidate patent, in the spirit of the International Patent Classification. The entity list is used to expand the query for document retrieval, and the ontology labels are injected into the matching prompt. On the PatentMatch benchmark, MemGraph reports an average 17.68% improvement over vanilla LLMs and 10.85% over vanilla RAG models, with the ontology component contributing most of the gain. A sympathetic reader would care because the method offers a training-free way to give LLMs the hierarchical categorical structure that keyword matching misses.

What carries the argument

The memory graph is the paper's central device: a conceptual network of entity nodes and ontology nodes said to live in the parametric memory of the LLM, with traversal performed by two autoregressive decoding prompts (Table 1). The entity traversal generates ZIR, the list of query-specific technical entities, which is concatenated to the query before retrieval. The ontology traversal generates ZGen, a three-level classification for the query and each candidate, which is concatenated into the matching instruction. These two latent variables are the mechanism that carries the argument: they connect retrieval and generation, supplying the hierarchical categorical signal that the authors claim is missing from keyword and RAG baselines.

What would settle it

Generate the ontology labels for the query and each candidate separately, with the ontology prompt seeing only that patent's own abstract and entities, never the other options; if matching accuracy then falls back to the vanilla RAG level, the reported gain is a pre-view effect of seeing all options, not evidence for the memory-graph ontology structure.

Watch

Extended reading notes

Core claim

The central claim is that a memory graph embedded in the parametric memory of an LLM can be traversed to yield entities and ontologies that function as hints for patent matching. The traversal is implemented with two prompt templates: an entity-traversal prompt that extracts up to ten technical entities from a patent abstract, and an ontology-traversal prompt that produces three-level classifications such as 'Food processing > Mixtures > Fluid fish feed' for the query and all four options. MemGraph then expands the retrieval query with the entities and prepends the ontology labels to the matching prompt. In experiments on the PatentMatch dataset across four LLM backbones, the method raises average accuracy from 61.6% for vanilla RAG to roughly 72.4%, the best single model being GLM-4-Chat9B at 81.8% on the English split; the ontology component (ZGen) accounts for a 9.2% average improvement while query expansion with entities (ZIR) contributes 3.8%.

Load-bearing premise

The paper assumes the improvement comes from the hierarchical ontology structure produced by the memory-graph prompts, rather than from the prompts merely making the model read and compare all candidate patents in a structured way before the final choice.

Editorial extensions

If this is right

  • The ontology component (ZGen) is the main driver, contributing 9.2% over vanilla RAG versus 3.8% for entity-based query expansion (ZIR), so fine-grained categorical context matters more than better retrieval in this setup.
  • MemGraph recovers much of the performance lost when noisy retrieved documents mislead the model, narrowing the gap in the Mem-Choice scenario by about 4%.
  • The method transfers across all four tested LLM backbones and across all eight IPC sections, suggesting it does not depend on one model's particular training.
  • MemGraph reduces perplexity on the ground-truth answer and produces reasoning that an external judge model prefers 61.1% of the time over vanilla LLM and RAG reasoning.

Reading between the lines

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

  • Because the ontology prompt shows all four options before the final choice, the reported gain may be inflated by a pre-view effect; a control with independent label generation would test whether the ontology structure itself is the causal ingredient.
  • If the ontology structure is what matters, the same traversal could transfer to any domain with a hierarchical taxonomy, such as medical coding, prior-art search, or legal document matching.
  • The method could be viewed as a structured, domain-aware variant of chain-of-thought; combining it with explicit CoT reasoning might yield further gains.
  • Because the method relies on the LLM's parametric memory, its benefits may depend on how much patent-related text the backbone saw during pretraining; the variation in gains across backbones (from +8.5 to +27.8 points) is consistent with that sensitivity.
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 MemGraph, a prompt-based method for patent matching that uses an LLM's parametric memory to (1) extract technical entities from the query patent to expand the query for dense retrieval, and (2) extract three-level ontologies for the query and each candidate patent to be re-injected into the final matching prompt. The method is evaluated on the PatentMatch dataset with four LLM backbones (Llama-3.1-8B, Qwen2-7B, GLM-4-9B, Qwen2.5-14B), reporting an average accuracy improvement of 17.68% over vanilla LLMs and 10.85% over vanilla RAG models. Ablations indicate that the ontology component (ZGen), not the entity-based query expansion (ZIR), is the main driver of the gains. The paper includes additional analyses of retrieval hit/miss scenarios, model uncertainty via perplexity, reasoning quality judged by GPT-4o, and a case study.

Significance. If the causal attribution is accepted, MemGraph is a practical, training-free contribution: it is simple, requires no domain-specific fine-tuning, and reports consistent gains across four open-weight LLMs. The authors provide code and data, and the decomposition into ZIR and ZGen is a useful analytical step. However, the central claim that the 'memory graph' mechanism is responsible for the gains is not yet isolated. The ontology traversal prompt itself gives the model a comparative, structured preview of all four candidates before the final matching decision, and no ablation separates this pre-view effect from the ontology/entity content. The 'memory graph' is also never materialized as a graph, since no edges or traversal paths are defined. These issues affect the paper's headline interpretation, so I recommend major revision rather than acceptance in the current form.

major comments (3)
  1. [§3.2, Table 1, Eq. (11)-(13)] The ontology traversal prompt (Table 1) provides the LLM with the query abstract, all four option abstracts, and their extracted entities, and explicitly instructs it to 'carefully read all abstracts and technical entities' and 'maintain consistency' while generating three-level classifications for every option. The resulting ZGen (Eq. 13) is then re-injected into the matching prompt. Consequently, the matching call has already seen a structured comparative summary of all candidates, produced by the same model, before making its final decision. MemGraph (OnlyZGen) still contains this pre-view. No ablation removes it, e.g., by using an independent classifier to generate ontologies, by withholding option information during ontology generation, or by including a control where the LLM produces a technical-field summary of each option without entity extraction or memory-graph framing. Without such a control, the empirical gains cannot be attributed to the memory-graph-derived ontologies rather than to the pre-view effect.
  2. [§3.2, Eq. (14)-(16)] The 'memory graph' is defined as G=(V,E) with entity nodes and ontology nodes, but edges E are never defined or used anywhere in the paper. 'Traversal' consists of two autoregressive extraction steps (Eqs. 15-16) with no graph structure, no adjacency, no paths, and no edge semantics. Calling this a graph traversal is therefore not supported by the formalization. Either define the edges and show how the extraction process uses graph structure, or revise the terminology to 'entity and ontology extraction from parametric memory'. This is load-bearing because the paper's title, abstract, and Section 5 interpret the improvement as coming from the memory graph.
  3. [§5, Tables 3-5] The evaluation is based on only 1,000 test instances, yet no error bars or confidence intervals are reported for any accuracy number, despite the statement that statistical significance is tested by permutation test (P<0.05). Several per-cell comparisons in Table 3 and Table 4 involve small differences (e.g., Qwen2-Instruct7B Chinese: Vanilla RAG 68.2 vs MemGraph 71.4) that may not be significant, and the IPC sub-analyses in Table 4 contain as few as 26 items per category (TEXT), making per-cell differences of 10-20 points statistically fragile. Please report significance levels or confidence intervals for the main comparisons and state the effective sample sizes for the sub-analyses, including the Mem-Choice/Hit-Choice/Miss-Choice partitions in Table 5.
minor comments (5)
  1. [Abstract, §5.1] The claims '17.68% improvement over baseline LLMs' and '10.85% improvement over vanilla RAG' are averages over four specific backbones; please state this explicitly in the abstract to avoid over-generalization.
  2. [Eq. (15)] The text says 'decoding the p-related entity v_e_k(p) at the i-th step', but the index should be k. This is a typo.
  3. [Figure 3(a)] Please specify how the reported perplexity is computed: whether it is averaged over all 1,000 cases or only over correctly predicted cases, and whether the ground-truth option's probability is used even when the model predicts incorrectly. This would clarify the uncertainty claim.
  4. [§5.4, Figure 3(b)] Using GPT-4o as a judge for reasoning quality is a reasonable approach, but the paper should acknowledge that the judge may favor longer or more structured responses; reporting inter-judge agreement or a second judge would strengthen this analysis.
  5. [Table 1] The two prompt templates are long and visually dense; consider moving them to an appendix or using a compact format to improve readability.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found: MemGraph is a self-contained prompting pipeline; the ontology/entity hints are LLM-generated context, not fitted inputs or derived-from-target quantities.

full rationale

MemGraph's derivation chain is empirical rather than formal, and no step reduces to its own inputs under the required evidentiary bar. The only produced quantities are ZIR (entity list, Eqs. 8-9) and ZGen (ontology list, Eqs. 12-13), both obtained by prompted LLM calls whose inputs are patent abstracts and extracted entities; neither uses the gold option label, and no parameter is fitted to the PatentMatch benchmark. Eqs. 11-13 only condition the final matching call on these generated lists as additional context, which is a legitimate two-stage prompting or intermediate-representation design, comparable to chain-of-thought or self-consistency. The matching probability is not set equal to the ontology output by construction, and the ontology output is not a fitted value. The 'memory graph' is loosely operationalized (edges E are never used), and the Table 1 ontology prompt does give the LLM a comparative preview of all options before the final call, so the Sec. 5.2 'OnlyZGen' ablation does not fully isolate the graph-structure mechanism from this pre-view effect; however, that is a confound in causal attribution, not a circular reduction. The paper's self-citations (e.g., refs. [33] and [57]) appear only as supportive background and are not load-bearing for the central claim. No circular step can be exhibited through an equation-level or fitted-parameter reduction.

Assumptions & free parameters 3 free parameters · 4 assumptions · 1 invented entities

The central claim rests on the empirical effectiveness of prompt-generated entities and classifications. No numeric parameters are fitted to the benchmark, but several design choices (top-k, entity count, ontology depth) are set by hand. The theoretical framing relies on unverified assumptions about LLM memory structure and the value of IPC-like ontologies.

free parameters (3)
  • top_k_retrieved = 3
    Number of retrieved patents fed to the generator; chosen without reported tuning or sensitivity analysis (Sec 4).
  • max_entities = 10
    Prompt instructs the LLM to list up to 10 entities; arbitrary cap without justification (Table 1).
  • ontology_depth = 3 levels
    Three-level technical classification (major > subcategory > specific class) chosen as the template; no comparison to 2 or 4 levels (Table 1).
assumptions (4)
  • domain assumption LLM parametric memory is structured as a traversable graph of concepts
    Sec 3.2 states the memory graph originates from parametric memory and that LLMs selectively visit nodes in autoregressive decoding; no direct evidence is provided, and the implementation is standard token generation. This is an empirical premise for the method's narrative.
  • domain assumption IPC-style hierarchical ontologies provide useful signal for patent similarity
    Sec 3.2 and Table 1 ask LLMs to generate three-level classifications; the paper assumes these categories improve matching without testing alternative categorical features.
  • domain assumption A 1,000-question multiple-choice test over 8 IPC sections is a reliable measure of patent matching ability
    Sec 4 describes PatentMatch with 1,000 questions; the paper reports no confidence intervals or repeated-run variance, treating accuracy as deterministic.
  • domain assumption BGE-Base dense retrieval on a 300k corpus is a suitable retriever for the RAG baseline
    Sec 4 sets Top-3 retrieval with BGE-Base; retrieval quality is not analyzed against alternatives.
invented entities (1)
  • Memory Graph
    purpose: To conceptualize LLM parametric memory as a traversable graph from which entities and ontologies are extracted
    The paper introduces the memory graph (Sec 3.2) as a graph-structured representation of parametric memory. No independent evidence of such a graph is provided; the traversal is implemented as ordinary LLM prompts (Eq. 15-16), making the graph entity a framing device rather than a measurable construct.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Enhancing the Patent Matching Capability of Large Language Models via the Memory Graph." pith.science (2026). https://pith.science/paper/3BDTTUT2

@misc{pith2026250414845,
  author       = {Pith},
  title        = {Pith review of: Enhancing the Patent Matching Capability of Large Language Models via the Memory Graph},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/3BDTTUT2}},
  note         = {Machine review of arXiv:2504.14845}
}
read the original abstract

Intellectual Property (IP) management involves strategically protecting and utilizing intellectual assets to enhance organizational innovation, competitiveness, and value creation. Patent matching is a crucial task in intellectual property management, which facilitates the organization and utilization of patents. Existing models often rely on the emergent capabilities of Large Language Models (LLMs) and leverage them to identify related patents directly. However, these methods usually depend on matching keywords and overlook the hierarchical classification and categorical relationships of patents. In this paper, we propose MemGraph, a method that augments the patent matching capabilities of LLMs by incorporating a memory graph derived from their parametric memory. Specifically, MemGraph prompts LLMs to traverse their memory to identify relevant entities within patents, followed by attributing these entities to corresponding ontologies. After traversing the memory graph, we utilize extracted entities and ontologies to improve the capability of LLM in comprehending the semantics of patents. Experimental results on the PatentMatch dataset demonstrate the effectiveness of MemGraph, achieving a 17.68% performance improvement over baseline LLMs. The further analysis highlights the generalization ability of MemGraph across various LLMs, both in-domain and out-of-domain, and its capacity to enhance the internal reasoning processes of LLMs during patent matching. All data and codes are available at https://github.com/NEUIR/MemGraph.

Figures

Figures reproduced from arXiv: 2504.14845 by the authors.

Figure 1
Figure 1. Illustration of Our MemGraph Method. The frame [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. The Illustration of Our MemGraph Method. [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Evaluation Results of Different Models in Patent [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

64 extracted references · 41 canonical work pages

  1. [1]

    Assad Abbas, Limin Zhang, and Samee U Khan. 2014. A Literature Review On the State-of-the-art in Patent Analysis. World Patent Information 37 (2014), 3–13

    work page 2014

  2. [2]

    Biju Paul Abraham and Soumyo D Moitra. 2001. Innovation Assessment through Patent Analysis. Technovation 21 (2001), 245–252

    work page 2001

  3. [3]

    Josh Achiam, Steven Adler, Sandhini Agarwal, Lama Ahmad, Ilge Akkaya, Flo- rencia Leoni Aleman, Diogo Almeida, Janko Altenschmidt, Sam Altman, et al

  4. [4]

    Sam Arts, Bruno Cassiman, and Juan Carlos Gomez. 2018. Text Matching to Measure Patent Similarity. Strategic Management Journal 39 (2018), 62–84

    work page 2018

  5. [5]

    Akari Asai, Zeqiu Wu, Yizhong Wang, Avi Sil, and Hannaneh Hajishirzi. 2024. Self-RAG: Learning to Retrieve, Generate, and Critique through Self-Reflection. In Proceedings of ICLR

    work page 2024

  6. [6]

    Grazia Sveva Ascione and Valerio Sterzi. 2024. A Comparative Analysis of Embedding Models for Patent Similarity. ArXiv preprint abs/2403.16630 (2024)

    arXiv 2024

  7. [7]

    Zilong Bai, Ruiji Zhang, Linqing Chen, Qijun Cai, Yuan Zhong, Cong Wang Yan Fang, Jie Fang, Jing Sun, et al. 2024. PatentGPT: A Large Language Model for Intellectual Property. ArXiv preprint abs/2404.18255 (2024)

    arXiv 2024

  8. [8]

    Yoshua Bengio, Réjean Ducharme, and Pascal Vincent. 2000. A Neural Probabilis- tic Language Model. In Proceedings of NeurIPS

    work page 2000

Show all 64 references
  1. [9]

    Deng Cai, Yan Wang, Lemao Liu, and Shuming Shi. 2022. Recent Advances in Retrieval-Augmented Text Generation. In Proceedings of SIGIR. 3417–3419

    2022

  2. [10]

    Gaetano Cascini and Manuel Zini. 2008. Measuring Patent Similarity by Com- paring Inventions Functional Trees. In Proceedings of the Second Topical Session on Computer-Aided Innovation. 31–42

    2008

  3. [11]

    Renukswamy Chikkamath, Deepak Rastogi, Mahesh Maan, and Markus Endres

  4. [12]

    Jung-Mei Chu, Hao-Cheng Lo, Jieh Hsiang, and Chun-Chieh Cho. 2024. Patent Response System Optimised for Faithfulness: Procedural Knowledge Embodiment with Knowledge Graph and Retrieval Augmented Generation. In Proceedings of the 1st Workshop on Towards Knowledgeable Language Models

    2024

  5. [13]

    Allan M Collins and Elizabeth F Loftus. 1975. A Spreading-activation Theory of Semantic Processing. Psychological Review 82 (1975), 407

    1975

  6. [14]

    Bruce Croft, Donald Metzler, and Trevor Strohman. 2010. Search Engines: Infor- mation Retrieval in Practice . Vol. 520. Addison-Wesley Reading

    2010

  7. [15]

    Abhimanyu Dubey, Abhinav Jauhri, Abhinav Pandey, Abhishek Kadian, Ahmad Al-Dahle, Aiesha Letman, Akhil Mathur, Alan Schelten, et al. 2024. The Llama 3 Herd of Models. ArXiv preprint abs/2407.21783 (2024)

    2024 arXiv

  8. [16]

    Furnas, Thomas K

    George W. Furnas, Thomas K. Landauer, Louis M. Gomez, and Susan T. Dumais

  9. [17]

    Yunfan Gao, Yun Xiong, Xinyu Gao, Kangxiang Jia, Jinliu Pan, Yuxi Bi, Yi Dai, Jiawei Sun, and Haofen Wang. 2023. Retrieval-Augmented Generation for Large Language Models: A Survey. ArXiv preprint abs/2312.10997 (2023)

    2023 arXiv

  10. [18]

    Eugene Garfield. 1966. Patent Citation Indexing and the Notions of Novelty, Similarity, and Relevance. Journal of Chemical Documentation 6 (1966), 63–65

    1966

  11. [19]

    Team GLM, Aohan Zeng, Bin Xu, Bowen Wang, Chenhui Zhang, Da Yin, Diego Rojas, Guanyu Feng, et al. 2024. ChatGLM: A Family of Large Language Models from GLM-130B to GLM-4 All Tools. ArXiv preprint abs/2406.12793 (2024)

    2024 arXiv

  12. [20]

    Stuart JH Graham, Alan C Marco, and Richard Miller. 2018. The USPTO Patent Examination Research Dataset: A Window on Patent Processing. Journal of Economics & Management Strategy 27 (2018), 554–578

    2018

  13. [21]

    Daniel S Hain, Roman Jurowetzki, Tobias Buchmann, and Patrick Wolf. 2022. A Text-embedding-based Approach to Measuring Patent-to-Patent Technological Similarity. Technological Forecasting and Social Change 177 (2022), 121559

    2022

  14. [22]

    Femi Halgin and Ahmed Taiye Mohammedand. 2024. Intelligent Patent Pro- cessing: Leveraging Retrieval-Augmented Generation for Enhanced Consultant Services. In Proceedings of International Symposium on Digital Transformation

    2024

  15. [23]

    Aaron Hurst, Adam Lerer, Adam P Goucher, Adam Perelman, Aditya Ramesh, Aidan Clark, AJ Ostrow, Akila Welihinda, Alan Hayes, Alec Radford, et al. 2024. GPT-4o System Card. ArXiv preprint abs/2410.21276 (2024)

    2024 arXiv

  16. [24]

    Kishore Vama Indukuri, Anurag Anil Ambekar, and Ashish Sureka. 2007. Simi- larity Analysis of Patent Claims using Natural Language Processing Techniques. In Proceedings of ICCMA, Vol. 4. IEEE, 169–175

    2007

  17. [25]

    Lekang Jiang and Stephan Goetz. 2024. Artificial Intelligence Exploring the Patent Field. ArXiv preprint abs/2403.04105 (2024)

    2024 arXiv

  18. [26]

    Lekang Jiang, Caiqi Zhang, Pascal A Scherz, and Stephan Goetz. 2024. Can Large Language Models Generate High-quality Patent Claims? ArXiv preprint abs/2406.19465 (2024)

    2024 arXiv

  19. [27]

    Zhengbao Jiang, Frank Xu, Luyu Gao, Zhiqing Sun, Qian Liu, Jane Dwivedi- Yu, Yiming Yang, Jamie Callan, and Graham Neubig. 2023. Active Retrieval Augmented Generation. In Proceedings of EMNLP. Singapore, 7969–7992

    2023

  20. [28]

    John L King. 2003. Patent Examination Procedures and Patent Quality. Patents in the Knowledge-based Economy 54 (2003)

    2003

  21. [29]

    Takeshi Kojima, Shixiang Shane Gu, Machel Reid, Yutaka Matsuo, and Yusuke Iwasawa. 2022. Large Language Models are Zero-Shot Reasoners. In Proceedings of NeurIPS

    2022

  22. [30]

    Ralf Krestel, Renukswamy Chikkamath, Christoph Hewel, and Julian Risch. 2021. A Survey on Deep Learning for Patent Analysis. World Patent Information 65 (2021)

    2021

  23. [31]

    Patrick S. H. Lewis, Ethan Perez, Aleksandra Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich Küttler, Mike Lewis, Wen-tau Yih, Tim Rock- täschel, Sebastian Riedel, and Douwe Kiela. 2020. Retrieval-Augmented Genera- tion for Knowledge-Intensive NLP Tasks. In Pr...

    2020

  24. [32]

    Loet Leydesdorff. 2008. Patent classifications as indicators of intellectual organi- zation. Journal of the American Society for Information Science and Technology 59, 10 (2008), 1582–1597

    2008

  25. [33]

    Xinze Li, Sen Mei, Zhenghao Liu, Yukun Yan, Shuo Wang, Shi Yu, Zheni Zeng, Hao Chen, Ge Yu, Zhiyuan Liu, et al . 2025. RAG-DDR: Optimizing Retrieval- Augmented Generation Using Differentiable Data Rewards. In Proceedings of ICLR

    2025

  26. [34]

    Mihai Lupu, Katja Mayer, Noriko Kando, and Anthony J Trippe. 2017. Current Challenges in Patent Information Retrieval . Vol. 37. Springer

    2017

  27. [35]

    Jonathan S Masur. 2010. Costly Screens and Patent Examination. Journal of Legal Analysis 2 (2010), 687–734

    2010

  28. [36]

    Shiwen Ni, Minghuan Tan, Yuelin Bai, Fuqiang Niu, Min Yang, Bowen Zhang, Ruifeng Xu, Xiaojun Chen, Chengming Li, and Xiping Hu. 2024. MoZIP: A Multilingual Benchmark to Evaluate Large Language Models in Intellectual Property. In Proceedings of LREC-COLING. 11658–11668

    2024

  29. [37]

    Hyunseok Park, Janghyeok Yoon, and Kwangsoo Kim. 2012. Identifying Patent In- fringement using SAO based Semantic Technological Similarities. Scientometrics 90 (2012)

    2012

  30. [38]

    Sergio Pelaez, Gaurav Verma, Barbara Ribeiro, and Philip Shapira. 2024. Large- Scale Text Analysis using Generative Language Models: A Case Study in Dis- covering Public Value Expressions in AI Patents. Quantitative Science Studies 5 (2024), 153–169

    2024

  31. [39]

    Zhuoyi Peng and Yi Yang. 2024. Connecting the Dots: Inferring Patent Phrase Similarity with Retrieved Phrase Graphs. In Findings of NAACL 2024 . Mexico City, Mexico, 1877–1890

    2024

  32. [40]

    Vinay Venkatesh Ramasesh, Aitor Lewkowycz, and Ethan Dyer. 2022. Effect of Scale on Catastrophic Forgetting in Neural Networks. In Proceedings of ICLR

    2022

  33. [41]

    Runtao Ren and Jian Ma. 2024. PatentGPT: A Large Language Model for Patent Drafting Using Knowledge-based Fine-tuning Method. ArXiv preprint abs/2409.00092 (2024)

    2024 arXiv

  34. [42]

    Julian Risch, Nicolas Alder, Christoph Hewel, and Ralf Krestel. 2020. Patentmatch: A Dataset for Matching Patent Claims & Prior Art. ArXiv preprint abs/2012.13919 (2020)

    2020 arXiv

  35. [43]

    Arousha Haghighian Roudsari, Jafar Afshar, Suan Lee, and Wookey Lee. 2021. Comparison and Analysis of Embedding Methods for Patent Documents. In Proceedings of BigComp. IEEE, 152–155

    2021

  36. [44]

    Sagar Srinivas Sakhinana, Venkataramana Runkana, et al. 2024. Towards Au- tomated Patent Workflows: AI-Orchestrated Multi-Agent Framework for In- tellectual Property Management and Analysis. In NeurIPS 2024 Workshop on Open-World Agents

    2024

  37. [45]

    L Siddharth and Jianxi Luo. 2024. Retrieval Augmented Generation using Engi- neering Design Knowledge. Knowledge-Based Systems 303 (2024), 112410

    2024

  38. [46]

    Hugo Touvron, Thibaut Lavril, Gautier Izacard, Xavier Martinet, Marie-Anne Lachaux, Timothée Lacroix, Baptiste Rozière, Naman Goyal, Eric Hambro, Faisal Azhar, et al. 2023. Llama: Open and Efficient Foundation Language Models. ArXiv preprint abs/2302.13971 (2023)

    2023 arXiv

  39. [47]

    Yuen-Hsien Tseng, Chi-Jen Lin, and Yu-I Lin. 2007. Text Mining Techniques for Patent Analysis. Information Processing & Management 43 (2007), 1216–1247

    2007

  40. [48]

    Anna Maria Villa and Manuel Wirz. 2022. A Sequential Patent Search Approach Combining Semantics and Artificial Intelligence to Identify Initial State-of-the- Art Documents. World Patent Information 68 (2022), 102096

    2022

  41. [49]

    Juanyan Wang, Sai Krishna Reddy Mudhiganti, and Manali Sharma. 2024. Patent- former: A Novel Method to Automate the Generation of Patent Applications. In Proceedings of EMNLP: Industry Track . 1361–1380

    2024

  42. [50]

    Jason Wei, Yi Tay, Rishi Bommasani, Colin Raffel, Barret Zoph, Sebastian Borgeaud, Dani Yogatama, Maarten Bosma, Denny Zhou, Donald Metzler, et al. 2022. Emergent Abilities of Large Language Models. ArXiv preprint abs/2206.07682 (2022)

    2022 arXiv

  43. [51]

    Chi, Quoc V

    Jason Wei, Xuezhi Wang, Dale Schuurmans, Maarten Bosma, Brian Ichter, Fei Xia, Ed H. Chi, Quoc V. Le, and Denny Zhou. 2022. Chain-of-Thought Prompting Elicits Reasoning in Large Language Models. In Proceeding of NeurIPS

    2022

  44. [52]

    Yilin Wen, Zifeng Wang, and Jimeng Sun. 2023. Mindmap: Knowledge Graph Prompting Sparks Graph of Thoughts in Large Language Models. ArXiv preprint abs/2308.09729 (2023)

    2023 arXiv

  45. [53]

    Thomas Wolf, Lysandre Debut, Victor Sanh, Julien Chaumond, Clement Delangue, Anthony Moi, Pierric Cistac, Tim Rault, Remi Louf, Morgan Funtowicz, Joe Davison, Sam Shleifer, Patrick von Platen, Clara Ma, Yacine Jernite, Julien Plu, Canwen Xu, Teven Le Scao, Sylvain Gugger, Mari...

    2020

  46. [54]

    Shitao Xiao, Zheng Liu, Peitian Zhang, and Niklas Muennighoff. 2023. C-Pack: Packaged Resources To Advance General Chinese Embedding. ArXiv preprint abs/2309.07597 (2023)

    2023 arXiv

  47. [55]

    Jian Xie, Kai Zhang, Jiangjie Chen, Renze Lou, and Yu Su. 2023. Adaptive Chameleon or Stubborn Sloth: Revealing the Behavior of Large Language Models in Knowledge Conflicts. ArXiv preprint abs/2305.13300 (2023)

    2023 arXiv

  48. [56]

    Rongwu Xu, Zehan Qi, Zhijiang Guo, Cunxiang Wang, Hongru Wang, Yue Zhang, and Wei Xu. 2024. Knowledge conflicts for llms: A survey. ArXiv preprint abs/2403.08319 (2024)

    2024 arXiv

  49. [57]

    Zhipeng Xu, Zhenghao Liu, Yukun Yan, Shuo Wang, Shi Yu, Zheni Zeng, Chaojun Xiao, Zhiyuan Liu, Ge Yu, and Chenyan Xiong. 2024. ActiveRAG: Autonomously Knowledge Assimilation and Accommodation through Retrieval-Augmented Agents. ArXiv preprint abs/2402.13547 (2024)

    2024

  50. [58]

    Xiaoibng Xue and W Bruce Croft. 2009. Transforming patents into prior-art queries. In Proceedings of SIGIR. 808–809

    2009

  51. [59]

    An Yang, Baosong Yang, Binyuan Hui, Bo Zheng, Bowen Yu, Chang Zhou, Cheng- peng Li, Chengyuan Li, Dayiheng Liu, Fei Huang, et al. 2024. Qwen2 Technical Report. ArXiv preprint abs/2407.10671 (2024)

    2024 arXiv

  52. [60]

    An Yang, Baosong Yang, Beichen Zhang, Binyuan Hui, Bo Zheng, Bowen Yu, Chengyuan Li, Dayiheng Liu, Fei Huang, Haoran Wei, et al . 2024. Qwen2. 5 Technical Report. ArXiv preprint abs/2412.15115 (2024)

    2024 arXiv

  53. [61]

    Wayne Xin Zhao, Kun Zhou, Junyi Li, Tianyi Tang, Xiaolei Wang, Yupeng Hou, Yingqian Min, Beichen Zhang, Junjie Zhang, Zican Dong, et al. 2023. A Survey of Large Language Models. ArXiv preprint abs/2303.18223 (2023)

    2023 arXiv

  54. [1987]

    The Vocabulary Problem in Human-system Communication. Commun. ACM 30 (1987), 964–971

    1987

  55. [2023]

    ArXiv preprint abs/2303.08774 (2023)

    GPT-4 Technical Report. ArXiv preprint abs/2303.08774 (2023)

    2023 arXiv

  56. [2024]

    World Patent Information 76 (2024), 102254

    Is Your Search Query Well-formed? A Natural Query Understanding for Patent Prior Art Search. World Patent Information 76 (2024), 102254

    2024

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.