Pith. sign in

REVIEW 4 major objections 4 minor 81 references

CancerKG.ORG A Web-scale, Interactive, Verifiable Knowledge Graph-LLM Hybrid for Assisting with Optimal Cancer Treatment and Care

T0 review · 4 major / 4 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read The paper claims that CancerKG, a web-scale knowledge-graph/large-language-model hybrid built largely automatically from 2.5 million peer-reviewed colorectal cancer publications, serves user needs better than any component alone, with the…

desk verdict CancerKG is a useful system description whose superiority and safety claims are unsupported; give it a serious referee to force an honest evaluation. read the letter →

arxiv 2501.00223 v1 pith:VLTD7Z3V submitted 2024-12-31 cs.AI cs.IRcs.LG

Michael Gubanov , Anna Pyayt , Aleksandra Karolak This is my paper · ORCID
classification cs.AIcs.IRcs.LG
keywords knowledgegraphlargelanguagemodelretrieval-augmentedgenerationcolorectalcancermedicalinformationretrievaltablesearchhallucinationmitigationfusion
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

CancerKG.ORG is a web-scale hybrid that joins a knowledge graph, built largely automatically from the latest peer-reviewed colorectal cancer literature after a small manual seed, with fine-tuned large language models; the paper's central claim is that this combination serves users better than any component alone. The authors argue that large language models hallucinate, forget, and train on outdated data, while manually curated medical knowledge graphs quickly become stale, so they make the verified graph act as a retrieval-augmented generation guardrail that keeps answers traceable to cited publications. If the claim holds, clinicians, researchers, and patients could get current, actionable cancer information without the staleness of curated databases or the fabrication risk of bare language models. Colorectal cancer is presented as a model domain, with the architecture claimed to transfer to other scientific and medical areas.

What carries the argument

The central object is the verified hierarchical knowledge graph: a scalable graph database seeded with a small vetted set of nodes and edges, then continuously enriched by a fusion process. Fusion matches the root of each newly extracted subtree to existing graph nodes through normalized term matching, amended by embedding-driven matching for terms never seen before; subtrees with several layers and low-confidence matches are kept separate or sent to expert review, and the fusion model is expected to learn from expert corrections over time. The other load-bearing mechanism is the retrieval loop around the graph: a conversational query parser splits a natural-language question into a structural table query and a text query, a table search engine ranks results with custom table embeddings, and a fine-tuned large language model generates the final answer using the retrieved evidence. Supporting machinery includes binary topical table classifiers built on fine-tuned biomedical language-model embeddings and the 3D meta-profile, a summary of a table cluster's horizontal and vertical metadata used for browsing and filtering.

What would settle it

Take a set of 100 clinical questions with known answers from recently published colorectal cancer trials, then check whether every cited graph node exists, matches the cited publication, and supports the language model's answer; if even a small percentage of answers cite nodes absent from the graph or contradict the cited paper, the verifiability and anti-hallucination claim collapses.

Watch

Extended reading notes

Core claim

The paper reports construction of one of the first web-scale knowledge-graph/large-language-model hybrids for a cancer domain, storing more than 2.5 million parsed, peer-reviewed publications and organizing them into a hierarchical, browsable graph. The discovery at the center of the paper is that an automated pipeline can keep such a graph current and can use it to discipline the language model: a small vetted graph is seeded manually, topical clusters of tables are extracted by trained classifiers, new knowledge is fused in by matching normalized terms and embedding vectors, and the resulting verified graph is used as a retrieval-augmented generation guardrail so answers point back to the literature. The authors report approximately 95% F-measure for the topical table classifiers under 10-fold cross-validation and describe the system as currently being evaluated for medical research and clinical information retrieval tasks at a major U.S. cancer center. Five interfaces are presented as serving different data modalities: interactive graph browsing, publication and structural table search engines, a conversational interface, and 3D meta-profiles.

Load-bearing premise

The load-bearing premise is that the automatic extraction-and-fusion pipeline keeps the graph medically correct: if embedding similarity plus occasional expert review of low-confidence merges cannot prevent wrong or outdated nodes from entering the graph, the verified guardrail would lock in and amplify errors rather than prevent them.

Editorial extensions

If this is right

  • Users can trace every answer to specific peer-reviewed publications, because the graph constrains generation to verified content.
  • Because ingestion is automatic, the system can stay current in a way fixed-training-cutoff language models and manually curated databases cannot.
  • Structural table search lets clinicians query concrete clinical attributes such as tumor size and lymph-node status instead of relying on keyword search over whole documents.
  • If the architecture is truly domain-independent, the same pipeline can be retrained for other fast-moving medical or scientific fields.
  • The verified-graph guardrail idea applies to any setting where language-model answers must be checked against an evolving corpus.

Reading between the lines

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

  • The decisive experiment the paper does not report is a head-to-head comparison on identical clinical questions between the hybrid, a bare large language model, and a keyword search engine, scored against expert-verified answers; that comparison would directly test the better-than-any-part-in-isolation claim.
  • Because the paper reports only that low-confidence fusions go to an expert and that mistakes are later learned, the unsupervised label is partial; measuring the fraction of graph updates that require human sign-off would reveal the true maintenance cost.
  • Embedding-driven fusion could merge or confuse distinct entities whose names are similar, so a targeted evaluation of entity identity under synonym and spelling variation would show whether graph precision holds at scale.
  • The same guardrail design could plausibly transfer to other clinical domains, but only if each domain can supply the seed nodes and table schemas the pipeline needs.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 4 minor

Summary. The paper describes CancerKG, a web-scale knowledge graph (KG) and large language model (LLM) hybrid for colorectal cancer, with automatic extraction and fusion from PubMed, multiple user interfaces (interactive KG, structural table search, conversational interface, 3D meta-profiles), and a GRU-based table topic classifier. The central claims are that the verified KG acts as a RAG guardrail preventing LLM hallucination and that the hybrid serves users better than an LLM, a KG, or a search engine in isolation. The only quantitative result is an approximately 95% F-measure for topical table classification (§4.3); no end-to-end evaluation, baseline comparison, user study, or hallucination measurement is provided.

Significance. If the central claims were substantiated, the system would be a valuable contribution: an automatically updated, verifiable medical KG-LLM hybrid that addresses hallucination and knowledge freshness at web scale. The paper also demonstrates engineering effort, including structural search over complex medical tables and interactive meta-profiles, and it builds on prior COVIDKG work. However, the current evidence does not support the load-bearing claims. The evaluation is confined to a single component classifier, trained and validated on the authors' own data, and the 'unsupervised' and 'verifiable' characterizations are contradicted by the design described in the manuscript. The paper is a plausible systems description, but it falls far short of demonstrating the claimed benefits.

major comments (4)
  1. [Abstract; §1; §3.3] The central comparative and safety claims are not evaluated. The abstract and introduction assert that the hybrid 'serves the user needs better than just an LLM, KG or a search-engine in isolation' and that the verified KG prevents hallucination, but no comparison against these baselines, no hallucination measurement, and no user study are reported. The only quantitative result (§4.3) is approximately 95% F-measure for topical table classification, which does not speak to answer correctness or KG-grounded generation. Add an end-to-end evaluation with expert or ground-truth assessment, comparing the hybrid against LLM-only, KG-only, and search-only conditions, and report factual accuracy and hallucination rates.
  2. [§2; §5.1; §5.2; §7] The 'unsupervised' and 'verifiable' characterizations are contradicted by the design. Section 2 requires a Data Scientist to initialize a seed KG with 10–20 nodes, §5.1 repeats this dependency, and §5.2 states that low-confidence fusions are evaluated by an expert. The conclusion in §7 that content is 'extracted and updated in unsupervised manner from PubMed.com' is therefore inaccurate. Moreover, the fusion process's correctness is unmeasured: there are no precision/recall figures for extracted nodes and edges, no counts or outcomes of expert interventions, and no evidence for the claim that mistakes 'will be learned by the fusion model.' Clarify the actual degree of supervision and report fusion accuracy on a validation set.
  3. [§4.3] The evaluation of the topical table classifier is circular with respect to the clustering claim. The positive training set is selected as tables within an empirical angle (18 degrees) of data-scientist-chosen centroid vectors, and the same centroids define the final clusters. This guarantees strong agreement between the classifier and the cluster definitions. The reported 'approximately 95% F-measure' also lacks error bars, statistical significance, and external validation. Evaluate the classifier on an independent, human-labeled test set and report variance.
  4. [§5.2; Abstract] The RAG guardrail claim assumes that the KG is a reliable repository of verified medical facts, but the manuscript does not measure KG correctness. The enrichment pipeline uses normalized NLP term matching and embedding-driven matching, with expert review only for low-confidence fusions; no estimate is given of how often expert review is needed or how accurate the resulting graph is. If the KG contains errors, grounding the LLM in it could reinforce false or stale information while giving an appearance of verifiability. Report KG accuracy on a sample of PubMed-derived facts with expert adjudication, and show that KG-grounded outputs are more accurate than those of the unaided LLM.
minor comments (4)
  1. [§2] The phrase 'Large Language Models (LMM)' should be 'Large Language Models (LLM)'.
  2. [References] References [2] and [3] appear mismatched: [2] is listed as the COVID-19 Infectious Disease Ontology, yet it is cited for LLM hallucinations, and [3] is COVIDScholar, cited for catastrophic forgetting. Verify the intended sources and correct the citations.
  3. [Throughout] Spelling of 'Moffitt' is inconsistent between the author affiliation and the body text; similarly, 'Llama 2' / 'LLama 2' variants should be normalized.
  4. [§4.2] The paper mentions a '100,000 dimensional feature space' and also '100K English terms in our vocabulary'; clarify whether the feature dimension equals the vocabulary size and how the 100K was selected beyond frequency sorting.

Circularity Check

2 steps flagged · score 6.0 of 10

CancerKG's 'verified KG guardrail' is asserted by definition from PubMed peer-review status, and the only reported accuracy (95% F-measure) is a classifier trained and evaluated on labels generated by the same embedding-distance criterion it is meant to validate.

  1. fitted input called prediction [Section 4.3, Training and Evaluation]
    "asking a clinical Data Scientist to select a representative (i.e. centroid) table for each topic ... Second, we created a composite embedding vector corresponding to each topical centroid table ... we take a centroid vector and select only the tables in the dataset within a 18 degrees from it ... Fourth, we train our GRU model as a binary topic classifier on these tables (as positively labeled) amended with the same number of random tables from the dataset (as negatively labeled). ... observed approximately 95% F-measure, when validated with 10-fold cross-validation."

    The reported 95% F-measure is offered as evidence for the topical clustering that seeds the KG, but the positive training labels were produced by the same centroid-embedding and 18-degree similarity rule that the system uses to create clusters. The classifier is trained to reproduce the embedding-similarity criterion that generated its own labels, so the evaluation measures internal self-consistency rather than agreement with any independent annotation or external benchmark. No fusion error rates, expert adjudication counts, or end-to-end KG precision are reported, yet this number is the paper's only quantitative validation of the KG-construction pipeline that underpins the 'verified KG' guardrail claim.

  2. self definitional [Abstract and Section 7 (Conclusion)]
    "To alleviate the LLMs shortcomings, the verified KG serves as a Retrieval Augmented Generation (RAG) guardrail. ... The content is extracted and updated in unsupervised manner from PubMed.com that contains vetted, peer-reviewed medical publications. Hence is verifiable and contains the most up to date medical practices."

    The guardrail claim depends entirely on the KG being 'verified,' but the only support offered for verification is the source: peer-reviewed PubMed publications. 'Verifiable' is thus equated, by definition, with 'from PubMed,' rather than demonstrated by measuring KG precision, fusion error rates, or answer accuracy. The paper's own Section 2 and Section 5.1 contradict the 'unsupervised' framing by requiring a Data Scientist to seed the KG, and Section 5.2 requires expert review of low-confidence fusions while acknowledging mistakes will only later be 'learned by the fusion model.' Consequently, the no-hallucination guardrail is an assertion restating its premise, not a derived or measured result.

full rationale

The paper is a system description rather than a derivation, so most of the load-bearing assertions (hybrid advantage, no hallucinations, usefulness at Moffitt) are unsupported claims rather than circular derivations. Two concrete circularities are present. First, the only quantitative result—95% F-measure for topical table classification—is computed on a training set whose positive labels were selected by the same embedding-distance criterion (centroid tables, 18-degree threshold) that the pipeline uses to form clusters; the 'prediction' therefore measures reproduction of its own labeling rule. Second, the central 'verified KG as RAG guardrail' claim defines 'verified' as 'peer-reviewed PubMed content' and then concludes 'Hence is verifiable,' without measuring the KG's correctness; this is a definitional substitution rather than evidence. The paper also leans on the authors' prior COVIDKG.ORG work for the architecture and classifier, but that self-citation is not the main source of circularity. Because the central guardrail claim reduces to a definition and the only evaluation reduces to a self-constructed fit, a partial-circularity score of 6 is appropriate. The remainder of the paper's validity gap (no external benchmark, no comparison against LLM-only answers, no fusion precision numbers) is a correctness/evidence concern, not circularity.

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

The central claims rest on human-chosen seed graphs, empirically tuned clustering thresholds, and an unvalidated assumption that automatic fusion produces a trustworthy KG. These are not derived from first principles and are not tested against an external gold standard.

free parameters (4)
  • Centroid angle threshold for table clustering = 18 degrees
    Section 4.3: 'within a 18 degrees from it (determined empirically)'; this threshold shapes which tables are assigned to each topic cluster.
  • Vocabulary size for feature space = 100,000 terms
    Section 4.2: vocabulary selected by frequency cutoff; dimensionality choice affects classification and was set to keep training time manageable.
  • Dense layer size in GRU classifier = 32 units
    Section 4.1: architecture choice for the binary table topic classifier; no ablation is provided.
  • Seed KG size = 10-20 nodes
    Section 2, Figure 1 item 1: a Data Scientist manually initializes the KG; this human choice anchors the graph hierarchy.
assumptions (4)
  • domain assumption PubMed peer-reviewed publications constitute vetted medical knowledge suitable for clinical guidance.
    Abstract and Section 5.2 rely on PubMed as the source of 'vetted' knowledge; peer review is treated as sufficient for correctness.
  • domain assumption Embedding-based and NLP matching of table metadata to graph nodes preserves clinical meaning across synonyms and new terms.
    Section 5.2 describes fusion through normalized NLP term matching and embedding vectors, assuming similarity implies semantic correctness.
  • domain assumption The seed KG extracted from PrimeKG and static ontologies provides a reliable structural backbone.
    Section 5.1 initializes the graph from PrimeKG and static ontologies without evaluating coverage or correctness for colorectal cancer.
  • ad hoc to paper A KG guardrail prevents LLM hallucination without introducing its own errors.
    Section 1 asserts the KG 'does not hallucinate' and acts as a RAG guardrail; no formal or empirical verification is provided for this claim.
invented entities (1)
  • 3D Meta-profile
    purpose: A visualization and dynamic filtering interface that summarizes table clusters by horizontal and vertical metadata and lets users drill down.
    Introduced in prior work [41, 63]; here it is presented as part of CancerKG but without user evaluation or external validation of its usefulness.

how reviews work

0 comments
Cite this review

Pith. "Pith review of CancerKG.ORG A Web-scale, Interactive, Verifiable Knowledge Graph-LLM Hybrid for Assisting with Optimal Cancer Treatment and Care." pith.science (2026). https://pith.science/paper/VLTD7Z3V

@misc{pith2026250100223,
  author       = {Pith},
  title        = {Pith review of: CancerKG.ORG A Web-scale, Interactive, Verifiable Knowledge Graph-LLM Hybrid for Assisting with Optimal Cancer Treatment and Care},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/VLTD7Z3V}},
  note         = {Machine review of arXiv:2501.00223}
}
read the original abstract

Here, we describe one of the first Web-scale hybrid Knowledge Graph (KG)-Large Language Model (LLM), populated with the latest peer-reviewed medical knowledge on colorectal Cancer. It is currently being evaluated to assist with both medical research and clinical information retrieval tasks at Moffitt Cancer Center, which is one of the top Cancer centers in the U.S. and in the world. Our hybrid is remarkable as it serves the user needs better than just an LLM, KG or a search-engine in isolation. LLMs as is are known to exhibit hallucinations and catastrophic forgetting as well as are trained on outdated corpora. The state of the art KGs, such as PrimeKG, cBioPortal, ChEMBL, NCBI, and other require manual curation, hence are quickly getting stale. CancerKG is unsupervised and is capable of automatically ingesting and organizing the latest medical findings. To alleviate the LLMs shortcomings, the verified KG serves as a Retrieval Augmented Generation (RAG) guardrail. CancerKG exhibits 5 different advanced user interfaces, each tailored to serve different data modalities better and more convenient for the user.

Figures

Figures reproduced from arXiv: 2501.00223 by the authors.

Figure 1
Figure 1. CancerKG.ORG Architecture [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. A Small Fragment of the Interactive CancerKG Knowledge Graph. [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. CancerKG Structural Search over Tables based on the LLM Conversational Interface. [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Deep-learning Architecture for Topical Table Classification [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

81 extracted references · 72 canonical work pages

  1. [1]

    https://www.bindingdb.org/rwd/bind/index.jsp

    BindingDB. https://www.bindingdb.org/rwd/bind/index.jsp

  2. [2]

    https://www.ebi.ac.uk/ols/ontologies/idocovid19

    The COVID -19 Infectious Disease Ontology. https://www.ebi.ac.uk/ols/ontologies/idocovid19

    work page

  3. [3]

    https://covidscholar.org/stats

    COVIDScholar. https://covidscholar.org/stats

    work page

  4. [4]

    https://go.drugbank.com/

    DrugBank. https://go.drugbank.com/

    work page

  5. [5]

    https://enamine.net/

    Enamine. https://enamine.net/

    work page

  6. [6]

    https://www.rcsb.org/

    Protein Data Bank (PDB). https://www.rcsb.org/

    work page

  7. [7]

    https://rdf4j.org/documentation/reference/sail/

    RDF4J. https://rdf4j.org/documentation/reference/sail/

    work page

  8. [8]

    https://www.researchrabbit.ai/

    researchrabbit. https://www.researchrabbit.ai/

    work page

Show all 81 references
  1. [9]

    https://covidsearch.sinequa.com/app/covid - search/#/home

    Sinequa. https://covidsearch.sinequa.com/app/covid - search/#/home

  2. [10]

    The Virus Infectious Disease Ontology

    [n.d.]. The Virus Infectious Disease Ontology. https://www.ebi.ac.uk/ols/ontologies/vido

  3. [11]

    [n.d.]. Zinc. https://zinc.docking.org/

  4. [12]

    online: http://www.mongodb.com

    2007. online: http://www.mongodb.com. http://www.mongodb.com

    2007

  5. [13]

    online: http://lucene.apache.org/core/

    2011. online: http://lucene.apache.org/core/. http://lucene.apache.org/core/

    2011

  6. [14]

    online:The National Science Foundation’s Innovation Corps (I - Corps™) program

    2022. online:The National Science Foundation’s Innovation Corps (I - Corps™) program. https://www.nsf.gov/news/special_reports/i - corps/

    2022

  7. [15]

    Aryn -Ai/Sycamore: Sycamore is an LLM -Powered Search and Analytics Platform for Unstructured Data

    2024. Aryn -Ai/Sycamore: Sycamore is an LLM -Powered Search and Analytics Platform for Unstructured Data. https://github.com/aryn - ai/sycamore

    2024

  8. [16]

    Martí n Abadi. 2015. TensorFlow: Large -Scale Machine Learning on Heteroge- neous Systems. http://tensorflow.org/ Software available from tensorflow.org

    2015

  9. [17]

    Ilyas, Michael Stonebraker, Paolo Papotti, and Mourad Ouzzani

    Zia Abedjan, John Morcos, Michael Gubanov, Ihab F. Ilyas, Michael Stonebraker, Paolo Papotti, and Mourad Ouzzani. 2014. DATAXFORMER: Leveraging the Web for Semantic Transformations. In CIDR

    2014

  10. [18]

    Meta AI. 2023. LLama 2 Large Language Model. https://ai.meta.com/llama/

    2023

  11. [19]

    Herna ndez, C

    Bogdan Alexe, Michael Gubanov, Mauricio A. Herna ndez, C. T. Howard Ho, Jen-Wei Huang, Yannis Katsis, Lucian Popa, Barna Saha, and Ioana Stanoi

  12. [20]

    Hernandez, Howard Ho, Jen - Wei Huang, Yannis Katsis, and Lucian Popa

    Bogdan Alexe, Michael Gubanov, Mauricio A. Hernandez, Howard Ho, Jen - Wei Huang, Yannis Katsis, and Lucian Popa. 2009. Simplifying Information Integra- tion: Object-Based Flow-of-Mappings Framework for Integration. In Business Intelligence for the Real Time Enterprise . Springer

    2009

  13. [21]

    Xin, Cheng Lian, Yin Huai, Davies Liu, Joseph K.Bradley, Xiangrui Meng, Tomer Kaftan, Michael J

    Michael Armbrust, Reynold S. Xin, Cheng Lian, Yin Huai, Davies Liu, Joseph K.Bradley, Xiangrui Meng, Tomer Kaftan, Michael J. Franklin, Ali Ghodsi, and Matei Zaharia. 2015. Spark SQL: Relational Data Processing in Spark. In ACM SIGMOD 2015

    2015

  14. [22]

    Sören Auer, Christian Bizer, Georgi Kobilarov, Jens Lehmann, Richard Cyga- niak, and Zachary Ives. 2007. DBpedia: A Nucleus for a Web of Open Data. In ISWC’07/ASWC’07

    2007

  15. [23]

    Bradley R. Bebee, Daniel Choi, Ankit Gupta, Andi Gutmans, Ankesh Khandelwal, Yigit Kiran, Sainath Mallidi, Bruce McGaughy, Mike Personick, Karthik Rajan, Si - mone Rondelli, Alexander Ryazanov, Michael Schmidt, Kunal Sengupta, Bryan B. Thompson, Divij Vaidya, and Shawn Wang. 2...

    2018

  16. [24]

    Zohra Bellahsene, Angela Bonifati, and Erhard Rahm. 2011. Schema Matching and Mapping. In Springer

    2011

  17. [25]

    Gross, Selcuk Onur Sumer, Bu lent Arman Aksoy, Anders Jacobsen, Caitlin J

    Ethan Cerami, Jianjiong Gao, Ugur Dogrusoz, Benjamin E. Gross, Selcuk Onur Sumer, Bu lent Arman Aksoy, Anders Jacobsen, Caitlin J. Byrne, Michael L. Heuer, Erik Larsson, Yevgeniy Antipin, Boris Reva, Arthur P. Goldberg, Chris Sander, and Nikolaus S chultz. 2012. The cBio Cance...

    2012 doi

  18. [26]

    Payal Chandak, Kexin Huang, and Marinka Zitnik. 2023. Building a knowledge graph to enable precision medicine. Scientific Data 10, 1 (2023), 67. https://doi.org/10.1038/s41597- 023- 01960- 3

    2023 doi

  19. [27]

    Chi, Jeff Dean, Jacob Devlin, Adam Roberts, Denny Zhou, Quoc V

    Hyung Won Chung, Le Hou, Shayne Longpre, Barret Zoph, Yi Tay, William Fedus, Yunxuan Li, Xuezhi Wang, Mostafa Dehghani, Siddhartha Brahma, Albert Web - son, Shixiang Shane Gu, Zhuyun Dai, Mirac Suzgun, Xinyun Chen, Aakanksha Chowdhery, Alex Castro -Ros, M arie Pellat, Kevin Ro...

  20. [28]

    E. F. Codd. 1983. A Relational Model of Data for Large Shared Data Banks. CACM 26, 1 (Jan. 1983), 64–69

    1983

  21. [29]

    Breda Corish. 2018. Medical knowledge doubles every few months; how can clinicians keep up. Elsevier Connect (2018). CancerKG.ORG–A Verifiable, Scalable Knowledge Graph-LLM hybrid for Assisting with Optimal Cancer Treatment and Care

    2018

  22. [30]

    Google Developers. 2012. Google Knowledge Graph. https://developers.google. com/knowledge- graph

    2012

  23. [31]

    Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. 2019. BERT: Pre -training of Deep Bidirectional Transformers for Language Understanding. In arXiv:1810.04805v2

    2019 arXiv

  24. [32]

    Christiane Fellbaum (Ed.). 1998. WordNet: An Electronic Lexical Database (Language, Speech, and Communication). The MIT Press

    1998

  25. [33]

    Bellis, A

    Anna Gaulton, Louisa J. Bellis, A. Patricia Bento, Jon Chambers, Mark Davies, Anne Hersey, Yvonne Light, Shaun McGlinchey, David Michalovich, Bissan Al- Lazikani, and John P. Overington. 2012. ChEMBL: A large -scale bioactivity database for drug discovery. Nucleic acids resear...

    2012 doi

  26. [34]

    Anna Lisa Gentile, Petar Ristoski, Steffen Eckel, Dominique Ritze, and Heiko Paulheim. 2017. Entity Matching on Web Tables: a Table Embeddings approach for Blocking. In EDBT

    2017

  27. [35]

    Michael Gubanov. 2017. Hybrid: A Large -scale In-memory Image Analytics System. In CIDR

    2017

  28. [36]

    Gubanov, C

    M. Gubanov, C. Jermaine, Z. Gao, and S. Luo. 2016. Hybrid: A Large -scale Linear- relational Database Management System. In MIT NEDB

    2016

  29. [37]

    Michael Gubanov, Chris Jermaine, Zekai Gao, and Shangyu Luo. 2016. Hybrid: A Large -scale Linear-relational Database Management System. In MIT Annual DB Conference

    2016

  30. [38]

    Michael Gubanov, Manju Priya, and Maksim Podkorytov. 2017. CognitiveDB: AnIntelligent Navigator for Large-scale Dark Structured Data. In WWW

    2017

  31. [39]

    Gubanov and A

    M. Gubanov and A. Pyayt. 2013. ReadFast: High -relevance Search -engine for BigText. In ACM CIKM

    2013

  32. [40]

    Gubanov and A

    M. Gubanov and A. Pyayt. 2014. Type-aware Web search. In EDBT

    2014

  33. [41]

    Michael Gubanov, Anna Pyayt, and Sophie Pavia. 2022. Visualizing and Querying Large -scale Structured Datasets by Learning Multi -layered 3D Meta-Profiles. In BigData. IEEE

    2022

  34. [42]

    Gubanov, A

    M. Gubanov, A. Pyayt, and L. Shapiro. 2011. ReadFast: Browsing large documents through UFO. In IRI

    2011

  35. [43]

    Michael Gubanov and Linda Shapiro. 2012. Using Unified Famous Objects (UFO) to Automate Alzheimer’s Disease Diagnostics. In BIBM

    2012

  36. [44]

    Michael Gubanov, Linda Shapiro, and Anna Pyayt. 2011. Learning Unified Famous Objects (UFO) to Bootstrap Information Integration. In IRI

    2011

  37. [45]

    Gubanov and M

    M. Gubanov and M. Stonebraker. 2014. Large -scale Semantic Profile Extraction. In EDBT

    2014

  38. [46]

    Gubanov and M

    M. Gubanov and M. Stonebraker. 2014. Text and Structured Data Fusion in Data Tamer at Scale. In ICDE

    2014

  39. [47]

    Gubanov and Philip A

    Michael N. Gubanov and Philip A. Bernstein. 2006. Structural text search and comparison using automatically extracted schema. In WebDB

    2006

  40. [48]

    Gubanov, Philip A

    Michael N. Gubanov, Philip A. Bernstein, and Alexander Moshchuk. 2008. Model Management Engine for Data Integration with Reverse -Engineering Support. In ICDE

    2008

  41. [49]

    Zhiheng Huang, Wei Xu, and Kai Yu. 2015. Bidirectional LSTM -CRF Models for Sequence Tagging. arXiv. https://doi.org/10.48550/ARXIV.1508.01991

    2015 doi arXiv

  42. [50]

    Kazi Islam and Michael Gubanov. 2021. Scalable Tabular Metadata Location and Classification in Large-scale Structured Datasets. In DEXA

    2021

  43. [51]

    L. C. Jain and L. R. Medsker. 1999. Recurrent Neural Networks: Design and Appli- cations (1st ed.). CRC Press, Inc., USA

    1999

  44. [52]

    Karen Sparck Jones. 1972. A statistical interpretation of term specificity and its application in retrieval. Journal of Documentation 28 (1972), 11–21

    1972

  45. [53]

    Bhimesh Kandibedala, Anna Pyayt, Chris Caballero, and Michael N. Gubanov. 2023. Scalable Hierarchical Metadata Classification in Heterogeneous Large- scale Datasets. In EDBT DOLAP 2023

    2023

  46. [54]

    Rituparna Khan and Michael Gubanov. 2018. Nested Dolls: Towards Unsupervised Clustering of Web Tables. In IEEE Big Data

    2018

  47. [55]

    Rituparna Khan and Michael Gubanov. 2018. Towards Unsupervised Web Tables Clustering. In IEEE BigData

    2018

  48. [56]

    Rituparna Khan and Michael Gubanov. 2020. Towards Tabular Embeddings, Training the Relational Models. In IEEE Big Data

    2020

  49. [57]

    Rituparna Khan and Michael Gubanov. 2020. WebLens: Towards Interactive Large-Scale Structured Data Profiling. In CIKM

    2020

  50. [58]

    Rituparna Khan and Michael Gubanov. 2020. WebLens: Towards Interactive Web-scale Data Integration, Training the Models. In IEEE Big Data

    2020

  51. [59]

    Anusha Kola, Harshal More, Sean Soderman, and Michael Gubanov. 2017. Gen- erating Unified Famous Objects (UFOs) from the classified object tables. In IEEE Big Data

    2017

  52. [60]

    Oliver Lehmberg, Dominique Ritze, Robert Meusel, and Christian Bizer

  53. [61]

    OpenAI. 2023. GPT-4 Large Language Model. https://openai.com/gpt - 4

    2023

  54. [62]

    Steven Ortiz, Caner Enbatan, Maksim Podkorytov, Dylan Soderman, and Michael Gubanov. 2017. Hybrid.JSON: High -velocity Parallel In -Memory Polystore JSON Ingest. In IEEE Bigdata

    2017

  55. [63]

    Sophie Pavia, Rituparna Khan, Anna Pyayt, and Michael Gubanov. 2022. Simpli- fying Access to Large -scale Structured Datasets by Meta -Profiling with Scalable Training Set Enrichment. In SIGMOD. ACM

    2022

  56. [64]

    Sophie Pavia, Montasir Shams, Rituparna Khan, Anna Pyayt, and Michael N.Gubanov. 2021. Learning Tabular Embeddings at Web Scale. In Big Data. IEEE

    2021

  57. [65]

    Thomas Pellissier Tanon, Gerhard Weikum, and Fabian Suchanek. 2020. YAGO4: A Reason -able Knowledge Base. In ESWC, Andreas Harth, Sabrina Kirrane, Axel-Cyrille Ngonga Ngomo, Heiko Paulheim, Anisa Rula, Anna Lisa Gentile, Peter Haase, and Michael Cochez (Eds.)

    2020

  58. [66]

    Rajib Rana. 2016. Gated Recurrent Unit (GRU) for Emotion Classification from Noisy Speech

    2016

  59. [67]

    Montasir Shams, Sophie Pavia, Rituparna Khan, Anna Pyayt, and Michael N.Gubanov. 2021. Towards Unveiling Dark Web Structured Data. In Big Data. IEEE

    2021

  60. [68]

    Hybrid.media: High Velocity Video Ingestion in an In - Memory Scalable Analytical Polystore

    Mark Simmons, Daniel Armstrong, Dylan Soderman, and Michael Gubanov.2017. Hybrid.media: High Velocity Video Ingestion in an In - Memory Scalable Analytical Polystore. In IEEE Bigdata

    2017

  61. [69]

    Amit Singhal. 2012. Introducing the KG: Things, Not Strings. In Google Blog

    2012

  62. [70]

    Sean Soderman, Anusha Kola, Maksim Podkorytov, Michael Geyer, and Michael Gubanov. 2018. Hybrid.AI: A Learning Search Engine for Large - scale Structured Data. In WWW

    2018

  63. [71]

    Irene Solaiman, Miles Brundage, Jack Clark, Amanda Askell, Ariel Herbert - Voss, Jeff Wu, Alec Radford, Gretchen Krueger, Jong Wook Kim, Sarah Kreps, Miles McCain, Alex Newhouse, Jason Blazakis, Kris McGuffie, and Jasmine Wang. 2019. Release Strategies and the Social Impacts o...

    2019 arXiv

  64. [72]

    Dan Su, Yan Xu, Tiezheng Yu, Farhad Bin Siddique, Elham Barezi, and Pascale Fung. 2020. CAiRE -COVID: A Question Answering and Query - focused Multi - Document Summarization System for COVID -19 Scholarly Information Manage- ment. In Proceedings of the 1st Workshop on NLP for ...

    2020

  65. [73]

    Santiago Villasenor, Tom Nguyen, Anusha Kola, Sean Soderman, and Michael Gubanov. 2017. Scalable spam classifier for web tables. In IEEE Big Data

    2017

  66. [74]

    Weld, Oren Etzioni, and Sebastian Kohlmeier

    Lucy Lu Wang, Kyle Lo, Yoganand Chandrasekhar, Russell Reas, Jiangjiang Yang, Doug Burdick, Darrin Eide, Kathryn Funk, Yannis Katsis, Rodney Kinney, Yunyao Li, Ziyang Liu, William Merrill, Paul Mooney, Dewey Murdick, Devvret Rishi, Jerry Sheehan, Zhihong Shen, Brandon Stilson,...

  67. [75]

    Nasser Zalmout, Chenwei Zhang, Xian Li, Yan Liang, and Xin Luna Dong

  68. [76]

    Bhimesh Kandibedala, Anna Pyayt, Nickolas Piraino, Chris Caballero, and Michael Gubanov. 2023. COVIDKG.ORG-a Web-scale COVID-19 Interactive, Trustworthy Knowledge Graph, Constructed and Interrogated for Bias using Deep-Learning. In EDBT 2023

    2023

  69. [2008]

    In BIRTE

    Simplifying Information Integration: Object -Based Flow -of- Mappings Framework for Integration. In BIRTE

  70. [2016]

    In WWW, Jacqueline Bourdeau, Jim Hendler, Roger Nkambou, Ian Horrocks, and Ben Y

    A Large Public Corpus of Web Tables containing Time and Context Metadata. In WWW, Jacqueline Bourdeau, Jim Hendler, Roger Nkambou, Ian Horrocks, and Ben Y. Zhao (Eds.)

  71. [2020]

    In arXiv, cs.DL 2004.10706

    CORD -19: The COVID -19 Open Research Dataset. In arXiv, cs.DL 2004.10706

    2004 arXiv

  72. [2021]

    In SIGKDD, Feida Zhu, Beng Chin Ooi, and Chunyan Miao (Eds.)

    All You Need to Know to Build a Product Knowledge Graph. In SIGKDD, Feida Zhu, Beng Chin Ooi, and Chunyan Miao (Eds.). ACM

  73. [2022]

    arXiv:2210.11416 [cs.LG]

    Scaling Ins truction-Finetuned Language Models. arXiv:2210.11416 [cs.LG]

    arXiv

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.