Pith. sign in

REVIEW 2 major objections 5 minor 50 references

Foundation Models for Anomaly Detection: Vision and Challenges

T0 review · 2 major / 5 minor · reviewed 2026-08-08 · deepseek-v4-flash

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

Pith's one-line read A taxonomy of three roles—encoder, detector, interpreter—organizes foundation-model anomaly detection.

desk verdict A useful survey whose encoder/detector/interpreter taxonomy organizes the field well, but the 'first comprehensive review' claim is asserted without a systematic search protocol or a serious comparison with prior surveys. read the letter →

arxiv 2502.06911 v2 pith:U3JLJRFF submitted 2025-02-10 cs.LG cs.AI

Jing Ren , Tao Tang , Hong Jia , Ziqi Xu , Haytham Fayek , Xiaodong Li , Suyu Ma , Xiwei Xu
show 1 more author
Feng Xia
This is my paper · ORCID
classification cs.LGcs.AI
keywords foundationmodelanomalydetectionlargelanguagetaxonomyencoderdetectorinterpreterexplainability
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 survey tries to organize the rapidly growing body of work that uses foundation models for anomaly detection. Its central claim is that every current method can be understood by the role the foundation model plays: encoding data into representations, directly detecting anomalies, or interpreting detected anomalies. The authors say this is the first comprehensive review to do so, and they use the three-role scheme to sort dozens of recent systems across images, video, time series, logs, graphs, and finance. The payoff of getting the taxonomy right is practical: practitioners can locate their problem in the scheme, and the field can see where methods are mature and where they are not.

What carries the argument

The load-bearing mechanism is the three-role taxonomy itself, which assigns a foundation model to one of three functional slots in the anomaly detection pipeline: encoder, detector, or interpreter. Because each slot has a distinct interface to the rest of the system—an embedding vector, a prompted prediction, or an explanation text—the taxonomy can organize methods that otherwise look unrelated. The survey also supplies sub-branches for each role, so the scheme is fine-grained enough to separate, for example, embeddings produced solely by an FM from embeddings fused with a graph neural network.

What would settle it

A literature search covering 2023–2025 that finds an earlier survey already using the encoder/detector/interpreter split would refute the 'first comprehensive review' claim. Likewise, finding a published FM-based anomaly detection system whose function is none of encoding, detecting, or interpreting would show the taxonomy is not exhaustive.

Watch

Extended reading notes

Core claim

The paper claims that foundation-model-based anomaly detection has reached a stage where a structured map is necessary and possible. It proposes that FMs take exactly three roles in the pipeline: as encoders, where they produce embeddings that downstream classifiers consume; as detectors, where they directly classify or localize anomalies from serialized or encoded input; and as interpreters, where they explain or verify anomaly reports. Within each role the survey distinguishes two subfamilies—FM-based versus hybrid embedding, serialization-based versus encoding-based detection, and detection-based versus verification-based explanation. The authors claim this taxonomy is the first tailored specifically to how FMs are applied to anomaly detection, and they argue that it brings the field's open challenges into view: efficiency, bias, explainability, and multimodality.

Load-bearing premise

The taxonomy's value depends on the literature search being complete, and on no earlier review already dividing FM-based anomaly detection into the same three roles.

Editorial extensions

If this is right

  • Any proposed method can be placed in the scheme by asking what the FM contributes—representations, decisions, or explanations—giving newcomers and reviewers a shared vocabulary.
  • The inventory shows that zero-shot and few-shot anomaly detection are already feasible across images, video, time series, logs, and finance without per-dataset retraining, with explanations delivered in natural language.
  • The open-challenge list points to concrete bottlenecks: model efficiency for real-time use, inherited bias, prompt-dependent interpretability, and the lack of genuinely multimodal anomaly detectors.
  • Because several surveyed systems use more than one FM or one FM in more than one role, the taxonomy implies that composite designs are common and that forcing every method into a single category would misrepresent the field.

Reading between the lines

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

  • A reader could push the taxonomy further by treating the three roles as composable building blocks and asking which role assignment is optimal for a given data modality; the survey itself does not rank role choices.
  • The taxonomy suggests a benchmarking program: fix the FM and vary its role on the same datasets to isolate where the value actually comes from—representations, decisions, or explanations.
  • If the taxonomy is right, progress in the field will look less like new architectures and more like better interfaces around existing FMs: prompt serialization, embedding fusion, and explanation verification.
  • The 'first comprehensive review' claim could be tested by checking whether later work adopts the encoder/detector/interpreter vocabulary; adoption would be evidence that the map is useful.
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

2 major / 5 minor

Summary. This survey reviews recent work on using foundation models (FMs) for anomaly detection (AD). It proposes a taxonomy that classifies FM usage into three roles—encoder, detector, and interpreter—and surveys roughly twenty recent methods across image, video, time-series, log, tabular, and graph data. The paper also discusses open challenges (efficiency, bias, explainability, multimodality) and outlines future research directions. The central claims are that this is the first comprehensive review of FM-based AD and that the three-role taxonomy is novel and organizes the field.

Significance. If the taxonomy is adopted, it provides a clear mental model for researchers entering the area and helps position new work. The survey covers a broad selection of recent methods and includes a useful summary table with model, FM type, fine-tuning, prompting, data domain, and code availability. The challenges section is sensible and well-grounded. The main weakness is that the 'first comprehensive' and 'novel taxonomy' claims are not substantiated by a transparent search protocol or a detailed comparison with existing surveys (Su et al. [39], Xu et al. [40]); the paper's value is still real, but the presentation overstates its novelty.

major comments (2)
  1. [Abstract and Introduction] The paper asserts that it presents 'the first comprehensive review' of FM-based anomaly detection and a 'novel taxonomy,' but no systematic literature search protocol is reported (e.g., databases, query strings, inclusion/exclusion criteria, or a PRISMA-style flow), and the differentiation from prior surveys Su et al. [39] and Xu et al. [40] is limited to a few sentences without a structured comparison table. Since comprehensiveness and non-redundancy are the central value claims of a survey, please add a transparent search protocol and a comparison with existing surveys, or temper the claims to be explicitly scoped (e.g., 'a review focusing on the encoder/detector/interpreter roles').
  2. [Proposed Taxonomy and Table 1] The taxonomy is presented as classifying current FM-based anomaly detection methods into three categories, but the paper admits several boundary cases: LogiCode [46], AnomalyRuler [41], and Audit-LLM [38] are placed in an 'Others' category, and AnomalyLLM [21] is described as a special case where the FM is not the encoder. This is honest but weakens the exhaustiveness implied by the 'novel taxonomy' claim. The authors should either extend the taxonomy with an explicit fourth role (e.g., 'assistant' or 'code generator') to accommodate these works, or clearly state that the three-role taxonomy covers the majority but not all works, and adjust the wording accordingly.
minor comments (5)
  1. [Introduction] The phrase 'notable examples including include GPT-4' contains a redundant 'including include' and should be corrected.
  2. [FM as Encoder, Eq. (3)] Equation (3) and the surrounding text contain notation errors: 'ϕGN N(·)' and 'ϕF M(cdot)' should be written with proper subscripts, e.g., 'ϕ_GNN(·)' and 'ϕ_FM(·)'.
  3. [Table 1] The 'FM type' column is inconsistent: rows [3] and [24] list 'Transformer' as a specific FM, although 'Transformer' is an architecture rather than a specific open-sourced model; row [16] lists 'ChatGPT' as an LVLM, but ChatGPT is typically an LLM. Please verify and correct these entries.
  4. [FM as Detector, Discussion] The statement that FMs 'cannot be directly applied as an anomaly detector without prompt learning and fine-tuning strategy' is too strong given the paper's own descriptions of LAVAD [42] and SIGLLM [1] as training-free or zero-shot methods; please qualify the claim (e.g., 'often require' or 'in many cases').
  5. [FM as Interpreter] The word 'human-understanable' in the Discussions paragraph is a typo and should be 'human-understandable'.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the survey's taxonomy is a classification scheme, not a derived result, and self-citations are background only.

full rationale

This paper is a literature survey. Its central contributions — the encoder/detector/interpreter taxonomy and the 'first comprehensive review' claim — are organizational and bibliographic statements, not predictions or fitted quantities. The taxonomy is introduced by definition in the 'Proposed Taxonomy' section, and the subsequent sections classify existing works according to that definition; no equation is derived from data, and no evaluation result is generated from a fitted parameter. The formal-looking equations (1)-(10) are generic templates describing how FM-based methods in each category process inputs; they are illustrative, not derivations. The authors' own prior works [22, 25, 33, 34] appear as background citations for anomaly detection and self-supervised learning; none of them supplies the taxonomy, and the taxonomy's novelty is asserted rather than derived from those papers. The absence of a systematic search protocol and a structured comparison with prior surveys [39, 40] is a rigor and verifiability concern about the 'first comprehensive' wording, but it is not a circular step under the stated criteria: the claim is not made true by the paper's own definitions or by self-citation. No specific circular reduction can be exhibited, so the circularity score is 0.

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

This survey contains no derivations, numerical fits, or new models, so there are no free parameters or invented entities. The central claims rest on two domain assumptions: the taxonomy is adequate (even though the text admits cross-cutting cases) and the literature selection is comprehensive (no search protocol is given).

assumptions (3)
  • domain assumption The three-role taxonomy (encoder, detector, interpreter) is a meaningful and sufficiently non-overlapping classification of foundation-model usage in anomaly detection.
    The entire survey is organized around this taxonomy; the text itself acknowledges that several systems (LogiCode, AnomalyRuler, Audit-LLM) do not fit exclusively into one category, which weakens the assumption.
  • domain assumption The surveyed set of approximately thirty papers is representative of the state of the art in FM-based anomaly detection.
    The 'comprehensive' claim relies on the selection being representative; no systematic search protocol is provided, so this assumption is unverified.
  • domain assumption The cited papers' reported performance and roles are accurately represented.
    A survey's value depends on accurate representation of the primary literature; the authors do not release any verification artifacts.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Foundation Models for Anomaly Detection: Vision and Challenges." pith.science (2026). https://pith.science/paper/U3JLJRFF

@misc{pith2026250206911,
  author       = {Pith},
  title        = {Pith review of: Foundation Models for Anomaly Detection: Vision and Challenges},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/U3JLJRFF}},
  note         = {Machine review of arXiv:2502.06911}
}
read the original abstract

As data continues to grow in volume and complexity across domains such as finance, manufacturing, and healthcare, effective anomaly detection is essential for identifying irregular patterns that may signal critical issues. Recently, foundation models (FMs) have emerged as a powerful tool for advancing anomaly detection. They have demonstrated unprecedented capabilities in enhancing anomaly identification, generating detailed data descriptions, and providing visual explanations. This survey presents the first comprehensive review of recent advancements in FM-based anomaly detection. We propose a novel taxonomy that classifies FMs into three categories based on their roles in anomaly detection tasks, i.e., as encoders, detectors, or interpreters. We provide a systematic analysis of state-of-the-art methods and discuss key challenges in leveraging FMs for improved anomaly detection. We also outline future research directions in this rapidly evolving field.

Figures

Figures reproduced from arXiv: 2502.06911 by the authors.

Figure 1
Figure 1. An illustration of applying foundation models to the three key stages of anomaly detection, playing roles as encoder, detector, and interpreter. Three cases are illustrated with image, tabular, and time series data, respectively. (a) FM-based Embedding (b) Hybrid Embedding [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. FM as encoder. WinCLIP+ is introduced to aggregate comple￾mentary information from WinCLIP and visual signals from normal samples. However, the per￾formance of WinCLIP+ is heavily dependent on extensive engineering on hundreds of manually defined prompts. To solve this problem, Anoma￾lyCLIP [48] uses an object-agnostic prompt tem￾plate to model the semantics of general abnor￾mality and normality, thus improving the … view at source ↗
Figure 3
Figure 3. FM as detector because FMs are often prompt-based, while the source data in most cases present different modal￾ities, such as time-series, image, and video. In this section, we classify the models broadly into serialization-based and encoding-based detection, depending on how data are preprocessed before inputted to FMs. Serialization-based Detection Most of the existing research attempts to employ the question-answ… view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: FM as interpreter on large corpora of normal data without spe￾cific training stages [35]. As for encoding-based detection, training an additional DL module and inserting it into FMs for joint training is challeng￾ing due to the problem of vanishing gradients in the ear…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

50 extracted references · 35 canonical work pages

  1. [39]

    Large language models for forecasting and anomaly detection: A systematic literature review

    Jing Su, Chufeng Jiang, Xin Jin, Yuxin Qiao, Tingsong Xiao, Hongda Ma, Rong Wei, Zhi Jing, Jiajun Xu, and Junhong Lin. Large language models for forecasting and anomaly detection: A systematic literature review. arXiv preprint arXiv:2402.10350, 2024

    arXiv 2024

  2. [40]

    Large language models for anomaly and out-of-distribution detection: A survey

    Ruiyao Xu and Kaize Ding. Large language models for anomaly and out-of-distribution detection: A survey. arXiv preprint arXiv:2409.01980, 2024

    arXiv 2024

  3. [46]

    Logicode: an llm-driven framework for logical anomaly detection

    Yiheng Zhang, Yunkang Cao, Xiaohao Xu, and Weim- ing Shen. Logicode: an llm-driven framework for logical anomaly detection. IEEE Transactions on Automation Science and Engineering, 2024

    work page 2024

  4. [41]

    Follow the rules: reasoning for video anomaly detection with large language mod- els

    Yuchen Yang, Kwonjoon Lee, Behzad Dariush, Yinzhi Cao, and Shao-Yuan Lo. Follow the rules: reasoning for video anomaly detection with large language mod- els. In European Conference on Computer Vision, pages 304–322. Springer, 2025

    work page 2025

  5. [38]

    Audit-llm: Multi-agent collaboration for log-based insider threat detection

    Chengyu Song, Linru Ma, Jianming Zheng, Jinzhi Liao, Hongyu Kuang, and Lin Yang. Audit-llm: Multi-agent collaboration for log-based insider threat detection. arXiv preprint arXiv:2408.08902, 2024

    arXiv 2024

  6. [21]

    Large language model guided knowl- edge distillation for time series anomaly detection

    Chen Liu, Shibo He, Qihang Zhou, Shizhong Li, and Wenchao Meng. Large language model guided knowl- edge distillation for time series anomaly detection. In Proceedings of the Thirty-Third International Joint Conference on Artificial Intelligence, pages 2162–2170, 2024

    work page 2024

  7. [1]

    Large language mod- els can be zero-shot anomaly detectors for time series? arXiv preprint arXiv:2405.14755, 2024

    Sarah Alnegheimish, Linh Nguyen, Laure Berti-Equille, and Kalyan Veeramachaneni. Large language mod- els can be zero-shot anomaly detectors for time series? arXiv preprint arXiv:2405.14755, 2024

    arXiv 2024

  8. [2]

    Gpt-lad: Leveraging large multimodal models for logical anomaly detection

    Yoojin An and Dongyeon Kang. Gpt-lad: Leveraging large multimodal models for logical anomaly detection. 10 AI MAGAZINE In ICASSP 2025-2025 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , pages 1–5. IEEE, 2025

    work page 2025

Show all 50 references
  1. [3]

    Advancing anomaly detection: Non-semantic financial data encoding with llms

    Alexander Bakumenko, Kate ˇrina Hlav´aˇckov´a-Schindler, Claudia Plant, and Nina C Hubig. Advancing anomaly detection: Non-semantic financial data encoding with llms. arXiv preprint arXiv:2406.03614, 2024

    2024

  2. [4]

    On the opportunities and risks of foundation models

    Rishi Bommasani, Drew A Hudson, Ehsan Adeli, Russ Altman, Simran Arora, Sydney von Arx, Michael S Bernstein, Jeannette Bohg, Antoine Bosselut, Emma Brunskill, et al. On the opportunities and risks of foundation models. arXiv preprint arXiv:2108.07258 , 2021

    2021 arXiv

  3. [5]

    Domain-controlled prompt learn- ing

    Qinglong Cao, Zhengqin Xu, Yuntian Chen, Chao Ma, and Xiaokang Yang. Domain-controlled prompt learn- ing. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 38, pages 936–944, 2024

    2024

  4. [6]

    Spiced: Syntactical bug and trojan pattern identification in a/ms circuits using llm-enhanced detection

    Jayeeta Chaudhuri, Dhruv Thapar, Arjun Chaudhuri, Farshad Firouzi, and Krishnendu Chakrabarty. Spiced: Syntactical bug and trojan pattern identification in a/ms circuits using llm-enhanced detection. In 2024 IEEE Physical Assurance and Inspection of Electronics (PAINE), pages ...

    2024

  5. [7]

    Can llms serve as time series anomaly detectors? arXiv preprint arXiv:2408.03475, 2024

    Manqing Dong, Hao Huang, and Longbing Cao. Can llms serve as time series anomaly detectors? arXiv preprint arXiv:2408.03475, 2024

    2024 arXiv

  6. [8]

    Llm-botguard: A novel framework for detecting llm-driven bots with mixture of experts and graph neural networks

    Jinglong Duan, Weihua Li, Quan Bai, Minh Nguyen, Xiaodan Wang, and Jianhua Jiang. Llm-botguard: A novel framework for detecting llm-driven bots with mixture of experts and graph neural networks. IEEE Transactions on Computational Social Systems, 2025

    2025

  7. [9]

    Semantic anomaly detection with large language mod- els

    Amine Elhafsi, Rohan Sinha, Christopher Agia, Edward Schmerling, Issa AD Nesnas, and Marco Pavone. Semantic anomaly detection with large language mod- els. Autonomous Robots, 47(8):1035–1055, 2023

    2023

  8. [10]

    Anomalygpt: Detecting industrial anomalies using large vision-language mod- els

    Zhaopeng Gu, Bingke Zhu, Guibo Zhu, Yingying Chen, Ming Tang, and Jinqiao Wang. Anomalygpt: Detecting industrial anomalies using large vision-language mod- els. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 38, pages 1932–1940, 2024

    1932

  9. [11]

    Dabl: Detecting semantic anomalies in business processes using large language models

    Wei Guan, Jian Cao, Jianqi Gao, Haiyan Zhao, and Shiyou Qian. Dabl: Detecting semantic anomalies in business processes using large language models. In Proceedings of the AAAI Conference on Artificial Intel- ligence, volume 39, pages 11735–11744, 2025

    2025

  10. [12]

    Anomaly detection on unstable logs with gpt models

    Fatemeh Hadadi, Qinghua Xu, Domenico Bianculli, and Lionel Briand. Anomaly detection on unstable logs with gpt models. arXiv preprint arXiv:2406.07467, 2024

    2024

  11. [13]

    Adapting visual-language models for generalizable anomaly detection in medical images

    Chaoqin Huang, Aofan Jiang, Jinghao Feng, Ya Zhang, Xinchao Wang, and Yanfeng Wang. Adapting visual-language models for generalizable anomaly detection in medical images. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 11375–11385, 2024

    2024

  12. [14]

    Win- clip: Zero-/few-shot anomaly classification and segmen- tation

    Jongheon Jeong, Yang Zou, Taewan Kim, Dongqing Zhang, Avinash Ravichandran, and Onkar Dabeer. Win- clip: Zero-/few-shot anomaly classification and segmen- tation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 19606–19616, 2023

    2023

  13. [15]

    Logicad: Explainable anomaly detection via vlm-based text feature extraction

    Er Jin, Qihui Feng, Yongli Mou, Gerhard Lakemeyer, Stefan Decker, Oliver Simons, and Johannes Stegmaier. Logicad: Explainable anomaly detection via vlm-based text feature extraction. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 39, pages 4129–4137, 2025

    2025

  14. [16]

    Unsupervised video anomaly detection based on similarity with predefined text descriptions

    Jaehyun Kim, Seongwook Yoon, Taehyeon Choi, and Sanghoon Sull. Unsupervised video anomaly detection based on similarity with predefined text descriptions. Sensors, 23(14):6256, 2023

    2023

  15. [17]

    The power of scale for parameter-efficient prompt tuning

    Brian Lester, Rami Al-Rfou, and Noah Constant. The power of scale for parameter-efficient prompt tuning. In Proceedings of the 2021 Conference on Empirical Meth- ods in Natural Language Processing, pages 3045–3059, 2021

    2021

  16. [18]

    Anomaly detection of tabular data using llms

    Aodong Li, Yunhan Zhao, Chen Qiu, Marius Kloft, Padhraic Smyth, Maja Rudolph, and Stephan Mandt. Anomaly detection of tabular data using llms. arXiv preprint arXiv:2406.16308, 2024

    2024 arXiv

  17. [19]

    Myriad: Large multimodal model by applying vision experts for industrial anomaly detec- tion

    Yuanze Li, Haolin Wang, Shihao Yuan, Ming Liu, Debin Zhao, Yiwen Guo, Chen Xu, Guangming Shi, and Wangmeng Zuo. Myriad: Large multimodal model by applying vision experts for industrial anomaly detec- tion. arXiv preprint arXiv:2310.19070, 2023

    arXiv

  18. [20]

    A survey of graph meets large language model: Progress and future direc- tions

    Yuhan Li, Zhixun Li, Peisong Wang, Jia Li, Xiangguo Sun, Hong Cheng, and Jeffrey Xu Yu. A survey of graph meets large language model: Progress and future direc- tions. In International Joint Conference on Artificial Intelligence (IJCAI), 2024

    2024

  19. [22]

    Deep graph learning for anomalous citation detec- tion

    Jiaying Liu, Feng Xia, Xu Feng, Jing Ren, and Huan Liu. Deep graph learning for anomalous citation detec- tion. IEEE Transactions on Neural Networks and Learning Systems, 33(6):2543–2557, 2022

    2022

  20. [23]

    Large language models can deliver accurate and interpretable time series anomaly detection

    Jun Liu, Chaoyun Zhang, Jiaxu Qian, Minghua Ma, Si Qin, Chetan Bansal, Qingwei Lin, Saravan Rajmo- han, and Dongmei Zhang. Large language models can deliver accurate and interpretable time series anomaly detection. arXiv preprint arXiv:2405.15370, 2024

    2024 arXiv

  21. [24]

    Anoma- lyllm: Few-shot anomaly edge detection for dynamic graphs using large language models

    Shuo Liu, Di Yao, Lanting Fang, Zhetao Li, Wenbin Li, Kaiyu Feng, XiaoWen Ji, and Jingping Bi. Anoma- lyllm: Few-shot anomaly edge detection for dynamic graphs using large language models. arXiv preprint arXiv:2405.07626, 2024

    2024 arXiv

  22. [25]

    Graph self-supervised learn- ing: A survey

    Yixin Liu, Ming Jin, Shirui Pan, Chuan Zhou, Yu Zheng, Feng Xia, and S Yu Philip. Graph self-supervised learn- ing: A survey. IEEE transactions on knowledge and data engineering, 35(6):5879–5900, 2022

    2022

  23. [26]

    Unified-io: A uni- fied model for vision, language, and multi-modal tasks

    Jiasen Lu, Christopher Clark, Rowan Zellers, Roozbeh Mottaghi, and Aniruddha Kembhavi. Unified-io: A uni- fied model for vision, language, and multi-modal tasks. In The Eleventh International Conference on Learning Representations, 2022

    2022

  24. [27]

    Video anomaly detection and explanation via large language models

    Hui Lv and Qianru Sun. Video anomaly detection and explanation via large language models. arXiv preprint arXiv:2401.05702, 2024. AI MAGAZINE 11

    2024 arXiv

  25. [28]

    Foundation models for generalist medical artificial intelligence

    Michael Moor, Oishi Banerjee, Zahra Shakeri Hos- sein Abad, Harlan M Krumholz, Jure Leskovec, Eric J Topol, and Pranav Rajpurkar. Foundation models for generalist medical artificial intelligence. Nature, 616(7956):259–265, 2023

    2023

  26. [29]

    A decade survey of transfer learning (2010–2020)

    Shuteng Niu, Yongxin Liu, Jian Wang, and Houb- ing Song. A decade survey of transfer learning (2010–2020). IEEE Transactions on Artificial Intelli- gence, 1(2):151–166, 2020

    2010

  27. [30]

    Enhancing anomaly detection in finan- cial markets with an llm-based multi-agent framework

    Taejin Park. Enhancing anomaly detection in finan- cial markets with an llm-based multi-agent framework. arXiv preprint arXiv:2403.19735, 2024

    2024 arXiv

  28. [31]

    Learning transferable visual models from natural lan- guage supervision

    Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sas- try, Amanda Askell, Pamela Mishkin, Jack Clark, et al. Learning transferable visual models from natural lan- guage supervision. In International conference on machine learning...

    2021

  29. [32]

    Scott E. Reed, Konrad Zolna, Emilio Parisotto, Ser- gio G ´omez Colmenarejo, Alexander Novikov, Gabriel Barth-Maron, Mai Gimenez, Yury Sulsky, Jackie Kay, Jost Tobias Springenberg, Tom Eccles, Jake Bruce, Ali Razavi, Ashley Edwards, Nicolas Heess, Yutian Chen, Raia Hadsell, Or...

    2022

  30. [33]

    Graph learning for anomaly ana- lytics: Algorithms, applications, and challenges

    Jing Ren, Feng Xia, Ivan Lee, Azadeh Noori Hoshyar, and Charu Aggarwal. Graph learning for anomaly ana- lytics: Algorithms, applications, and challenges. ACM Transactions on Intelligent Systems and Technology , 14(2), February 2023

    2023

  31. [34]

    Deep video anomaly detection: Opportunities and challenges

    Jing Ren, Feng Xia, Yemeng Liu, and Ivan Lee. Deep video anomaly detection: Opportunities and challenges. In International Conference on Data Mining Workshops (ICDMW), pages 959–966, 2021

    2021

  32. [35]

    Aad-llm: Adaptive anomaly detec- tion using large language models

    Alicia Russell-Gilbert, Alexander Sommers, Andrew Thompson, Logan Cummins, Sudip Mittal, Shahram Rahimi, Maria Seale, Joseph Jaboure, Thomas Arnold, and Joshua Church. Aad-llm: Adaptive anomaly detec- tion using large language models. In 2024 IEEE International Conference on B...

    2024

  33. [36]

    Face it yourselves: An llm-based two-stage strategy to localize configuration errors via logs

    Shiwen Shan, Yintong Huo, Yuxin Su, Yichen Li, Dan Li, and Zibin Zheng. Face it yourselves: An llm-based two-stage strategy to localize configuration errors via logs. In Proceedings of the 33rd ACM SIGSOFT Inter- national Symposium on Software Testing and Analysis , pages 13–25, 2024

    2024

  34. [37]

    Real-time anomaly detection and reactive plan- ning with large language models

    Rohan Sinha, Amine Elhafsi, Christopher Agia, Matthew Foutter, Edward Schmerling, and Marco Pavone. Real-time anomaly detection and reactive plan- ning with large language models. In Proceedings of Robotics: Science and Systems, 2024

    2024

  35. [42]

    Harnessing large lan- guage models for training-free video anomaly detec- tion

    Luca Zanella, Willi Menapace, Massimiliano Mancini, Yiming Wang, and Elisa Ricci. Harnessing large lan- guage models for training-free video anomaly detec- tion. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 18527–18536, 2024

    2024

  36. [43]

    Holmes-vad: Towards unbiased and explainable video anomaly detection via multi-modal llm

    Huaxin Zhang, Xiaohao Xu, Xiang Wang, Jialong Zuo, Chuchu Han, Xiaonan Huang, Changxin Gao, Yue- huan Wang, and Nong Sang. Holmes-vad: Towards unbiased and explainable video anomaly detection via multi-modal llm. arXiv preprint arXiv:2406.12235 , 2024

    2024 arXiv

  37. [44]

    Transfer adaptation learn- ing: A decade survey

    Lei Zhang and Xinbo Gao. Transfer adaptation learn- ing: A decade survey. IEEE Transactions on Neural Networks and Learning Systems, 2022

    2022

  38. [45]

    Scalalog: Scalable log-based failure diagnosis using llm

    Lingzhe Zhang, Tong Jia, Mengxi Jia, Yifan Wu, Hongyi Liu, and Ying Li. Scalalog: Scalable log-based failure diagnosis using llm. In ICASSP 2025-2025 IEEE International Conference on Acoustics, Speech and Sig- nal Processing (ICASSP), pages 1–5. IEEE, 2025

    2025

  39. [47]

    A comprehensive survey on pre- trained foundation models: A history from bert to chat- gpt

    Ce Zhou, Qian Li, Chen Li, Jun Yu, Yixin Liu, Guangjing Wang, Kai Zhang, Cheng Ji, Qiben Yan, Lifang He, et al. A comprehensive survey on pre- trained foundation models: A history from bert to chat- gpt. International Journal of Machine Learning and Cybernetics, pages 1–65, 2024

    2024

  40. [48]

    Anomalyclip: Object-agnostic prompt learning for zero-shot anomaly detection

    Qihang Zhou, Guansong Pang, Yu Tian, Shibo He, and Jiming Chen. Anomalyclip: Object-agnostic prompt learning for zero-shot anomaly detection. InThe Twelfth International Conference on Learning Representations, 2024

    2024

  41. [49]

    Do llms understand visual anoma- lies? uncovering llm’s capabilities in zero-shot anomaly detection

    Jiaqi Zhu, Shaofeng Cai, Fang Deng, Beng Chin Ooi, and Junran Wu. Do llms understand visual anoma- lies? uncovering llm’s capabilities in zero-shot anomaly detection. In Proceedings of the 32nd ACM Interna- tional Conference on Multimedia, pages 48–57, 2024

    2024

  42. [50]

    Toward general- ist anomaly detection via in-context residual learning with few-shot sample prompts

    Jiawen Zhu and Guansong Pang. Toward general- ist anomaly detection via in-context residual learning with few-shot sample prompts. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 17826–17836, 2024

    2024

Pith tools

Reviewed August 8, 2026 · model on record in the stance chip above.