Pith. sign in

REVIEW 4 major objections 4 minor 170 references

A layered architecture for log analysis in complex IT systems

T0 review · 4 major / 4 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read A three-layer log analysis architecture claims anomaly detection F1 scores of 0.98-1.0 and root cause candidates in the top 10 for 90-98% of failures.

desk verdict A coherent thesis-style integration of the author's earlier log-analysis methods; the packaging is genuinely new, but the headline F1/RCA numbers rest on an unvalidated PU-learning premise and on an evaluation chapter I cannot actually check in the provided text. read the letter →

arxiv 2509.08698 v1 pith:BIZLEHHY submitted 2025-08-29 cs.LG cs.AI

Thorsten Wittkopp This is my paper · ORCID
classification cs.LGcs.AI
keywords loganalysisanomalydetectionrootcausePUlearningweaksupervisiontransformerencoderlabelingAIOps
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 proposes a three-layer architecture for log analysis that lets DevOps teams go from raw log files to root cause candidates without manually labeling data. The first layer autonomously labels each log line as normal or anomalous, using only rough failure timestamps from monitoring systems; the second layer detects anomalies with one method that can be trained unsupervised, weakly supervised, or supervised; the third layer ranks the detected anomalies to find the minimal set of log lines describing the failure. The central claim is that the anomaly detection layer achieves F1 scores between 0.98 and 1.0 on public and industry datasets across all three training scenarios, and the root cause analysis layer places 90-98% of root cause log lines within its top 10 candidates. If these numbers hold, the architecture would remove the main bottleneck in AIOps—expensive expert labeling and per-system threshold tuning—while giving operators a short, actionable list of log lines to inspect.

What carries the argument

The load-bearing mechanism is PU learning over failure time windows, combined with a transformer encoder whose anomaly score is the vector norm of the $[CLS]$ token. Positive class $P$ contains all log lines outside the monitoring-derived failure windows; unknown class $U$ contains all lines inside them. The objective is $\frac{1}{m}\sum_i ((1-\tilde y_i)\|z_i\|^2 + \tilde y_i (|P|/(|P|+|U|))^2/\|z_i\|)$, so normal lines are pulled toward the origin and unknown lines are pushed away, with the push strength scaled by the ratio of class sizes. Iterative training smooths scores with $\tanh(\max(0,\|z_i\|-m))$ to turn them into pseudo-labels for the next model, and the inference stage builds a decision boundary by augmenting log lines and observing how scores shift. This same machinery is reused for labeling, for anomaly detection under all three training paradigms, and—after a clustering-based balancing step—for root cause ranking.

What would settle it

On one of the paper's public datasets (for example BGL), compare the labeling-layer F1 when the failure time windows are exactly the ones used in the evaluation versus windows shifted by a few seconds or widened by a factor of two. If the near-unity F1 scores degrade sharply under realistic monitoring jitter, the claim that the method needs only rough failure estimates is refuted.

Watch

Extended reading notes

Core claim

The thesis's core claim is that the entire log analysis workflow can be built on one repeated pattern: split the log lines by estimated failure time windows, treat everything outside the windows as the positive class $P$ (presumed normal) and everything inside as the unknown class $U$ (presumed anomalous but actually mixed), and train a transformer encoder to assign each line an anomaly score equal to the norm of its $[CLS]$ embedding. The loss pushes $P$ scores toward zero and pushes $U$ scores away from zero, with a class-ratio weight that keeps the model stable when the unknown class is large. An iterative loop converts the previous model's scores into smoothed pseudo-labels via $\tanh(\max(0,\|z_i\|-m))$ and retrains, which the paper argues removes the bias of the initial inaccurate window labels. The paper also derives a decision boundary from data augmentation rather than from labeled validation data, and reports F1 scores of 0.98-1.0 for unsupervised, weakly supervised, and supervised training on three public datasets plus industry data; for root cause analysis, it reports that 90-98% of root cause log lines appear in the top 10 ranked candidates after balancing the training data.

Load-bearing premise

The load-bearing premise is that the log lines outside the estimated failure time windows are all truly normal, so the positive class $P$ contains no hidden anomalies; if monitoring timestamps are imprecise or a failure leaks log lines outside the window, every layer inherits that error.

Editorial extensions

If this is right

  • DevOps teams could obtain per-line anomaly labels automatically from monitoring timestamps, removing the manual labeling bottleneck.
  • One anomaly detector would cover all three training regimes, so a team can start unsupervised and later add labels without replacing the model.
  • Operators would receive a shortlist of at most 10 candidate lines for most failures, reducing manual log reading during incident response.
  • The anomaly taxonomy would let teams inspect which anomaly types dominate their logs and choose detection methods accordingly.
  • The autonomous labeling layer could bootstrap supervised training on systems that currently have no labels at all.

Reading between the lines

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

  • If the reported scores depend strongly on the purity of the positive class, then the architecture's real-world ceiling is set by monitoring accuracy, not by the model; a quick test is to inject a small fraction of anomalies into $P$ and measure F1 decay.
  • The same PU-plus-transformer recipe may transfer to other weakly labeled event sequences (traces, metrics, CI/CD logs), where failure windows are even noisier than in the datasets used here.
  • The 'top 10 candidates' framing suggests root cause analysis is better posed as ranking than classification; the paper's balancing step implies that rare root cause lines, not common ones, are what limit recall.
  • The architecture's modularity implies each layer can be used alone; the labeling layer, in particular, could serve as a data-cleaning step for any downstream log model, not just the one proposed.
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

4 major / 4 minor

Summary. The dissertation-style manuscript introduces a three-layer log-analysis architecture. Layer 1 (Log Investigation) contributes an autonomous labeling method based on PU learning with failure-time windows and an anomaly taxonomy dividing anomalies into template, attribute, and contextual types. Layer 2 (Anomaly Detection) contributes a transformer-encoder method that can be trained unsupervised, weakly supervised, or supervised, together with a decision-boundary procedure for inference. Layer 3 (Root Cause Analysis) aims to identify a small, ranked set of log lines describing the origin and propagation of a failure. The headline empirical claims are F1-scores of 0.98-1.0 for anomaly detection across all three training modes and 90-98% of root-cause log lines appearing within the top-10 candidates.

Significance. If the results hold, the architecture would be practically significant: it would allow per-line anomaly labeling without manual effort, a single detection method spanning multiple supervision regimes, and a ranked root-cause shortlist for DevOps teams. The manuscript's strengths include the explicit formalization of the PU-learning setup in Section 2.4.2, the detailed preprocessing pipeline in Section 6.3, the careful discussion of decision boundaries as a first-class inference problem, and the fact that the components build on six peer-reviewed publications. The significance is currently conditional, however: the provided arXiv text is truncated after Section 6.4.2, so the evaluation that supports the abstract's numbers is not available for review, and the central PU objective in Eq. (5.5) has a mathematical issue that needs attention.

major comments (4)
  1. [§2.4.2, §5.1.2, §7.2] The architecture's load-bearing premise is Assumption 1 in Section 2.4.2: the positive class P (log lines outside the failure-time windows) consists of true normal samples. This is stated as an assumption but never validated. If monitoring fault times are imprecise, or if failures emit log lines outside the chosen ±δ windows, then P is silently polluted with anomalies, and every downstream score, label, and RCA ranking inherits that bias. The claimed F1 0.98-1.0 and RCA recall@10 numbers are therefore conditional on an unvalidated property. Please provide a direct measurement of P purity (the fraction of ground-truth anomaly lines that fall into P) as a function of δ, and a sensitivity analysis of AD F1 and RCA recall@10 to δ on the public datasets with known ground truth; reporting F1 at different δ values (e.g., Table 8.7) does not by itself establish that P is anomaly-free.
  2. [§5.1.5, Eq. (5.5)] The unknown-class term in the objective function is b(zi) = q^2 / ||zi||, making the total loss in Eq. (5.5) monotonically decreasing as ||zi|| grows. The loss is therefore unbounded below: a model can reduce it without limit by inflating output norms, and no regularization, weight decay, or norm constraint is described. This makes the labeling method and the weakly supervised anomaly-detection variant ill-posed, because the anomaly score has no finite optimum. The objective should be replaced or supplemented with a bounded/margin-based formulation, and the boundedness of the resulting loss should be explicitly established.
  3. [§4.1.3, Definition 6] The formal definition of Root Cause Analysis uses Lroot = arg max over subsets S of Σ_{li∈S} Impact(li), but the function Impact(li) is never defined in Chapter 4 or in any of the available text. Since the RCA recall@10 claim is a headline result, the objective function must be specified precisely: what is Impact(li), how is it computed from the anomaly scores and service information, and how is the subset maximization carried out? Without these details the RCA method in Chapter 7 and the reported 90-98% recall@10 cannot be reproduced.
  4. [§8 (Evaluation)] The provided manuscript is truncated after Section 6.4.2, so Tables 8.6-8.11 and Figures 8.2-8.10, which are referenced as supporting the abstract's headline numbers, are not available for review. As submitted, the central empirical claims are unverifiable. A complete evaluation must include: dataset statistics, hyperparameter settings (δ, τ, S, R, α, a, b), baseline configurations, standard deviations across repeated runs, statistical significance tests, and a clear statement of which public splits are used. Without this material, the F1 range 0.98-1.0 and the RCA top-10 recall cannot be assessed.
minor comments (4)
  1. [§5.1.2] The sentence 'Whereas P contains presumably normal log lines, the label normal1 is assigned' contains a typo; it should read 'the label normal, i.e., 0, is assigned' to be consistent with the formal definition in Section 2.4.2.
  2. [§5.2.3, Eq. (5.7)] The set notation in Eq. (5.7) appears malformed: the closing bracket is ']' instead of '}', and the example c10 = {l8, l9, l11} refers to log lines whereas the formula defines contexts over template ids txj. Please align the notation.
  3. [§6.4.1, Definitions 7-11] The terms 'Positive Training Data' and 'Unknown Training Data' conflict with the earlier PU-learning notation, where P denotes the normal class and U denotes the failure-window class. In Definition 9, 'positive' means 'single-class' rather than 'anomalous', which is confusing; please rename or explicitly disambiguate.
  4. [§3.2.3] There is a duplicated word in the sentence introducing LogBD: 'Temporal Convolutional Networks Temporal Convolutional Networkss' should be 'Temporal Convolutional Networks'.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the central AD and labeling derivations are evaluated against independent ground truth, and the PU-window premise is an explicit assumption rather than a hidden reduction.

full rationale

The dissertation's core derivations are not circular by the paper's own equations. The log-labeling and anomaly-detection methods (Chapters 5 and 6) use PU learning with failure-time windows only to construct noisy training classes P and U; the reported F1 evaluations on public datasets are scored against independent ground-truth labels, and the weak-supervision objective (Eq. 5.5) learns from the noisy split rather than importing the target labels. The RCA definition (Definition 6) is formally underspecified because Impact(li) is not defined there, and the truncated evaluation prevents fully verifying whether the RCA ground truth is independent of the window-based U; however, no passage in the provided manuscript equates the RCA target to the PU training input by construction. The PU assumption that P contains only true normal samples is stated explicitly (Section 2.4.2) and is a testable premise; concerns about imprecise monitoring windows are correctness and robustness risks, not circularity. Self-citations to the author's prior peer-reviewed papers describe the same methods but are not used as the load-bearing justification for the central claims; external citations support the transformer and PU-learning foundations. Thus no specific reduction of a prediction to its own input is exhibited.

Assumptions & free parameters 6 free parameters · 6 assumptions · 0 invented entities

The central claims rest on two explicit PU assumptions (purity of the normal class, distinctness of anomalies in the unknown class) plus the availability and quality of failure time estimates from monitoring, and on a per-line content classification setup with several per-system free parameters (δ, S, R, α, thresholds, context bounds). The taxonomy and AD scoring are algebraically sound descriptive statistics, but the architecture-level results inherit every one of these choices.

free parameters (6)
  • δ (failure time window half-width) = δ = ±10000 ms featured; swept in Tables 8.6 and 8.7
    Defines which log lines enter the unknown class U for labeling, AD, and RCA; directly controls the noise level of the training signal and the resulting F1.
  • Labeling/decision threshold τ = not specified numerically
    Step 6 of the labeling pipeline in section 5.1.3 ('threshold determined') maps final anomaly scores to binary labels; the selection rule is not given.
  • S (max tokens per log line) = 75th percentile guideline
    Truncation and padding length in section 6.3; set per system, affects which tokens the model can attend to.
  • R (regex replacement rules) = not fixed; per system
    Expert-chosen placeholder rules in section 6.3 (e.g., [HEX], [NUM]); the thesis explicitly says an expert must choose the minimum number of rules depending on the dominant anomaly type.
  • α (augmentation token replacement count) = not fixed; example α=2 in Figure 6.6
    Number of replaced tokens when generating synthetic deviations to compute the decision boundary in section 6.5.2; the boundary location and hence the F1 depend on it.
  • a, b (context boundaries) = example a=2, b=1
    Context window for the context anomaly score in Eq. 5.7; the taxonomy classification results depend on this choice.
assumptions (6)
  • domain assumption The approximated normal class P consists of true normal samples.
    Stated in section 2.4.2 as an assumption of the PU training strategy and reused in sections 5.1.2 and 7.2; if monitoring failure times are wrong, P contains anomalies.
  • domain assumption Real abnormal samples hidden in U have characteristics that differ from normal samples.
    Second PU assumption in section 2.4.2; required for the model to separate the classes from content alone.
  • domain assumption Failure times are roughly known from monitoring systems or similar sources.
    Input to all three layers (sections 4.1, 5.1.1); the thesis notes metrics cannot pinpoint exact log lines, so the method inherits the alert quality.
  • domain assumption Log anomalies of interest for AD are detectable from a single log line's content (point anomalies).
    Chapter 6 opening states the AD method concentrates on point anomalies (template and attribute), excluding contextual anomalies; the 'any system' claim is scoped by this.
  • domain assumption At least half of all log samples are normal.
    Section 5.1.6 assumes anomalies are rare in order to compute the median for score smoothing; if this fails, the iterative labels are biased.
  • standard math Transformer encoder representation is adequate for log line classification (Vaswani et al. background), plus positional encoding details.
    Standard architecture adopted in section 2.4.1; no new architectural proof is offered.

how reviews work

0 comments
Cite this review

Pith. "Pith review of A layered architecture for log analysis in complex IT systems." pith.science (2026). https://pith.science/paper/BIZLEHHY

@misc{pith2026250908698,
  author       = {Pith},
  title        = {Pith review of: A layered architecture for log analysis in complex IT systems},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/BIZLEHHY}},
  note         = {Machine review of arXiv:2509.08698}
}
read the original abstract

In the evolving IT landscape, stability and reliability of systems are essential, yet their growing complexity challenges DevOps teams in implementation and maintenance. Log analysis, a core element of AIOps, provides critical insights into complex behaviors and failures. This dissertation introduces a three-layered architecture to support DevOps in failure resolution. The first layer, Log Investigation, performs autonomous log labeling and anomaly classification. We propose a method that labels log data without manual effort, enabling supervised training and precise evaluation of anomaly detection. Additionally, we define a taxonomy that groups anomalies into three categories, ensuring appropriate method selection. The second layer, Anomaly Detection, detects behaviors deviating from the norm. We propose a flexible Anomaly Detection method adaptable to unsupervised, weakly supervised, and supervised training. Evaluations on public and industry datasets show F1-scores between 0.98 and 1.0, ensuring reliable anomaly detection. The third layer, Root Cause Analysis, identifies minimal log sets describing failures, their origin, and event sequences. By balancing training data and identifying key services, our Root Cause Analysis method consistently detects 90-98% of root cause log lines within the top 10 candidates, providing actionable insights for mitigation. Our research addresses how log analysis methods can be designed and optimized to help DevOps resolve failures efficiently. By integrating these three layers, the architecture equips teams with robust methods to enhance IT system reliability.

Figures

Figures reproduced from arXiv: 2509.08698 by the authors.

Figure 2
Figure 2. [PITH_FULL_IMAGE:figures/full_fig_p028_2.png] view at source ↗
Figure 2.1
Figure 2.1. The AIOps lifecycle inspired by Gartner decision making based on the data analyzed. This phase involves the utilization of au￾tomation technologies to respond quickly and intelligently to the insights gained from the analyzing stage based on the monitored information and data. Automation facilitates the execution of predefined actions or the initiation of dynamic responses to optimize IT operations, mitigate risks, … view at source ↗
Figure 2.2
Figure 2.2. Example of embedding vectors in a two dimensional space. [PITH_FULL_IMAGE:figures/full_fig_p035_2_2.png] view at source ↗
Figures from the paper (51 more)
Figure 2
Figure 2. Figure 2 [PITH_FULL_IMAGE:figures/full_fig_p040_2.png]
Figure 2.3
Figure 2.3. Figure 2.3: Transformer encoder architecture. by the application of positional encoding to each embedding, due to the fact that the transformer lacks the capability to recognize the sequence of tokens. Let L = {l1, l2, . . . , ln} be a sequence of log lines, ti the tokenized con…
Figure 2
Figure 2. Figure 2 [PITH_FULL_IMAGE:figures/full_fig_p042_2.png]
Figure 4
Figure 4. Figure 4 [PITH_FULL_IMAGE:figures/full_fig_p058_4.png]
Figure 4.1
Figure 4.1. Figure 4.1: Schematic representation of our layered architecture, illus [PITH_FULL_IMAGE:figures/full_fig_p059_4_1.png]
Figure 5.1
Figure 5.1. Figure 5.1: We use rough estimates for failure times provided by mon [PITH_FULL_IMAGE:figures/full_fig_p068_5_1.png]
Figure 5
Figure 5. Figure 5 [PITH_FULL_IMAGE:figures/full_fig_p069_5.png]
Figure 5.2
Figure 5.2. Figure 5.2: 6 steps to obtain labels for single log lines [PITH_FULL_IMAGE:figures/full_fig_p070_5_2.png]
Figure 2
Figure 2. Figure 2 [PITH_FULL_IMAGE:figures/full_fig_p071_2.png]
Figure 5.3
Figure 5.3. Figure 5.3: Anomaly scores are smoothed for training. [PITH_FULL_IMAGE:figures/full_fig_p074_5_3.png]
Figure 5
Figure 5. Figure 5 [PITH_FULL_IMAGE:figures/full_fig_p075_5.png]
Figure 5.4
Figure 5.4. Figure 5.4: Two examples for point anomalies. On the left side: Point [PITH_FULL_IMAGE:figures/full_fig_p076_5_4.png]
Figure 5.5
Figure 5.5. Figure 5.5: Two examples for point anomalies in written text. [PITH_FULL_IMAGE:figures/full_fig_p076_5_5.png]
Figure 5.6
Figure 5.6. Figure 5.6: One example of a contextual anomaly in time series and two [PITH_FULL_IMAGE:figures/full_fig_p077_5_6.png]
Figure 5.7
Figure 5.7. Figure 5.7: Taxonomy for anomalies in log data. an abnormal log message that is characterized by its own content. The log line could be classified as anomalous only by investigating its content and without observing its context. Hence, the unusual behavior of a log message is de…
Figure 5
Figure 5. Figure 5 [PITH_FULL_IMAGE:figures/full_fig_p078_5.png]
Figure 5.8
Figure 5.8. Figure 5.8: Mining process of the different anomaly types. [PITH_FULL_IMAGE:figures/full_fig_p078_5_8.png]
Figure 6
Figure 6. Figure 6 [PITH_FULL_IMAGE:figures/full_fig_p085_6.png]
Figure 6.1
Figure 6.1. Figure 6.1: Structure of our general anomaly detection method including [PITH_FULL_IMAGE:figures/full_fig_p085_6_1.png]
Figure 6
Figure 6. Figure 6 [PITH_FULL_IMAGE:figures/full_fig_p086_6.png]
Figure 6.2
Figure 6.2. Figure 6.2: Preprocessing steps for every single log line. [PITH_FULL_IMAGE:figures/full_fig_p087_6_2.png]
Figure 6
Figure 6. Figure 6 [PITH_FULL_IMAGE:figures/full_fig_p089_6.png]
Figure 6
Figure 6. Figure 6 [PITH_FULL_IMAGE:figures/full_fig_p092_6.png]
Figure 6
Figure 6. Figure 6 [PITH_FULL_IMAGE:figures/full_fig_p094_6.png]
Figure 6.3
Figure 6.3. Figure 6.3: Part of the objective function for unsupervised learning to [PITH_FULL_IMAGE:figures/full_fig_p095_6_3.png]
Figure 6
Figure 6. Figure 6 [PITH_FULL_IMAGE:figures/full_fig_p098_6.png]
Figure 6
Figure 6. Figure 6 [PITH_FULL_IMAGE:figures/full_fig_p099_6.png]
Figure 6.4
Figure 6.4. Figure 6.4: Illustration of the various possibilities for setting a decision [PITH_FULL_IMAGE:figures/full_fig_p099_6_4.png]
Figure 6.5
Figure 6.5. Figure 6.5: Illustration of the various possibilities for setting a decision [PITH_FULL_IMAGE:figures/full_fig_p100_6_5.png]
Figure 6
Figure 6. Figure 6 [PITH_FULL_IMAGE:figures/full_fig_p101_6.png]
Figure 6.6
Figure 6.6. Figure 6.6: Data augmentation for a log message with [PITH_FULL_IMAGE:figures/full_fig_p102_6_6.png]
Figure 7.1
Figure 7.1. Figure 7.1: Illustration of the problem and the desired solution for root [PITH_FULL_IMAGE:figures/full_fig_p109_7_1.png]
Figure 7.2
Figure 7.2. Figure 7.2: Illustrating the training process with false labeled data and [PITH_FULL_IMAGE:figures/full_fig_p110_7_2.png]
Figure 7.3
Figure 7.3. Figure 7.3: Balancing the training data [PITH_FULL_IMAGE:figures/full_fig_p112_7_3.png]
Figure 7.4
Figure 7.4. Figure 7.4: Steps for selecting root cause candidates. [PITH_FULL_IMAGE:figures/full_fig_p114_7_4.png]
Figure 7.5
Figure 7.5. Figure 7.5: Find main involved services for each root cause. [PITH_FULL_IMAGE:figures/full_fig_p115_7_5.png]
Figure 8.1
Figure 8.1. Figure 8.1: Number of log lines in the Investigation Time Windows and [PITH_FULL_IMAGE:figures/full_fig_p123_8_1.png]
Figure 8
Figure 8. Figure 8 [PITH_FULL_IMAGE:figures/full_fig_p125_8.png]
Figure 8.2
Figure 8.2. Figure 8.2: Performance after iterative training for [PITH_FULL_IMAGE:figures/full_fig_p126_8_2.png]
Figure 8.3
Figure 8.3. Figure 8.3: Percentage of anomalies by type at different thresholds. [PITH_FULL_IMAGE:figures/full_fig_p130_8_3.png]
Figure 8
Figure 8. Figure 8 [PITH_FULL_IMAGE:figures/full_fig_p133_8.png]
Figure 8.4
Figure 8.4. Figure 8.4: F1-Scores for predicting attribute, template, and contextual [PITH_FULL_IMAGE:figures/full_fig_p134_8_4.png]
Figure 8
Figure 8. Figure 8 [PITH_FULL_IMAGE:figures/full_fig_p135_8.png]
Figure 8.5
Figure 8.5. Figure 8.5: Evaluation of our decision boundary against the best possible [PITH_FULL_IMAGE:figures/full_fig_p136_8_5.png]
Figure 8
Figure 8. Figure 8 [PITH_FULL_IMAGE:figures/full_fig_p136_8.png]
Figure 8.6
Figure 8.6. Figure 8.6: F1-Scores of different methods in an unsupervised training [PITH_FULL_IMAGE:figures/full_fig_p137_8_6.png]
Figure 8.7
Figure 8.7. Figure 8.7: F1-Scores of different methods in a weak-supervised training [PITH_FULL_IMAGE:figures/full_fig_p139_8_7.png]
Figure 8.8
Figure 8.8. Figure 8.8: F1-Scores of different methods in a supervised training sce [PITH_FULL_IMAGE:figures/full_fig_p141_8_8.png]
Figure 8
Figure 8. Figure 8 [PITH_FULL_IMAGE:figures/full_fig_p144_8.png]
Figure 8.9
Figure 8.9. Figure 8.9: Fraction of root case log lines at 10/20/50 returned candi [PITH_FULL_IMAGE:figures/full_fig_p144_8_9.png]
Figure 8.10
Figure 8.10. Figure 8.10: Average precision and recall of all approaches with and [PITH_FULL_IMAGE:figures/full_fig_p145_8_10.png]

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

170 extracted references · 75 canonical work pages

  1. [1]

    Reliability engineering: Old problems and new challenges

    Enrico Zio. “Reliability engineering: Old problems and new challenges”. In:Reli- ability engineering & system safety94.2 (2009), pp. 125–141

    2009

  2. [2]

    Basicconcepts and taxonomy of dependable and secure computing

    AlgirdasAvizienis,J-CLaprie,BrianRandell,andCarlLandwehr.“Basicconcepts and taxonomy of dependable and secure computing”. In:IEEE transactions on dependable and secure computing1.1 (2004), pp. 11–33

    2004

  3. [3]

    Why do computers stop and what can be done about it?

    Jim Gray. “Why do computers stop and what can be done about it?” In:Sympo- sium on reliability in distributed software and database systems. Los Angeles, CA, USA. 1986, pp. 3–12

    1986

  4. [4]

    The internet of things: A survey

    Luigi Atzori, Antonio Iera, and Giacomo Morabito. “The internet of things: A survey”. In:Computer networks 54.15 (2010), pp. 2787–2805

    2010

  5. [5]

    Transforming business using digital innovations: The ap- plication of AI, blockchain, cloud and data analytics

    Shahriar Akter, Katina Michael, Muhammad Rajib Uddin, Grace McCarthy, and Mahfuzur Rahman. “Transforming business using digital innovations: The ap- plication of AI, blockchain, cloud and data analytics”. In:Annals of Operations Research (2022), pp. 1–33

    2022

  6. [6]

    BigDataAnalyticsin Healthcare: exploring the role of Machine Learning in Predicting patient outcomes and improving Healthcare Delivery

    FedericoDelGiorgioSolfaandFernandoRogelioSimonato.“BigDataAnalyticsin Healthcare: exploring the role of Machine Learning in Predicting patient outcomes and improving Healthcare Delivery”. In:International Journal of Computations, Information and Manufacturing (IJCIM)3 (2023)

    2023

  7. [7]

    Digital transformation and strategy in the banking sector: Evaluating the acceptance rate of e-services

    Fotis Kitsios, Ioannis Giatsidis, and Maria Kamariotou. “Digital transformation and strategy in the banking sector: Evaluating the acceptance rate of e-services”. In: Journal of Open Innovation: Technology, Market, and Complexity7.3 (2021), p. 204

    2021

  8. [8]

    The Impact of Social Media Use on Social Interaction in Contemporary Society

    Hizbul Khootimah Azzaakiyyah. “The Impact of Social Media Use on Social Interaction in Contemporary Society”. In: Technology and Society Perspectives (TACIT) 1.1 (2023), pp. 1–9. 133 134 BIBLIOGRAPHY

    2023

Show all 170 references
  1. [9]

    Cloud computing: state-of-the-art and research challenges

    Qi Zhang, Lu Cheng, and Raouf Boutaba. “Cloud computing: state-of-the-art and research challenges”. In:Journal of internet services and applications1 (2010), pp. 7–18

    2010

  2. [10]

    Analyzing the use of concept maps in computer science: A systematic mapping study

    Vinicius dos Santos, Erica F de Souza, Katia R Felizardo, and Nandamudi L Vi- jaykumar. “Analyzing the use of concept maps in computer science: A systematic mapping study”. In:Informatics in Education16.2 (2017), pp. 257–288

    2017

  3. [11]

    Rapid business and IT change: drivers for strategic information systems planning?

    Henry E Newkirk, Albert L Lederer, and Alice M Johnson. “Rapid business and IT change: drivers for strategic information systems planning?” In:European Journal of Information Systems17 (2008), pp. 198–218

    2008

  4. [12]

    Debugging reinvented: asking and answering why and why not questions about program behavior

    Amy J Ko and Brad A Myers. “Debugging reinvented: asking and answering why and why not questions about program behavior”. In:Proceedings of the 30th in- ternational conference on Software engineering. 2008, pp. 301–310

    2008

  5. [13]

    Assessing de- pendability with software fault injection: A survey

    Roberto Natella, Domenico Cotroneo, and Henrique S Madeira. “Assessing de- pendability with software fault injection: A survey”. In:ACM Computing Surveys (CSUR) 48.3 (2016), pp. 1–55

    2016

  6. [14]

    The organization of complex systems

    Herbert A Simon and Herbert A Simon. “The organization of complex systems”. In:Models of discovery: And other topics in the methods of science(1977),pp.245– 261

    1977

  7. [15]

    DevOps capabilities, practices, and challenges: Insights from a case study

    Mali Senapathi, Jim Buchan, and Hady Osman. “DevOps capabilities, practices, and challenges: Insights from a case study”. In:Proceedings of the 22nd Interna- tional Conference on Evaluation and Assessment in Software Engineering 2018. 2018, pp. 57–67

    2018

  8. [16]

    Addison-Wesley Professional, 2015

    Len Bass, Ingo Weber, and Liming Zhu.DevOps: A software architect’s perspec- tive. Addison-Wesley Professional, 2015

    2015

  9. [17]

    Evolving from tra- ditional systems to AIOps: design, implementation and measurements

    Shijun Shen, Jiuling Zhang, Daochao Huang, and Jun Xiao. “Evolving from tra- ditional systems to AIOps: design, implementation and measurements”. In:2020 IEEE International Conference on Advances in Electrical Engineering and Com- puter Applications (AEECA). IEEE. 2020, pp. 276–280

    2020

  10. [18]

    Logmine: Fast pattern recognition for log analytics

    Hossein Hamooni, Biplob Debnath, Jianwu Xu, Hui Zhang, Guofei Jiang, and Ab- dullah Mueen. “Logmine: Fast pattern recognition for log analytics”. In:ACM In- ternational on Conference on Information and Knowledge Management (CIKM). 2016

    2016

  11. [19]

    Aiops: real-world challenges and research innovations

    Yingnong Dang, Qingwei Lin, and Peng Huang. “Aiops: real-world challenges and research innovations”. In:2019 IEEE/ACM 41st International Conference on Software Engineering: Companion Proceedings (ICSE-Companion). IEEE. 2019, pp. 4–5. BIBLIOGRAPHY 135

    2019

  12. [20]

    Artificialintelligence for it operations (aiops) workshop white paper

    JasminBogatinovski,SashoNedelkoski,AlexanderAcker,FlorianSchmidt,Thors- tenWittkopp,SoerenBecker,JorgeCardoso,andOdejKao.“Artificialintelligence for it operations (aiops) workshop white paper”. In:arXiv preprint arXiv:2101.- 06054 (2021)

    2021

  13. [21]

    A systematic mapping study in AIOps

    Paolo Notaro, Jorge Cardoso, and Michael Gerndt. “A systematic mapping study in AIOps”. In:International Conference on Service-Oriented Computing. Springer. 2020, pp. 110–123

    2020

  14. [22]

    Towards aiops in edge computing environments

    Soeren Becker, Florian Schmidt, Anton Gulenko, Alexander Acker, and Odej Kao. “Towards aiops in edge computing environments”. In:2020 IEEE International Conference on Big Data (Big Data). IEEE. 2020, pp. 3470–3475

    2020

  15. [23]

    A survey on log research of aiops: Methods and trends

    Jiang Zhaoxue, Li Tong, Zhang Zhenguo, Ge Jingguo, You Junling, and Li Liangx- iong. “A survey on log research of aiops: Methods and trends”. In:Mobile Networks and Applications26.6 (2021), pp. 2353–2364

    2021

  16. [24]

    A system architecture for real-time anomaly detection in large-scale nfv systems

    Anton Gulenko, Marcel Wallschläger, Florian Schmidt, Odej Kao, and Feng Liu. “A system architecture for real-time anomaly detection in large-scale nfv systems”. In: Procedia Computer Science94 (2016), pp. 491–496

    2016

  17. [25]

    Operational-log analysis for big data systems: Challenges and solutions

    Andriy Miranskyy, Abdelwahab Hamou-Lhadj, Enzo Cialini, and Alf Larsson. “Operational-log analysis for big data systems: Challenges and solutions”. In: IEEE Software33.2 (2016), pp. 52–59

    2016

  18. [26]

    Advances and challenges in log analysis

    Adam J. Oliner, Archana Ganapathi, and Wei Xu. “Advances and challenges in log analysis”. In:Commun. ACM 55.2 (2012), pp. 55–61

    2012

  19. [27]

    Beehive: Large-scale log analysis for detecting suspicious activity in enterprise networks

    Ting-Fang Yen, Alina Oprea, Kaan Onarlioglu, Todd Leetham, William Robert- son, Ari Juels, and Engin Kirda. “Beehive: Large-scale log analysis for detecting suspicious activity in enterprise networks”. In:Proceedings of the 29th annual com- puter security applications conferen...

    2013

  20. [28]

    Detecting Large-Scale System Problems by Mining Console Logs

    Wei Xu, Ling Huang, Armando Fox, David A. Patterson, and Michael I. Jordan. “Detecting Large-Scale System Problems by Mining Console Logs”. In:Proceedings of the 27th International Conference on Machine Learning (ICML-10), June 21- 24, 2010, Haifa, Israel. Omnipress, 2010, pp. 37–46

    2010

  21. [29]

    Mining temporal invariants from partially ordered logs

    Ivan Beschastnikh, Yuriy Brun, Michael D Ernst, Arvind Krishnamurthy, and Thomas E Anderson. “Mining temporal invariants from partially ordered logs”. In: Managing Large-scale Systems via the Analysis of System Logs and the Application of Machine Learning Techniques. 2011, pp. 1–10

    2011

  22. [30]

    Mining program workflow from inter- leaved logs

    J-GLQF Shengqi and YANG Jiang LI. “Mining program workflow from inter- leaved logs”. In:Proceedings of the 16th ACM SIGKDD international conference on Knowledge discovery and data mining (KDD’10). 2010. 136 BIBLIOGRAPHY

    2010

  23. [31]

    Efficient fault detection and diagnosis in complex software systems with informa- tion-theoreticmonitoring

    Miao Jiang, Mohammad A Munawar, Thomas Reidemeister, and Paul AS Ward. “Efficient fault detection and diagnosis in complex software systems with informa- tion-theoreticmonitoring”.In: IEEE Transactions on Dependable and Secure Com- puting 8.4 (2011), pp. 510–522

    2011

  24. [32]

    Execution anomaly detection in distributed systems through unstructured log analysis

    Qiang Fu, Jian-Guang Lou, Yi Wang, and Jiang Li. “Execution anomaly detection in distributed systems through unstructured log analysis”. In:2009 ninth IEEE international conference on data mining. IEEE. 2009, pp. 149–158

    2009

  25. [33]

    Kafka: A distributed messaging sys- tem for log processing

    Jay Kreps, Neha Narkhede, Jun Rao, et al. “Kafka: A distributed messaging sys- tem for log processing”. In: Proceedings of the NetDB. Vol. 11. 2011. Athens, Greece. 2011, pp. 1–7

    2011

  26. [34]

    LogBERT: Log Anomaly Detection via BERT

    Haixuan Guo, Shuhan Yuan, and Xintao Wu. “LogBERT: Log Anomaly Detection via BERT”. In:International Joint Conference on Neural Networks, IJCNN 2021, Shenzhen, China, July 18-22, 2021. IEEE, 2021, pp. 1–8

    2021

  27. [35]

    Deeplog: Anomaly de- tection and diagnosis from system logs through deep learning

    Min Du, Feifei Li, Guineng Zheng, and Vivek Srikumar. “Deeplog: Anomaly de- tection and diagnosis from system logs through deep learning”. In:Proceedings of the 2017 ACM SIGSAC Conference on Computer and Communications Security. 2017, pp. 1285–1298

    2017

  28. [36]

    Self-supervised log parsing

    Sasho Nedelkoski, Jasmin Bogatinovski, Alexander Acker, Jorge Cardoso, and Odej Kao. “Self-supervised log parsing”. In: Machine Learning and Knowledge Discovery in Databases: Applied Data Science Track: European Conference, ECML PKDD 2020, Ghent, Belgium, September 14–18, 2020...

    2020

  29. [37]

    Log filtering and interpretation for root cause analysis

    Hamzeh Zawawy, Kostas Kontogiannis, and John Mylopoulos. “Log filtering and interpretation for root cause analysis”. In:IEEE International Conference on Soft- ware Maintenance (ICSM). 2010

    2010

  30. [38]

    LogLAB: Attention-Based Labeling of Log Data Anomalies via Weak Supervi- sion

    Thorsten Wittkopp, Philipp Wiesner, Dominik Scheinert, and Alexander Acker. “LogLAB: Attention-Based Labeling of Log Data Anomalies via Weak Supervi- sion”. In: Service-Oriented Computing - 19th International Conference, ICSOC 2021, Virtual Event, November 22-25, 2021, Proceed...

    2021

  31. [39]

    Landscape of Automated Log Anal- ysis: A Systematic Literature Review and Mapping Study

    Łukasz Korzeniowski and Krzysztof Goczyła. “Landscape of Automated Log Anal- ysis: A Systematic Literature Review and Mapping Study”. In: IEEE Access (2022)

    2022

  32. [40]

    Log-based abnormal task detection and root cause analysis for spark

    Siyang Lu, BingBing Rao, Xiang Wei, Byungchul Tak, Long Wang, and Liqiang Wang. “Log-based abnormal task detection and root cause analysis for spark”. In: IEEE International Conference on Web Services (ICWS). 2017. BIBLIOGRAPHY 137

    2017

  33. [41]

    A2Log: Attentive Augmented Log Anomaly Detection

    Thorsten Wittkopp, Alexander Acker, Sasho Nedelkoski, Jasmin Bogatinovski, Dominik Scheinert, Wu Fan, and Odej Kao. “A2Log: Attentive Augmented Log Anomaly Detection”. In: 55th Hawaii International Conference on System Sci- ences, HICSS 2022, Virtual Event / Maui, Hawaii, USA,...

    2022

  34. [42]

    PULL: Reactive Log Anomaly Detection Based On Iterative PU Learning

    Thorsten Wittkopp, Dominik Scheinert, Philipp Wiesner, Alexander Acker, and Odej Kao. “PULL: Reactive Log Anomaly Detection Based On Iterative PU Learning”. In:56th Hawaii International Conference on System Sciences, HICSS 2023, Maui, Hawaii, USA, January 3-6, 2023. ScholarSpa...

    2023

  35. [43]

    SwissLog: Robust and Unified Deep Learning Based Log Anomaly Detection for Diverse Faults

    Xiaoyun Li, Pengfei Chen, Linxiao Jing, Zilong He, and Guangba Yu. “SwissLog: Robust and Unified Deep Learning Based Log Anomaly Detection for Diverse Faults”. In:31st IEEE International Symposium on Software Reliability Engineer- ing, ISSRE 2020, Coimbra, Portugal, October 12...

    2020

  36. [44]

    LogAnomaly: Un- supervised Detection of Sequential and Quantitative Anomalies in Unstructured Logs

    Weibin Meng, Ying Liu, Yichen Zhu, Shenglin Zhang, Dan Pei, Yuqing Liu, Yihao Chen, Ruizhi Zhang, Shimin Tao, Pei Sun, and Rong Zhou. “LogAnomaly: Un- supervised Detection of Sequential and Quantitative Anomalies in Unstructured Logs”. In: Proceedings of the Twenty-Eighth Inte...

    2019

  37. [45]

    LogRule: Ef- ficient Structured Log Mining for Root Cause Analysis

    Paolo Notaro, Soroush Haeri, Jorge Cardoso, and Michael Gerndt. “LogRule: Ef- ficient Structured Log Mining for Root Cause Analysis”. In:IEEE Transactions on Network and Service Management(2023)

    2023

  38. [46]

    LADRA: Log-based abnormal task detection and root-cause analysis in big data processing with Spark

    Siyang Lu, Xiang Wei, Bingbing Rao, Byungchul Tak, Long Wang, and Liqiang Wang. “LADRA: Log-based abnormal task detection and root-cause analysis in big data processing with Spark”. In: Future Generation Computer Systems 95 (2019)

    2019

  39. [47]

    Log- master: Mining event correlations in logs of large-scale cluster systems

    Xiaoyu Fu, Rui Ren, Jianfeng Zhan, Wei Zhou, Zhen Jia, and Gang Lu. “Log- master: Mining event correlations in logs of large-scale cluster systems”. In:2012 IEEE 31st Symposium on Reliable Distributed Systems. IEEE. 2012, pp. 71–80

    2012

  40. [48]

    Log-based predictive maintenance

    Ruben Sipos, Dmitriy Fradkin, Fabian Moerchen, and Zhuang Wang. “Log-based predictive maintenance”. In:Proceedings of the 20th ACM SIGKDD international conference on knowledge discovery and data mining. 2014, pp. 1867–1876

    2014

  41. [49]

    Experience Report: System Log Analysis for Anomaly Detection

    Shilin He, Jieming Zhu, Pinjia He, and Michael R. Lyu. “Experience Report: System Log Analysis for Anomaly Detection”. In:27th IEEE International Sym- posium on Software Reliability Engineering, ISSRE 2016, Ottawa, ON, Canada, October 23-27, 2016. IEEE Computer Society, 2016, ...

    2016

  42. [50]

    How to im- prove cloud services availability? Investigating the impact of power and it subsys- tems failures

    Daniel Rosendo, Guto Leoni, Demis Gomes, André Moreira, Glauco Gonçalves, Patricia Endo, Judith Kelner, Djamel Sadok, and Mozhgan Mahloo. “How to im- prove cloud services availability? Investigating the impact of power and it subsys- tems failures”. In:Hawaii International Con...

    2018

  43. [51]

    A survey of aiops methods for failure management

    Paolo Notaro, Jorge Cardoso, and Michael Gerndt. “A survey of aiops methods for failure management”. In:ACM Transactions on Intelligent Systems and Tech- nology (TIST)12.6 (2021), pp. 1–45

    2021

  44. [52]

    AI-Governance and Levels of Automation for AIOps-supported System Administration

    Anton Gulenko, Alexander Acker, Odej Kao, and Feng Liu. “AI-Governance and Levels of Automation for AIOps-supported System Administration”. In: IEEE International Conference on Computer Communications and Networks (ICCCN). 2020

    2020

  45. [53]

    AIOps-Artificial Intelligence for IT Operations: Todays Challenges of new Technologies and new Methodologies in IT Operations

    Martin Andenmatten. “AIOps-Artificial Intelligence for IT Operations: Todays Challenges of new Technologies and new Methodologies in IT Operations”. In: HMD Praxis der Wirtschaftsinformatik56 (2019), pp. 332–344

    2019

  46. [54]

    On the future of cloud engineering

    David Bermbach, Abhishek Chandra, Chandra Krintz, Aniruddha Gokhale, Alek- sanderSlominski,LauritzThamsen,EvertonCavalcante,TianGuo,IvonaBrandic, and Rich Wolski. “On the future of cloud engineering”. In:2021 IEEE Interna- tional conference on cloud engineering (IC2E). IEEE. 2...

    2021

  47. [55]

    Distributed systems observability: a guide to building robust sys- tems

    Cindy Sridharan. Distributed systems observability: a guide to building robust sys- tems. O’Reilly Media, 2018

    2018

  48. [56]

    A Generic Review on Anomaly Detection

    Nishu Bansal and Swimpy Pahuja. “A Generic Review on Anomaly Detection”. In:Proceedings of 3rd International Conference on Machine Learning, Advances in Computing, Renewable Energy and Communication. Singapore: Springer Nature Singapore, 2022, pp. 495–506

    2022

  49. [57]

    A review of ma- chine learning and deep learning techniques for anomaly detection in IoT data

    Redhwan Al-amri, Raja Kumar Murugesan, Mustafa Man, Alaa Fareed Abdula- teef, Mohammed A Al-Sharafi, and Ammar Ahmed Alkahtani. “A review of ma- chine learning and deep learning techniques for anomaly detection in IoT data”. In: Applied Sciences11.12 (2021), p. 5320

    2021

  50. [58]

    Aiops:Amul- tivocal literature review

    LaxmiRijal,RicardoColomo-Palacios,andMarySánchez-Gordón.“Aiops:Amul- tivocal literature review”. In:Artificial Intelligence for Cloud and Edge Computing (2022), pp. 31–50

    2022

  51. [59]

    Optimal IT service availability: Shorter outages, or fewer?

    Ulrik Franke. “Optimal IT service availability: Shorter outages, or fewer?” In: IEEE Transactions on Network and Service Management9.1 (2011), pp. 22–33. BIBLIOGRAPHY 139

    2011

  52. [60]

    A joint study of the challenges, opportunities, and roadmap of mlops and aiops: A systematic survey

    Josu Diaz-De-Arcaya, Ana I Torre-Bastida, Gorka Zárate, Raúl Miñón, and Aitor Almeida. “A joint study of the challenges, opportunities, and roadmap of mlops and aiops: A systematic survey”. In:ACM Computing Surveys56.4 (2023), pp. 1– 30

    2023

  53. [61]

    Microrca: Root cause local- ization of performance issues in microservices

    Li Wu, Johan Tordsson, Erik Elmroth, and Odej Kao. “Microrca: Root cause local- ization of performance issues in microservices”. In:NOMS 2020-2020 IEEE/IFIP Network Operations and Management Symposium. IEEE. 2020, pp. 1–9

    2020

  54. [62]

    CloudRCA: A root cause analysis framework for cloud computing platforms

    Yingying Zhang, Zhengxiong Guan, Huajie Qian, Leili Xu, Hengbo Liu, Qingsong Wen, Liang Sun, Junwei Jiang, Lunting Fan, and Min Ke. “CloudRCA: A root cause analysis framework for cloud computing platforms”. In:Proceedings of the 30th ACM International Conference on Information...

    2021

  55. [63]

    Real-time root cause analysis for complex technical systems

    Jan Eric Larsson and Joseph DeBor. “Real-time root cause analysis for complex technical systems”. In: 2007 IEEE 8th Human Factors and Power Plants and HPRCT 13th Annual Meeting. IEEE. 2007, pp. 156–163

    2007

  56. [64]

    Qulog: Data-driven approach for log instruction quality assessment

    Jasmin Bogatinovski, Sasho Nedelkoski, Alexander Acker, Jorge Cardoso, and Odej Kao. “Qulog: Data-driven approach for log instruction quality assessment”. In: Proceedings of the 30th IEEE/ACM International Conference on Program Comprehension. 2022, pp. 275–286

    2022

  57. [65]

    Drain: An online log parsing approach with fixed depth tree

    Pinjia He, Jieming Zhu, Zibin Zheng, and Michael R Lyu. “Drain: An online log parsing approach with fixed depth tree”. In:IEEE International Conference on Web Services (ICWS). 2017

    2017

  58. [66]

    A tax- onomy of anomalies in log data

    Thorsten Wittkopp, Philipp Wiesner, Dominik Scheinert, and Odej Kao. “A tax- onomy of anomalies in log data”. In:International Conference on Service-Oriented Computing. Springer. 2021, pp. 153–164

    2021

  59. [67]

    Bert: Pre- training of deep bidirectional transformers for language understanding

    Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. “Bert: Pre- training of deep bidirectional transformers for language understanding”. In:arXiv preprint arXiv:1810.04805 (2018)

    2018 arXiv

  60. [68]

    Fast, consistent tokenization of natural language text

    Lincoln A. Mullen, Kenneth Benoit, Os Keyes, Dmitry Selivanov, and Jeffrey Arnold. “Fast, consistent tokenization of natural language text”. In:Journal of Open Source Software3.23 (2018), p. 655

    2018

  61. [69]

    Natural lan- guage processing: state of the art, current trends and challenges

    Diksha Khurana, Aditya Koli, Kiran Khatter, and Sukhdev Singh. “Natural lan- guage processing: state of the art, current trends and challenges”. In:Multim. Tools Appl.82.3 (2023), pp. 3713–3744. 140 BIBLIOGRAPHY

    2023

  62. [70]

    Tools and benchmarks for automated log parsing

    Jieming Zhu, Shilin He, Jinyang Liu, Pinjia He, Qi Xie, Zibin Zheng, and Michael R. Lyu. “Tools and benchmarks for automated log parsing”. In:Proceedings of the 41st International Conference on Software Engineering: Software Engineering in Practice, ICSE (SEIP) 2019, Montreal,...

    2019

  63. [71]

    A survey on automated log analysis for reliability engineering

    Shilin He, Pinjia He, Zhuangbin Chen, Tianyi Yang, Yuxin Su, and Michael R Lyu. “A survey on automated log analysis for reliability engineering”. In:ACM computing surveys (CSUR)54.6 (2021), pp. 1–37

    2021

  64. [72]

    Unsupervised learning based on artificial neural network: A review

    Happiness Ugochi Dike, Yimin Zhou, Kranthi Kumar Deveerasetty, and Qingtian Wu. “Unsupervised learning based on artificial neural network: A review”. In:2018 IEEE International Conference on Cyborg and Bionic Systems (CBS).IEEE.2018, pp. 322–327

    2018

  65. [73]

    Unsupervised K-means clustering algo- rithm

    Kristina P Sinaga and Miin-Shen Yang. “Unsupervised K-means clustering algo- rithm”. In:IEEE access8 (2020), pp. 80716–80727

    2020

  66. [74]

    Principal Component Analysis.SpringerSeriesinStatistics.Springer,

    IanT.Jolliffe. Principal Component Analysis.SpringerSeriesinStatistics.Springer,

  67. [75]

    An analysis of the t-sne al- gorithm for data visualization

    Sanjeev Arora, Wei Hu, and Pravesh K Kothari. “An analysis of the t-sne al- gorithm for data visualization”. In:Conference on learning theory. PMLR. 2018, pp. 1455–1462

    2018

  68. [76]

    Neural Unsupervised Domain Adaptation in NLP - A Survey

    Alan Ramponi and Barbara Plank. “Neural Unsupervised Domain Adaptation in NLP - A Survey”. In:arXiv preprint arXiv:2006.00632(2020)

    2020 arXiv

  69. [77]

    Small data challenges in big data era: A survey of recent progress on unsupervised and semi-supervised methods

    Guo-Jun Qi and Jiebo Luo. “Small data challenges in big data era: A survey of recent progress on unsupervised and semi-supervised methods”. In:IEEE Trans- actions on Pattern Analysis and Machine Intelligence(2020)

    2020

  70. [78]

    How to cope with change?- preserving validity of predictive services over time

    Lucas Baier, Niklas Kühl, and Gerhard Satzger. “How to cope with change?- preserving validity of predictive services over time”. In:Proceedings of the 52nd Hawaii International Conference on System Sciences. 2019

    2019

  71. [79]

    A brief introduction to weakly supervised learning

    Zhi-Hua Zhou. “A brief introduction to weakly supervised learning”. In:National science review5.1 (2018)

    2018

  72. [80]

    Learning from positive and unlabeled data: A survey

    Jessa Bekker and Jesse Davis. “Learning from positive and unlabeled data: A survey”. In:Machine Learning109.4 (2020)

    2020

  73. [81]

    Introduction to Semi-Supervised Learn- ing

    Xiaojin Zhu and Andrew B. Goldberg. “Introduction to Semi-Supervised Learn- ing”. In: Synthesis Lectures on Artificial Intelligence and Machine Learning (2009)

    2009

  74. [82]

    Multiple instance learning: A survey of problem characteristics and appli- cations

    Marc-André Carbonneau, Veronika Cheplygina, Eric Granger, and Ghyslain Gag- non. “Multiple instance learning: A survey of problem characteristics and appli- cations”. In:Pattern Recognition77 (2018), pp. 329–353. BIBLIOGRAPHY 141

    2018

  75. [83]

    Partially supervised classifi- cation of text documents

    Bing Liu, Wee Sun Lee, Philip S Yu, and Xiaoli Li. “Partially supervised classifi- cation of text documents”. In:ICML. Sydney, NSW. 2002

    2002

  76. [84]

    Decentralized Federated Learning Pre- serves Model and Data Privacy

    Thorsten Wittkopp and Alexander Acker. “Decentralized Federated Learning Pre- serves Model and Data Privacy”. In:Service-Oriented Computing - ICSOC 2020 Workshops - AIOps, CFTIC, STRAPS, AI-PA, AI-IOTS, and Satellite Events, Dubai, United Arab Emirates, December 14-17, 2020, P...

    2020

  77. [85]

    Time series data augmentation for deep learning: A survey

    Qingsong Wen, Liang Sun, Fan Yang, Xiaomin Song, Jingkun Gao, Xue Wang, and Huan Xu. “Time series data augmentation for deep learning: A survey”. In: arXiv preprint arXiv:2002.12478(2020)

    2020 arXiv

  78. [86]

    Robust log-based anomaly detection on unstable log data

    Xu Zhang, Yong Xu, Qingwei Lin, Bo Qiao, Hongyu Zhang, Yingnong Dang, Chunyu Xie, Xinsheng Yang, Qian Cheng, Ze Li, et al. “Robust log-based anomaly detection on unstable log data”. In:ESEC/FSE. 2019

    2019

  79. [87]

    Self-supervised learning: Generative or contrastive

    Xiao Liu, Fanjin Zhang, Zhenyu Hou, Li Mian, Zhaoyu Wang, Jing Zhang, and Jie Tang. “Self-supervised learning: Generative or contrastive”. In:IEEE transactions on knowledge and data engineering35.1 (2021), pp. 857–876

    2021

  80. [88]

    Not all unlabeled data are equal: Learning to weight data in semi-supervised learning

    Zhongzheng Ren, Raymond Yeh, and Alexander Schwing. “Not all unlabeled data are equal: Learning to weight data in semi-supervised learning”. In:Advances in Neural Information Processing Systems33 (2020), pp. 21786–21797

    2020

  81. [89]

    MIT press, 2018

    Mehryar Mohri, Afshin Rostamizadeh, and Ameet Talwalkar.Foundations of ma- chine learning. MIT press, 2018

    2018

  82. [90]

    Christopher M Bishop and Nasser M Nasrabadi.Pattern recognition and machine learning. Vol. 4. 4. Springer, 2006

    2006

  83. [91]

    The elements of statistical learning: data mining, inference, and prediction

    Trevor Hastie, Robert Tibshirani, Jerome H Friedman, and Jerome H Friedman. The elements of statistical learning: data mining, inference, and prediction. Vol. 2. Springer, 2009

    2009

  84. [92]

    Recurrent neu- ral network attention mechanisms for interpretable system log anomaly detection

    Andy Brown, Aaron Tuor, Brian Hutchinson, and Nicole Nichols. “Recurrent neu- ral network attention mechanisms for interpretable system log anomaly detection”. In: Proceedings of the first workshop on machine learning for computing systems. 2018, pp. 1–8

    2018

  85. [93]

    Deep learning for anomaly detec- tion: A survey

    Raghavendra Chalapathy and Sanjay Chawla. “Deep learning for anomaly detec- tion: A survey”. In:arXiv preprint arXiv:1901.03407(2019)

    2019 arXiv

  86. [94]

    BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding

    Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. “BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding”. In: NAACL-HLT. Association for Computational Linguistics, 2019. 142 BIBLIOGRAPHY

    2019

  87. [95]

    Attention is All you Need

    Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N. Gomez, Lukasz Kaiser, and Illia Polosukhin. “Attention is All you Need”. In: NeurIPS. 2017

    2017

  88. [96]

    Convolutional Sequence to Sequence Learning

    JonasGehring,MichaelAuli,DavidGrangier,DenisYarats,andYannN.Dauphin. “Convolutional Sequence to Sequence Learning”. In:ICML. PMLR, 2017

    2017

  89. [97]

    ImageNet Classification with Deep Convolutional Neural Networks

    Alex Krizhevsky, Ilya Sutskever, and Geoffrey E. Hinton. “ImageNet Classification with Deep Convolutional Neural Networks”. In: (2012), pp. 1106–1114

    2012

  90. [98]

    Revisiting Unreasonable Effectiveness of Data in Deep Learning Era

    Chen Sun, Abhinav Shrivastava, Saurabh Singh, and Abhinav Gupta. “Revisiting Unreasonable Effectiveness of Data in Deep Learning Era”. In:IEEE International Conference on Computer Vision, ICCV 2017, Venice, Italy, October 22-29, 2017. IEEE Computer Society, 2017, pp. 843–852

    2017

  91. [99]

    Time Series Data Augmentation for Deep Learning: A Survey

    Qingsong Wen, Liang Sun, Fan Yang, Xiaomin Song, Jingkun Gao, Xue Wang, and Huan Xu. “Time Series Data Augmentation for Deep Learning: A Survey”. In: IJCAI. ijcai.org, 2021

    2021

  92. [100]

    Snorkel: rapid training data creation with weak supervision

    Alexander Ratner, Stephen H. Bach, Henry R. Ehrenberg, Jason A. Fries, Sen Wu, and Christopher Ré. “Snorkel: rapid training data creation with weak supervision”. In: VLDB J.29.2-3 (2020), pp. 709–730

    2020

  93. [101]

    Learning classifiers from only positive and unla- beled data

    Charles Elkan and Keith Noto. “Learning classifiers from only positive and unla- beled data”. In:Proceedings of the 14th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, Las Vegas, Nevada, USA, August 24- 27, 2008. ACM, 2008, pp. 213–220

    2008

  94. [102]

    Building text classifiers using positive and unlabeled examples

    Bing Liu, Yang Dai, Xiaoli Li, Wee Sun Lee, and Philip S Yu. “Building text classifiers using positive and unlabeled examples”. In:ICDM. IEEE. 2003

    2003

  95. [103]

    Anomaly detection: A survey

    Varun Chandola, Arindam Banerjee, and Vipin Kumar. “Anomaly detection: A survey”. In:ACM computing surveys (CSUR)41.3 (2009)

    2009

  96. [104]

    A Classification Framework for Anomaly Detection

    Ingo Steinwart, Don Hush, and Clint Scovel. “A Classification Framework for Anomaly Detection.” In:Journal of Machine Learning Research6.2 (2005)

    2005

  97. [105]

    A tax- onomy and platform for anomaly detection

    Gheorghe Sebestyen, Anca Hangan, Zoltan Czako, and Gyorgy Kovacs. “A tax- onomy and platform for anomaly detection”. In:AQTR. IEEE. 2018

    2018

  98. [106]

    A review on outlier/anomalydetectionintimeseriesdata

    Ane Blazquez-Garcia, Angel Conde, Usue Mori, and Jose A Lozano. “A review on outlier/anomalydetectionintimeseriesdata”.In: arXiv preprint arXiv:2002.04236 (2020)

    2020 arXiv

  99. [107]

    Structured Com- parative Analysis of Systems Logs to Diagnose Performance Problems

    Karthik Nagaraj, Charles Edwin Killian, and Jennifer Neville. “Structured Com- parative Analysis of Systems Logs to Diagnose Performance Problems”. In:NSDI. USENIX Association, 2012. BIBLIOGRAPHY 143

    2012

  100. [108]

    Deep learning–based text classification: a comprehen- sive review

    Shervin Minaee, Nal Kalchbrenner, Erik Cambria, Narjes Nikzad, Meysam Chen- aghlu, and Jianfeng Gao. “Deep learning–based text classification: a comprehen- sive review”. In:ACM computing surveys (CSUR)54.3 (2021), pp. 1–40

    2021

  101. [109]

    Text classification algorithms: A survey

    Kamran Kowsari, Kiana Jafari Meimandi, Mojtaba Heidarysafa, Sanjana Mendu, Laura Barnes, and Donald Brown. “Text classification algorithms: A survey”. In: Information 10.4 (2019)

    2019

  102. [110]

    Induction of decision trees

    J. Ross Quinlan. “Induction of decision trees”. In:Machine learning1.1 (1986)

    1986

  103. [111]

    A survey of decision tree classifier methodology

    S Rasoul Safavian and David Landgrebe. “A survey of decision tree classifier methodology”.In: IEEE transactions on systems, man, and cybernetics21.3(1991)

    1991

  104. [112]

    Failure Diagnosis Using Decision Trees

    Mike Y. Chen, Alice X. Zheng, Jim Lloyd, Michael I. Jordan, and Eric A. Brewer. “Failure Diagnosis Using Decision Trees”. In:1st International Conference on Au- tonomic Computing (ICAC 2004), 17-19 May 2004, New York, NY, USA. IEEE Computer Society, 2004, pp. 36–43

    2004

  105. [113]

    Random decision forests

    Tin Kam Ho. “Random decision forests”. In: Third International Conference on Document Analysis and Recognition, ICDAR 1995, August 14 - 15, 1995, Mon- treal, Canada. Volume I. IEEE Computer Society, 1995, pp. 278–282

    1995

  106. [114]

    David W Hosmer Jr, Stanley Lemeshow, and Rodney X Sturdivant.Applied lo- gistic regression. Vol. 398. John Wiley & Sons, 2013

    2013

  107. [115]

    A comparative study of the Binary Logistic Regression (BLR) and Artificial Neural Network (ANN) mod- els for GIS-based spatial predicting landslides at a regional scale

    J Dou, H Yamagishi, Z Zhu, AP Yunus, and CW Chen. “A comparative study of the Binary Logistic Regression (BLR) and Artificial Neural Network (ANN) mod- els for GIS-based spatial predicting landslides at a regional scale”. In:Landslide Dynamics: ISDR-ICL Landslide Interactive T...

    2018

  108. [116]

    Large-Scale Bayesian LogisticRegressionforTextCategorization

    Alexander Genkin, David D. Lewis, and David Madigan. “Large-Scale Bayesian LogisticRegressionforTextCategorization”.In: Technometrics49.3(2007),pp.291– 304

    2007

  109. [117]

    Support vector machines

    Marti A. Hearst, Susan T Dumais, Edgar Osuna, John Platt, and Bernhard Scholkopf. “Support vector machines”. In:IEEE Intelligent Systems and their ap- plications 13.4 (1998), pp. 18–28

    1998

  110. [118]

    One-class SVMs for document classifica- tion

    Larry M Manevitz and Malik Yousef. “One-class SVMs for document classifica- tion”. In:Journal of machine Learning research2.Dec (2001)

    2001

  111. [119]

    BoosTexter: A boosting-based system for text categorization

    Robert E Schapire and Yoram Singer. “BoosTexter: A boosting-based system for text categorization”. In:Machine learning39.2 (2000)

    2000

  112. [120]

    Large scale multi-label text classification of a hierarchical dataset using Rocchio algorithm

    BJ Sowmya, KG Srinivasa, et al. “Large scale multi-label text classification of a hierarchical dataset using Rocchio algorithm”. In:CSITSS. IEEE. 2016. 144 BIBLIOGRAPHY

    2016

  113. [121]

    Text categorization using Rocchio algorithm and random forest algo- rithm

    S Thamarai Selvi, P Karthikeyan, A Vincent, V Abinaya, G Neeraja, and R Deepika. “Text categorization using Rocchio algorithm and random forest algo- rithm”. In:ICoAC. IEEE. 2017

    2017

  114. [122]

    Failure pre- diction in ibm bluegene/l event logs

    Yinglung Liang, Yanyong Zhang, Hui Xiong, and Ramendra Sahoo. “Failure pre- diction in ibm bluegene/l event logs”. In:ICDM. IEEE. 2007

    2007

  115. [123]

    Anomaly detection from log files using data mining techniques

    Jakub Breier and Jana Branišová. “Anomaly detection from log files using data mining techniques”. In: Information Science and Applications. Springer, 2015, pp. 449–457

    2015

  116. [124]

    MiningInvariants from Console Logs for System Problem Detection

    Jian-GuangLou,QiangFu,ShengqiYang,YeXu,andJiangLi.“MiningInvariants from Console Logs for System Problem Detection”. In: 2010 USENIX Annual Technical Conference, Boston, MA, USA, June 23-25, 2010.USENIXAssociation, 2010, pp. 1–14

    2010

  117. [125]

    Event Logs for the Analysis of Software Failures: A Rule-Based Approach

    Marcello Cinque, Domenico Cotroneo, and Antonio Pecchia. “Event Logs for the Analysis of Software Failures: A Rule-Based Approach”. In:IEEE Transactions on Software Engineering39.6 (2013), pp. 806–821

    2013

  118. [126]

    Relational Syn- thesis of Text and Numeric Data for Anomaly Detection on Computing System Logs

    Elisabeth Baseman, Sean Blanchard, Zongze Li, and Song Fu. “Relational Syn- thesis of Text and Numeric Data for Anomaly Detection on Computing System Logs”. In:15th IEEE International Conference on Machine Learning and Applica- tions, ICMLA 2016, Anaheim, CA, USA, December 18-...

    2016

  119. [127]

    Dynamic log file analysis: An unsupervised cluster evolution approach for anomaly detection

    Max Landauer, Markus Wurzenberger, Florian Skopik, Giuseppe Settanni, and Peter Filzmoser. “Dynamic log file analysis: An unsupervised cluster evolution approach for anomaly detection”. In:Computers & Security 79 (2018), pp. 94– 116

    2018

  120. [128]

    Improving Log-Based Anomaly Detection with Component-Aware Analysis

    Kun Yin, Meng Yan, Ling Xu, Zhou Xu, Zhao Li, Dan Yang, and Xiaohong Zhang. “Improving Log-Based Anomaly Detection with Component-Aware Analysis”. In: IEEE International Conference on Software Maintenance and Evolution, ICSME 2020, Adelaide, Australia, September 28 - October 2...

    2020

  121. [129]

    Enhanced network anomaly detection based on deep neural networks

    Sheraz Naseer, Yasir Saleem, Shehzad Khalid, Muhammad Khawar Bashir, Jihun Han, Muhammad Munwar Iqbal, and Kijun Han. “Enhanced network anomaly detection based on deep neural networks”. In:IEEE access6 (2018), pp. 48231– 48246

    2018

  122. [130]

    A Hybrid Autoencoder and Density Estimation Model for Anomaly Detection

    Van Loi Cao, Miguel Nicolau, and James McDermott. “A Hybrid Autoencoder and Density Estimation Model for Anomaly Detection”. In:Parallel Problem Solv- ing from Nature - PPSN XIV - 14th International Conference, Edinburgh, UK, BIBLIOGRAPHY 145 September 17-21, 2016, Proceedings...

    2016

  123. [131]

    AnomalyDetectionUsingAutoencoderswith Nonlinear Dimensionality Reduction

    MayuSakuradaandTakehisaYairi.“AnomalyDetectionUsingAutoencoderswith Nonlinear Dimensionality Reduction”. In:Proceedings of the MLSDA 2014 2nd Workshop on Machine Learning for Sensory Data Analysis, Gold Coast, Australia, QLD, Australia, December 2, 2014. ACM, 2014, p. 4

    2014

  124. [132]

    nLSALog: An Anomaly Detection Framework for Log Sequence in Security Management

    Ruipeng Yang, Dan Qu, Ying Gao, Yekui Qian, and Yongwang Tang. “nLSALog: An Anomaly Detection Framework for Log Sequence in Security Management”. In: IEEE Access7 (2019), pp. 181152–181164

    2019

  125. [133]

    Detecting anomaly in big data system logs using convolutional neural network

    Siyang Lu, Xiang Wei, Yandong Li, and Liqiang Wang. “Detecting anomaly in big data system logs using convolutional neural network”. In:2018 IEEE 16th Intl Conf on Dependable, Autonomic and Secure Computing, 16th Intl Conf on Pervasive Intelligence and Computing, 4th Intl Conf ...

    2018

  126. [134]

    Self-Attentive Classification-Based Anomaly Detection in Unstructured Logs

    S. Nedelkoski, J. Bogatinovski, A. Acker, J. Cardoso, and O. Kao. “Self-Attentive Classification-Based Anomaly Detection in Unstructured Logs”. In:2020 IEEE International Conference on Data Mining (ICDM). 2020, pp. 1196–1201

    2020

  127. [135]

    Unsupervised log message anomaly detec- tion

    Amir Farzad and T. Aaron Gulliver. “Unsupervised log message anomaly detec- tion”. In:ICT Express(2020)

    2020

  128. [136]

    ADA: Adaptive Deep Log Anomaly Detector

    Yali Yuan, Sripriya Srikant Adhatarao, Mingkai Lin, Yachao Yuan, Zheli Liu, and Xiaoming Fu. “ADA: Adaptive Deep Log Anomaly Detector”. In:IEEE Interna- tional Conference on Computer Communications (INFOCOM). 2020

    2020

  129. [137]

    Log clustering based problem identification for online service systems

    Qingwei Lin, Hongyu Zhang, Jian-Guang Lou, Yu Zhang, and Xuewei Chen. “Log clustering based problem identification for online service systems”. In:Interna- tional Conference on Software Engineering (ICSE). ACM, 2016

    2016

  130. [138]

    Semi-supervisedlearn- ing

    OlivierChapelle,BernhardScholkopf,andAlexanderZien.“Semi-supervisedlearn- ing”. In:IEEE Transactions on Neural Networks20.3 (2009), pp. 542–542

    2009

  131. [139]

    Discriminative unsupervised feature learning with exemplar convolutional neural networks

    Dosovitskiy Alexey, Philipp Fischer, Jost Tobias, Martin Riedmiller Springenberg, and Thomas Brox. “Discriminative unsupervised feature learning with exemplar convolutional neural networks”. In:IEEE TPAMI 38.9 (2016), pp. 1734–1747

    2016

  132. [140]

    Semi-supervised log-based anomaly detection via probabilis- tic label estimation

    Lin Yang, Junjie Chen, Zan Wang, Weijing Wang, Jiajun Jiang, Xuyuan Dong, and Wenbin Zhang. “Semi-supervised log-based anomaly detection via probabilis- tic label estimation”. In:2021 IEEE/ACM 43rd International Conference on Soft- ware Engineering (ICSE). IEEE. 2021, pp. 1448...

    2021

  133. [141]

    Device-agnostic log anomaly classification with partial labels

    Weibin Meng, Ying Liu, Shenglin Zhang, Dan Pei, Hui Dong, Lei Song, and Xu- long Luo. “Device-agnostic log anomaly classification with partial labels”. In:2018 IEEE/ACM 26th International Symposium on Quality of Service (IWQoS). IEEE. 2018

    2018

  134. [142]

    Loglg: Weakly supervised log anomaly detection via log-event graph construction

    Hongcheng Guo, Yuhui Guo, Jian Yang, Jiaheng Liu, Zhoujun Li, Tieqiao Zheng, Liangfan Zheng, Weichao Hou, and Bo Zhang. “Loglg: Weakly supervised log anomaly detection via log-event graph construction”. In:International Conference on Database Systems for Advanced Applications....

    2023

  135. [143]

    A weakly supervised anomaly detection method based on deep anomaly scoring network

    Xin Xie, Zixi Li, Yuhui Huang, and Dengquan Wu. “A weakly supervised anomaly detection method based on deep anomaly scoring network”. In:Signal, Image and Video Processing17.8 (2023), pp. 3903–3911

    2023

  136. [144]

    Weakly Supervised Anomaly Detection via Knowledge-Data Alignment

    Haihong Zhao, Chenyi Zi, Yang Liu, Chen Zhang, Yan Zhou, and Jia Li. “Weakly Supervised Anomaly Detection via Knowledge-Data Alignment”. In:Proceedings of the ACM on Web Conference 2024. 2024, pp. 4083–4094

    2024

  137. [145]

    Deep weakly-supervised anomaly detection

    Guansong Pang, Chunhua Shen, Huidong Jin, and Anton van den Hengel. “Deep weakly-supervised anomaly detection”. In:Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining. 2023, pp. 1795–1807

    2023

  138. [146]

    Data programming: Creating large training sets, quickly

    Alexander J Ratner, Christopher M De Sa, Sen Wu, Daniel Selsam, and Christo- pher Re. “Data programming: Creating large training sets, quickly”. In:Advances in neural information processing systems29 (2016)

    2016

  139. [147]

    Isolation forest

    Fei Tony Liu, Kai Ming Ting, and Zhi-Hua Zhou. “Isolation forest”. In:2008 eighth ieee international conference on data mining. IEEE. 2008

    2008

  140. [148]

    Iso- lation forest based anomaly detection: A systematic literature review

    Wahid Salman Al Farizi, Indriana Hidayah, and Muhammad Nur Rizal. “Iso- lation forest based anomaly detection: A systematic literature review”. In:2021 8th International Conference on Information Technology, Computer and Electrical Engineering (ICITACEE). IEEE. 2021, pp. 118–122

    2021

  141. [149]

    LogBD: A Log Anomaly Detec- tion Method Based on Pretrained Models and Domain Adaptation

    Shuxian Liu, Le Deng, Huan Xu, and Wei Wang. “LogBD: A Log Anomaly Detec- tion Method Based on Pretrained Models and Domain Adaptation”. In:Applied Sciences 13.13 (2023), p. 7739

    2023

  142. [150]

    Learning from positive and unlabeled examples

    Fabien Letouzey, François Denis, and Rémi Gilleron. “Learning from positive and unlabeled examples”. In:International Conference on Algorithmic Learning The- ory. Springer. 2000, pp. 71–85

    2000

  143. [151]

    Detecting positive and negative deceptive opinions using PU- learning

    Donato Hernández Fusilier, Manuel Montes-y-Gómez, Paolo Rosso, and Rafael Guzmán-Cabrera. “Detecting positive and negative deceptive opinions using PU- learning”. In:Information Process Management51.4 (2015), pp. 433–443

    2015

  144. [152]

    A bagging SVM to learn from positive and unlabeled examples

    Fantine Mordelet and Jean-Philippe Vert. “A bagging SVM to learn from positive and unlabeled examples”. In:Pattern Recognition Letters37 (2014), pp. 201–209. BIBLIOGRAPHY 147

    2014

  145. [153]

    hPSD: A Hybrid PU-Learning-Based Spammer Detection Model for Product Reviews

    Zhiang Wu, Jie Cao, Yaqiong Wang, Youquan Wang, Lu Zhang, and Junjie Wu. “hPSD: A Hybrid PU-Learning-Based Spammer Detection Model for Product Reviews”. In:IEEE Trans. Cybern.50.4 (2020), pp. 1595–1606

    2020

  146. [154]

    PULearninginPayload- based Web Anomaly Detection

    YuxuanLuo,ShaoyinCheng,ChongLiu,andFanJiang.“PULearninginPayload- based Web Anomaly Detection”. In:Third International Conference on Security of Smart Cities, Industrial Control System and Communications, SSIC 2018, Shang- hai, China, October 18-19, 2018. IEEE, 2018, pp. 1–5

    2018

  147. [155]

    What Supercomputers Say: A Study of Five System Logs

    Adam Oliner and Jon Stearley. “What Supercomputers Say: A Study of Five System Logs”. In:DSN. 2007

    2007

  148. [156]

    Automated log audits for privacy compliance validation: a literature survey

    Jenni Reuben, Leonardo A Martucci, and Simone Fischer-Hübner. “Automated log audits for privacy compliance validation: a literature survey”. In:Privacy and Identity Management. Time for a Revolution? 10th IFIP WG 9.2, 9.5, 9.6/11.7, 11.4, 11.6/SIG 9.2. 2 International Summer S...

    2016

  149. [157]

    Hands-onAIOps

    NavinSabharwalandGBhardwaj.“Hands-onAIOps”.In: Apress eBooks, Springer (2022)

    2022

  150. [158]

    Monitor- ing and mitigating software aging on IBM cloud controller system

    Harish Sukhwani, Rivalino Matias, Kishor S Trivedi, and Andy Rindos. “Monitor- ing and mitigating software aging on IBM cloud controller system”. In:ISSREW. IEEE. 2017

    2017

  151. [159]

    Condi- tional Anomaly Detection

    Xiuyao Song, Mingxi Wu, Christopher M. Jermaine, and Sanjay Ranka. “Condi- tional Anomaly Detection”. In:IEEE Trans. Knowl. Data Eng.19.5 (2007)

    2007

  152. [160]

    Pretrained transformers for text ranking: BERT and beyond

    Andrew Yates, Rodrigo Nogueira, and Jimmy Lin. “Pretrained transformers for text ranking: BERT and beyond”. In:Proceedings of the 14th ACM International Conference on web search and data mining. 2021, pp. 1154–1156

    2021

  153. [161]

    LogRCA: Log-based Root Cause Analysis for Distributed Services

    Thorsten Wittkopp, Philipp Wiesner, and Odej Kao. “LogRCA: Log-based Root Cause Analysis for Distributed Services”. In: European Conference on Parallel Processing. Springer. 2024

    2024

  154. [162]

    Semi-Autonomous Labeling of Unstructured Maintenance Log Data for Diagnostic Root Cause Anal- ysis

    MichaelESharp,ThurstonBSexton,andMichaelPBrundage.“Semi-Autonomous Labeling of Unstructured Maintenance Log Data for Diagnostic Root Cause Anal- ysis”. In:International Conference Advances in Production Management Systems (APMS). 2016

    2016

  155. [163]

    A Quantitative Causal Analysis for Network Log Data

    Richard Jarry, Satoru Kobayashi, and Kensuke Fukuda. “A Quantitative Causal Analysis for Network Log Data”. In: IEEE 45th Annual Computers, Software, and Applications Conference, COMPSAC 2021, Madrid, Spain, July 12-16, 2021. IEEE, 2021, pp. 1437–1442. 148 BIBLIOGRAPHY

    2021

  156. [164]

    Progressing from Anomaly Detection to Automated Log Labeling and Pioneering Root Cause Analysis

    Thorsten Wittkopp, Alexander Acker, and Odej Kao. “Progressing from Anomaly Detection to Automated Log Labeling and Pioneering Root Cause Analysis”. In: IEEE International Conference on Data Mining, ICDM 2023 - Workshops, Shang- hai, China, December 4, 2023. IEEE, 2023, pp. 1231–1239

    2023

  157. [165]

    Survey on deep learning with class imbalance

    Justin Johnson and Taghi Khoshgoftaar. “Survey on deep learning with class imbalance”. In:Journal of Big Data6 (Mar. 2019)

    2019

  158. [166]

    Automaticclusteringusingnature- inspired metaheuristics: A survey

    AdanJose-GarciaandWilfridoGomez-Flores.“Automaticclusteringusingnature- inspired metaheuristics: A survey”. In:Applied Soft Computing41 (2016), pp. 192– 213

    2016

  159. [167]

    Probabilistic Outputs for Support Vector Machines and Comparisons to Regularized Likelihood Methods

    John Platt. “Probabilistic Outputs for Support Vector Machines and Comparisons to Regularized Likelihood Methods”. In:Advances in Large-Margin Classifiers10 (1999)

    1999

  160. [168]

    On Calibration of Modern Neural Networks

    Chuan Guo, Geoff Pleiss, Yu Sun, and Kilian Q. Weinberger. “On Calibration of Modern Neural Networks”. In:International Conference on Machine Learning (ICML). 2017

    2017

  161. [169]

    BIRCH: an efficient data clustering method for very large databases

    Tian Zhang, Raghu Ramakrishnan, and Miron Livny. “BIRCH: an efficient data clustering method for very large databases”. In:ACM sigmod record25.2 (1996), pp. 103–114

    1996

  162. [1986]

    isbn: 978-1-4757-1906-2

    1906

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.