Pith. sign in

REVIEW 2 major objections 5 minor 80 references

Explainable AI for Comprehensive Risk Assessment for Financial Reports: A Lightweight Hierarchical Transformer Network Approach

T0 review · 2 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read TinyXRA, a lightweight hierarchical transformer, claims to rank U.S.

desk verdict Solid engineering, but the triplet loss trains the model to rank favorable firms as risky on two of four measures, so the headline claims do not survive contact with the paper's own definitions. read the letter →

arxiv 2506.23767 v2 pith:URYMMZPQ submitted 2025-06-30 q-fin.RM cs.LG

Xue Wen Tan , Stanley Kok This is my paper · ORCID
classification q-fin.RMcs.LG
keywords TinyXRAExplainableAIfinancialriskassessment10-Kreportshierarchicaltransformertripletrankinglossattentionwordcloudmeasures
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 sets out to show that a small, efficient transformer can take the Management's Discussion and Analysis (MD&A) section of a company's 10-K filing and rank the company's financial risk on four measures—standard deviation, skewness, kurtosis, and the Sortino ratio—rather than relying on volatility alone. The motivation is practical: financial institutions often lack the hardware for large language models and are required to explain their risk decisions, so a lightweight model with transparent attention is valuable if it can match or beat heavier approaches. The authors report that TinyXRA achieves top-two performance on F1, Spearman's rho, and Kendall's tau across seven test years (2018–2024) against XRR, TF-IDF, and two small LLMs, and that removing the words and sentences it attends most to degrades accuracy, which they read as evidence that the explanations are faithful. A central design choice is triplet ranking loss, which enforces that high-, medium-, and low-risk documents are separated with a margin, improving over pairwise ranking.

What carries the argument

The machinery is a hierarchical attention network: TinyBERT produces sentence embeddings and word-level attention from the [CLS] token, a bidirectional GRU adds sentence context, a sentence-level attention layer weights sentences, and the weighted sum becomes a document vector scored by a linear layer. A triplet ranking loss with margin 0.1 pulls high-risk scores above medium-risk scores and medium above low, using anchors sampled from the medium group. For explanations, a dynamic word cloud multiplies word-level and sentence-level attention, applies a per-document quantile threshold, and subtracts a token's maximum frequency in other risk groups so only group-distinctive terms remain.

What would settle it

Rerun the training and evaluation for skewness and Sortino with the risk order reversed in the triplet loss—treating low skewness and low Sortino as the high-risk end—and compare Spearman's rho and Kendall's tau with the paper's reported values; if the positive correlations collapse or change sign, the central accuracy claim for those two measures is an artifact of the label mapping.

Watch

Extended reading notes

Core claim

The central claim is that TinyXRA can replace volatility-only risk assessment with a single architecture that reads prose and predicts a richer risk profile, while staying small enough to run on an 11GB-VRAM GPU and producing explanations that survive quantitative scrutiny. The paper argues that interpretability does not cost accuracy: the attention weights that explain predictions are also the mechanism that lets the model outperform the prior state-of-the-art XRR, and the contextual embeddings from TinyBERT outperform XRR's static embeddings. It further claims that the triplet loss captures both direction and magnitude of risk differences, giving a real but modest gain over pairwise loss, and that the attention word clouds, built from sentence-level and word-level attention with group-specific filtering, isolate risk vocabulary that aligns with economic conditions, such as 'deposits' and 'dividends' appearing in low-risk word clouds during a rising-rate year.

Load-bearing premise

The weakest load-bearing premise is that a higher bin always means higher risk when building triplets, but for skewness and the Sortino ratio a higher value is better for investors, so the model may be trained to rank safer firms as riskier on two of the four tasks.

Editorial extensions

If this is right

  • Financial firms with a consumer-grade GPU could screen thousands of 10-K filings in real time, because the encoder is TinyBERT and the design runs within 11GB VRAM.
  • Risk assessment can move beyond volatility: the same model outputs rankings on downside-focused measures such as the Sortino ratio and distribution-shape measures such as skewness and kurtosis.
  • Attention-based explanations can be treated as faithful: removal experiments show that deleting the most-attended words or sentences degrades F1, Spearman, and Kendall consistently across years.
  • Triplet ranking loss is preferable to pairwise and cross-entropy losses for ordinal risk bins, improving F1 with small but non-negligible effect sizes.
  • The seven-year rolling-origin results indicate the model generalizes across different market regimes, including the COVID-19 shock in 2020 and the post-COVID recovery.

Reading between the lines

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

  • The label-ordering choice for skewness and Sortino is a testable extension: re-running with the risk order reversed would show whether the positive correlations reported for those two measures survive.
  • The word clouds could be turned into an early-warning monitor: year-over-year shifts in group-distinctive terms might flag emerging systemic risks before balance-sheet data reflect them.
  • Since the architecture is not finance-specific beyond its labels, the same lightweight hierarchical design could be applied to other long regulatory filings, such as 10-Q reports, with minimal adaptation.
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

2 major / 5 minor

Summary. The paper proposes TinyXRA, a lightweight hierarchical transformer for financial risk assessment from the MD&A sections of 10-K reports. Risk labels are derived from four stock-return measures (standard deviation, skewness, kurtosis, Sortino ratio), discretized into three percentile bins. The model uses TinyBERT sentence embeddings, a bidirectional GRU with sentence-level attention, and a triplet ranking loss, and it provides explanations through attention-based word clouds and heatmaps. The authors report experiments on a rolling-origin design over 2018–2024 and claim state-of-the-art predictive accuracy across all risk measures and years, together with faithfulness evaluations of the attention explanations.

Significance. If the claims were supported, TinyXRA would be a practically useful contribution: a lightweight, explainable model for financial risk ranking with competitive accuracy and transparent word- and sentence-level explanations. The paper includes a broad experimental setup, seven test years, multiple baselines, and public code. However, the central SOTA claim is contradicted by the paper's own tables, where TF-IDF and sometimes XRR outperform TinyXRA on several metrics and years. In addition, the triplet loss and evaluation metrics are oriented inconsistently with the risk semantics for skewness and the Sortino ratio, as acknowledged in Section 7. These issues undermine the two main contributions claimed in the abstract and introduction.

major comments (2)
  1. [§5.3.2, Tables 3–6, Table 9] The claim that TinyXRA achieves state-of-the-art predictive accuracy across seven test years is contradicted by the paper's own results. For instance, TF-IDF beats TinyXRA on F1 for standard deviation in 2021 (75.4 vs 74.1) and 2022 (75.8 vs 75.1) in Table 3; on skewness F1 in 2024 (46.3 vs 44.4) and 2018 (49.1 vs 46.0) in Table 4; and on kurtosis Spearman in 2023 (38.0 vs 35.2) and 2018 (39.4 vs 39.0) in Table 5. Table 6 shows TF-IDF winning several Sortino cells as well. The paper itself acknowledges in §5.3.2 that TF-IDF 'performs remarkably well, particularly under the skewness and kurtosis measures, consistently ranking close to or better than more advanced neural models.' The consolidated Cohen's d values in Table 9 aggregate over all risk measures and years, which masks these per-cell losses and cannot support the unqualified statement that TinyXRA 'consistently demonstrates superior predictive performance' (§5.3.2) or the abstract's 'state-of-the-art' claim. This is a load-bearing discrepancy between the stated contribution and the reported evidence.
  2. [§4.4.2, Eq. (19); §3.3; §7] The triplet loss in Eq. (19) treats label 2 as the 'High Risk' positive sample and label 0 as the 'Low Risk' negative sample for all four risk measures. Section 7, however, states that for skewness and the Sortino ratio, a higher bin is generally more favorable to investors. Because the bins are defined as percentile ranges of the raw measure (§3.3), label 2 corresponds to the lowest-risk firms for these two measures. Training with the triplet loss as written therefore forces the model to assign higher logits to safer firms on two of the four tasks. The Spearman and Kendall correlations in Tables 4 and 6 are computed against the numeric labels 0, 1, 2 without reversing the order for skewness and Sortino, so a positive correlation indicates agreement with the raw measure's percentile, not with financial risk. This inverts the risk semantics for half of the 'comprehensive' assessment and invalidates the favorable results reported for skewness and Sortino. The authors need to either reverse the label order for these measures in the loss and evaluation, or provide a clear justification for treating higher skewness and higher Sortino as higher risk, which would contradict the paper's own Section 7.
minor comments (5)
  1. [§6.3, Appendix C] The faithfulness test removes the top-k% attended words/sentences and measures performance drop. Because the same attention scores are used both to select the removed elements and to compute the document representation, the test is partially self-referential. Adding a control condition that removes the lowest-attended elements (or random elements) would strengthen the claim that the observed degradation is due to the importance of the attended content rather than simply the reduction of input length.
  2. [§3.3, Table 1] The paper describes the task as 'risk quartile classification' in the abstract and as 'quartiles' in §4.4, but the actual binning is three percentile groups (0–30%, 30–70%, 70–100%). Please use consistent terminology, such as 'terciles' or 'percentile bins', throughout.
  3. [§5.1.1, §5.3.2] The description of the F1 metric in §5.1.1 says Micro F1 is equivalent to accuracy, but the paper only reports Macro F1. Please clarify how the predicted logits are converted to discrete bins for F1 computation, including the exact thresholds used, since this affects the interpretation of the reported F1 values.
  4. [Appendix E, Table 10] The hyperparameter table lists 'Qwen2.5-0.5B' while the main text and experiment descriptions use 'Qwen2-0.5B'. Please reconcile the naming to avoid confusion about which model was actually evaluated.
  5. [§8.3] The limitations section candidly notes that earlier years are affected by missing data for delisted or merged companies and by the use of the latest CIK-to-ticker mapping. These issues likely bias the earlier test years, yet the experimental analysis in §5.3.2 does not address or quantify this potential bias. A sensitivity analysis or explicit discussion of how the data gaps affect the 2018–2020 results would be helpful.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: risk labels are externally derived from stock returns and predictions are evaluated on held-out years, so the accuracy claim is self-contained.

full rationale

The paper's central claim is that TinyXRA ranks companies by financial risk better than benchmarks across seven test years. The labels are not produced by the model: they are computed from external post-event Fama-French residual returns (Section 3.2) and binned into percentile groups (Table 1), and evaluation uses a rolling-origin split in which each test year's labels are unseen during training (Table 2). The triplet ranking loss (Eq. 19) is a training objective defined over those external labels, not a source of labels. The reported Spearman's Rho and Kendall's Tau compare raw predicted logits to held-out ground-truth bins, so the accuracy numbers are empirical outcomes rather than identities. The attention-removal faithfulness test in Section 6.3 is an internal consistency check and is not a derivation of the risk-ranking result. Although Section 7 notes that for skewness and the Sortino ratio "a higher bin number is generally more favorable for investors," which creates a potential semantic mismatch in how the bins are labeled as risk, this is a validity or direction-of-benefit concern, not a circularity: it does not make the model's prediction equal to its input by construction. No load-bearing argument relies on a self-citation; the principal baseline XRR (Lin et al. 2021) is an independent prior work, not by the present authors. Therefore the paper shows no significant circularity.

Assumptions & free parameters 5 free parameters · 5 assumptions · 0 invented entities

The central claim depends on several hand-chosen hyperparameters (margin, truncation lengths, quantile cutoffs) and on domain assumptions about the predictive content of MD&A text and the temporal integrity of the label windows. The most consequential assumption, the uniform risk ordering across all four measures, is contradicted by the paper's own statements about skewness and Sortino, making it an ad hoc assumption that breaks the training objective.

free parameters (5)
  • triplet margin delta = 0.1
    Hand-chosen margin in Eq. 19; no sensitivity analysis is reported, and the central ranking claim could depend on this value.
  • word cloud selection quantile K = 30%
    The (1-K) quantile threshold for token selection in Eq. 21 is set to 30% by default; changing it changes the word-cloud content and the faithfulness results.
  • sentence length cutoff l = 40 tokens
    Documents are split into sentences truncated at 40 tokens (Section 4.1.1); this truncation could discard risk-relevant content.
  • document length cutoff L = 350 sentences
    Each document is truncated to 350 sentences (Section 4.1.1); longer MD&A sections lose content, which may affect both performance and explanations.
  • percentile bins for labels = 30th and 70th percentiles
    Risk values are binned into 0-30%, 30-70%, 70-100% (Section 3.3); the choice of cutoffs is taken from prior credit-risk practice and is not varied.
assumptions (5)
  • domain assumption Fama-French three-factor model residuals capture firm-specific risk relevant to future returns.
    Eq. 2's residuals are used to compute all four risk labels; if the model is misspecified, the labels themselves are biased. (Section 3.2)
  • domain assumption The MD&A text of a 10-K filing contains information that predicts the four return-based risk statistics computed over the following 252 trading days.
    This is the core learnability assumption; the paper provides no theoretical justification. (Section 3.1, Section 5)
  • ad hoc to paper A single ordinal mapping from risk level to label (0=low, 2=high) applies to all four risk measures.
    Eq. 19 assumes label 2 is always 'High Risk', but Section 7 states higher bins are favorable for skewness and Sortino, making the mapping wrong for those measures. This is effectively assumed without verification.
  • domain assumption The rolling-origin split keeps training and test labels temporally disjoint.
    Table 2 splits by filing year, but labels use 252 days of post-filing returns, so training labels for late-year filings can extend into the test-year window. The paper assumes no leakage but does not verify the label windows.
  • domain assumption Missing price data for delisted and merged companies does not systematically bias the risk labels.
    Section 8.3 states that Yahoo Finance lacks historical prices for some companies and the CIK-to-ticker map is current only, which the authors acknowledge may bias earlier years.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Explainable AI for Comprehensive Risk Assessment for Financial Reports: A Lightweight Hierarchical Transformer Network Approach." pith.science (2026). https://pith.science/paper/URYMMZPQ

@misc{pith2026250623767,
  author       = {Pith},
  title        = {Pith review of: Explainable AI for Comprehensive Risk Assessment for Financial Reports: A Lightweight Hierarchical Transformer Network Approach},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/URYMMZPQ}},
  note         = {Machine review of arXiv:2506.23767}
}
read the original abstract

Every publicly traded U.S. company files an annual 10-K report containing critical insights into financial health and risk. We propose Tiny eXplainable Risk Assessor (TinyXRA), a lightweight and explainable transformer-based model that automatically assesses company risk from these reports. Unlike prior work that relies solely on the standard deviation of excess returns (adjusted for the Fama-French model), which indiscriminately penalizes both upside and downside risk, TinyXRA incorporates skewness, kurtosis, and the Sortino ratio for more comprehensive risk assessment. We leverage TinyBERT as our encoder to efficiently process lengthy financial documents, coupled with a novel dynamic, attention-based word cloud mechanism that provides intuitive risk visualization while filtering irrelevant terms. This lightweight design ensures scalable deployment across diverse computing environments with real-time processing capabilities for thousands of financial documents which is essential for production systems with constrained computational resources. We employ triplet loss for risk quartile classification, improving over pairwise loss approaches in existing literature by capturing both the direction and magnitude of risk differences. Our TinyXRA achieves state-of-the-art predictive accuracy across seven test years on a dataset spanning 2013-2024, while providing transparent and interpretable risk assessments. We conduct comprehensive ablation studies to evaluate our contributions and assess model explanations both quantitatively by systematically removing highly attended words and sentences, and qualitatively by examining explanation coherence. The paper concludes with findings, practical implications, limitations, and future research directions. Our code is available at https://github.com/Chen-XueWen/TinyXRA.

Figures

Figures reproduced from arXiv: 2506.23767 by the authors.

Figure 5
Figure 5. [PITH_FULL_IMAGE:figures/full_fig_p039_5.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

80 extracted references · 69 canonical work pages

  1. [1]

    , " * write output.state after.block = add.period write newline

    ENTRY address author booktitle chapter doi edition editor eid howpublished institution isbn issn journal key month note number organization pages publisher school series title type url volume year label extra.label sort.label short.list INTEGERS output.state before.all mid.sentence after.sentence after.block FUNCTION init.state.consts #0 'before.all := #1...

  2. [2]

    write newline

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

  3. [3]

    Aikman D, Alessandri P, Eklund B, Gai P, Kapadia S, Martin E, Mora N, Sterne G, Willison M (2011) Funding Liquidity Risk in a Quantitative Model of Systemic Stability . Alfaro R, ed., Financial Stability, Monetary Policy, and Central Banking , volume 15 of Central Banking, Analysis, and Economic Policies Book Series, chapter 12, 371--410 (Central Bank of Chile)

    work page 2011

  4. [4]

    Akhtar MS, Kumar A, Ghosal D, Ekbal A, Bhattacharyya P (2017) A multilayer perceptron based ensemble technique for fine-grained financial sentiment analysis. Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing, 540--546 (Copenhagen, Denmark: Association for Computational Linguistics), ://dx.doi.org/10.18653/v1/D17-1057

    work page doi:10.18653/v1/d17-1057 2017

  5. [5]

    Alemania FFSA (2018) Big Data Meets Artificial Intelligence: Challenges and Implications for the Supervision and Regulation of Financial Services (Federal Financial Supervision Authority), ://books.google.com.sg/books?id=QU9RzwEACAAJ

    work page 2018

  6. [6]

    The Journal of Finance 23(4):589--609, ://dx.doi.org/https://doi.org/10.1111/j.1540-6261.1968.tb00843.x

    Altman EI (1968) Financial ratios, discriminant analysis and the prediction of corporate bankruptcy. The Journal of Finance 23(4):589--609, ://dx.doi.org/https://doi.org/10.1111/j.1540-6261.1968.tb00843.x

    arXiv 1968

  7. [7]

    MIS Quarterly 30(1):79--97, ISSN 02767783

    Arnold V, Clark N, Collier PA, Leech SA, Sutton SG (2006) The differential use and effect of knowledge-based system explanations in novice and expert judgment decisions. MIS Quarterly 30(1):79--97, ISSN 02767783

    work page 2006

  8. [8]

    MIS Quarterly Executive 19(4):259--278

    Asatiani A, Malo P, Nagb l PR, Penttinen E, Rinta-Kahila T, Salovaara A (2020) Challenges of explaining the behavior of black-box ai systems. MIS Quarterly Executive 19(4):259--278

    work page 2020

Show all 80 references
  1. [9]

    European Journal of Operational Research 202(3):789--801, ISSN 0377-2217, ://dx.doi.org/https://doi.org/10.1016/j.ejor.2009.06.023

    Balakrishnan R, Qiu XY, Srinivasan P (2010) On the predictive ability of narrative disclosures in annual reports. European Journal of Operational Research 202(3):789--801, ISSN 0377-2217, ://dx.doi.org/https://doi.org/10.1016/j.ejor.2009.06.023

    2010 doi

  2. [10]

    Management Science 67(7):4004--4031

    Bellstam G, Bhagat S, Cookson JA (2021) A text-based analysis of corporate innovation. Management Science 67(7):4004--4031

    2021

  3. [11]

    American Economic Review 97(2):159--165

    Brunnermeier MK, Gollier C, Parker JA (2007) Optimal beliefs, asset prices, and the preference for skewed returns. American Economic Review 97(2):159--165

    2007

  4. [12]

    Computational Economics 57:203--216, ://dx.doi.org/10.1007/s10614-020-10042-0

    Bussmann N, Giudici P, Marinelli D, McNeil AJ (2021) Explainable machine learning in credit risk management. Computational Economics 57:203--216, ://dx.doi.org/10.1007/s10614-020-10042-0

    2021 doi

  5. [13]

    Journal of Financial Research 39(2):145--178

    Chiang IHE (2016) Skewness and coskewness in bond returns. Journal of Financial Research 39(2):145--178

    2016

  6. [14]

    2005 IEEE computer society conference on computer vision and pattern recognition (CVPR'05), volume 1, 539--546 (IEEE)

    Chopra S, Hadsell R, LeCun Y (2005) Learning a similarity metric discriminatively, with application to face verification. 2005 IEEE computer society conference on computer vision and pattern recognition (CVPR'05), volume 1, 539--546 (IEEE)

    2005

  7. [15]

    Christoffersen PF, Diebold FX (2000) How relevant is volatility forecasting for financial risk management? Review of Economics and Statistics 82(1):12--22

    2000

  8. [16]

    NIPS 2014 Workshop on Deep Learning, December 2014

    Chung J, Gulcehre C, Cho K, Bengio Y (2014) Empirical evaluation of gated recurrent neural networks on sequence modeling. NIPS 2014 Workshop on Deep Learning, December 2014

    2014

  9. [17]

    Optimization Challenges in Complex, Networked and Risky Systems, 269--281 (INFORMS)

    Das AS, Gupta A, Singh G, Subramaniam LV (2016) Mining qualitative attributes to assess corporate performance. Optimization Challenges in Complex, Networked and Risky Systems, 269--281 (INFORMS)

    2016

  10. [18]

    Contemporary Accounting Research 29(3):804--837

    Davis AK, Tama-Sweet I (2012) Managers’ use of language across alternative disclosure outlets: earnings press releases versus md&a. Contemporary Accounting Research 29(3):804--837

    2012

  11. [19]

    Physica A: Statistical Mechanics and Its Applications 622:128886

    De Domenico F, Livan G, Montagna G, Nicrosini O (2023) Modeling and simulation of financial returns under non-gaussian distributions. Physica A: Statistical Mechanics and Its Applications 622:128886

    2023

  12. [20]

    NAACL-HLT (1) , 4171--4186 (Association for Computational Linguistics)

    Devlin J, Chang M, Lee K, Toutanova K (2019) BERT: pre-training of deep bidirectional transformers for language understanding. NAACL-HLT (1) , 4171--4186 (Association for Computational Linguistics)

    2019

  13. [21]

    International Joint Conference on Artificial Intelligence

    Ding X, Zhang Y, Liu T, Duan J (2015) Deep learning for event-driven stock prediction. International Joint Conference on Artificial Intelligence

    2015

  14. [22]

    2018 41st International Convention on Information and Communication Technology, Electronics and Microelectronics (MIPRO), 0210--0215, ://dx.doi.org/10.23919/MIPRO.2018.8400040

    Došilović FK, Brčić M, Hlupić N (2018) Explainable artificial intelligence: A survey. 2018 41st International Convention on Information and Communication Technology, Electronics and Microelectronics (MIPRO), 0210--0215, ://dx.doi.org/10.23919/MIPRO.2018.8400040

    2018

  15. [23]

    ACM Computing Surveys 56(9):1--42

    Du K, Xing F, Mao R, Cambria E (2024) Financial sentiment analysis: Techniques and applications. ACM Computing Surveys 56(9):1--42

    2024

  16. [24]

    Journal of Accounting and Economics 64(2):221--245, ISSN 0165-4101, ://dx.doi.org/https://doi.org/10.1016/j.jacceco.2017.07.002

    Dyer T, Lang M, Stice-Lawrence L (2017) The evolution of 10-k textual disclosure: Evidence from latent dirichlet allocation. Journal of Accounting and Economics 64(2):221--245, ISSN 0165-4101, ://dx.doi.org/https://doi.org/10.1016/j.jacceco.2017.07.002

    2017 doi

  17. [25]

    The European Journal of Finance 6(1):18--38

    Eftekhari B, Pedersen CS, Satchell SE (2000) On the volatility of measures of financial risk: an investigation using returns from european markets. The European Journal of Finance 6(1):18--38

    2000

  18. [26]

    Journal of Financial Economics 33(1):3--56, ISSN 0304-405X, ://dx.doi.org/https://doi.org/10.1016/0304-405X(93)90023-5

    Fama EF, French KR (1993) Common risk factors in the returns on stocks and bonds. Journal of Financial Economics 33(1):3--56, ISSN 0304-405X, ://dx.doi.org/https://doi.org/10.1016/0304-405X(93)90023-5

    1993 doi

  19. [27]

    Frontiers in artificial intelligence 5:779799

    Fritz-Morgenthal S, Hein B, Papenbrock J (2022) Financial risk management and explainable, trustworthy, responsible ai. Frontiers in artificial intelligence 5:779799

    2022

  20. [28]

    Comparative Economic Studies 53:511--534

    Gabrisch H, Orlowski LT (2011) Extreme risks in financial markets and monetary policies of the euro-candidates. Comparative Economic Studies 53:511--534

    2011

  21. [29]

    ACM Computing Surveys (CSUR) 55(1):1--30

    Ge W, Lalbakhsh P, Isai L, Lenskiy A, Suominen H (2022) Neural network--based financial volatility forecasting: A systematic review. ACM Computing Surveys (CSUR) 55(1):1--30

    2022

  22. [30]

    Frontiers in Artificial Intelligence 4:752558

    Gramegna A, Giudici P (2021) Shap and lime: an evaluation of discriminative power in credit risk. Frontiers in Artificial Intelligence 4:752558

    2021

  23. [31]

    (2024) The llama 3 herd of models

    Grattafiori A, Dubey A, Jauhri A, Pandey A, Kadian A, Al-Dahle A, Letman A, Mathur A, Schelten A, Vaughan A, et al. (2024) The llama 3 herd of models. arXiv preprint arXiv:2407.21783

    2024 arXiv

  24. [32]

    Neural Computing and Applications 29:343--358

    H \'a jek P (2018) Combining bag-of-words and sentiment features of annual reports to predict abnormal stock returns. Neural Computing and Applications 29:343--358

    2018

  25. [33]

    Similarity-based pattern recognition: third international workshop, SIMBAD 2015, Copenhagen, Denmark, October 12-14, 2015

    Hoffer E, Ailon N (2015) Deep metric learning using triplet network. Similarity-based pattern recognition: third international workshop, SIMBAD 2015, Copenhagen, Denmark, October 12-14, 2015. Proceedings 3, 84--92 (Springer)

    2015

  26. [34]

    Contemporary Accounting Research 40(2):806--841

    Huang AH, Wang H, Yang Y (2023) Finbert: A large language model for extracting information from financial text. Contemporary Accounting Research 40(2):806--841

    2023

  27. [35]

    The Journal of Finance 53(3):1111--1130

    Ito T, Lyons RK, Melvin MT (1998) Is there private information in the fx market? the tokyo experiment. The Journal of Finance 53(3):1111--1130

    1998

  28. [36]

    Findings of the Association for Computational Linguistics: EMNLP 2020, 4163--4174

    Jiao X, Yin Y, Shang L, Jiang X, Chen X, Li L, Wang F, Liu Q (2020) Tinybert: Distilling bert for natural language understanding. Findings of the Association for Computational Linguistics: EMNLP 2020, 4163--4174

    2020

  29. [37]

    Information Systems Research 20(4):527--546, ISSN 10477047, 15265536

    Kayande U, Bruyn AD, Lilien GL, Rangaswamy A, van Bruggen GH (2009) How incorporating feedback mechanisms in a dss affects dss evaluations. Information Systems Research 20(4):527--546, ISSN 10477047, 15265536

    2009

  30. [38]

    Biometrika 30(1/2):81--93, ISSN 00063444

    Kendall MG (1938) A new measure of rank correlation. Biometrika 30(1/2):81--93, ISSN 00063444

    1938

  31. [39]

    Biometrika 33(3):239--251

    Kendall MG (1945) The treatment of ties in ranking problems. Biometrika 33(3):239--251

    1945

  32. [40]

    Kogan S, Levin D, Routledge BR, Sagi JS, Smith NA (2009) Predicting risk from financial reports with regression. Proceedings of Human Language Technologies: The 2009 Annual Conference of the North American Chapter of the Association for Computational Linguistics, 272–280, NAAC...

    2009

  33. [41]

    Proceedings of the IEEE 86(11):2278--2324, ://dx.doi.org/10.1109/5.726791

    Lecun Y, Bottou L, Bengio Y, Haffner P (1998) Gradient-based learning applied to document recognition. Proceedings of the IEEE 86(11):2278--2324, ://dx.doi.org/10.1109/5.726791

    1998 doi

  34. [42]

    Journal of Accounting Research 48(5):1049--1102, ://dx.doi.org/https://doi.org/10.1111/j.1475-679X.2010.00382.x

    Li F (2010 a ) The information content of forward-looking statements in corporate filings a naïve bayesian machine learning approach. Journal of Accounting Research 48(5):1049--1102, ://dx.doi.org/https://doi.org/10.1111/j.1475-679X.2010.00382.x

    2010 arXiv

  35. [43]

    Journal of accounting research 48(5):1049--1102

    Li F (2010 b ) The information content of forward-looking statements in corporate filings—a na \" ve bayesian machine learning approach. Journal of accounting research 48(5):1049--1102

    2010

  36. [44]

    Machine Learning and Knowledge Discovery in Databases

    Lin TW, Sun RY, Chang HL, Wang CJ, Tsai MF (2021) Xrr: Explainable risk ranking for financial reports. Machine Learning and Knowledge Discovery in Databases. Applied Data Science Track: European Conference, ECML PKDD 2021, Bilbao, Spain, September 13–17, 2021, Proceedings, Par...

    2021 doi

  37. [45]

    Entropy 23(1), ISSN 1099-4300

    Linardatos P, Papastefanopoulos V, Kotsiantis S (2021) Explainable ai: A review of machine learning interpretability methods. Entropy 23(1), ISSN 1099-4300

    2021

  38. [46]

    Management Science 68(2):1573--1589

    Liu Y (2022) The short-run and long-run components of idiosyncratic volatility and stock returns. Management Science 68(2):1573--1589

    2022

  39. [47]

    The Journal of finance 66(1):35--65

    Loughran T, McDonald B (2011) When is a liability not a liability? textual analysis, dictionaries, and 10-ks. The Journal of finance 66(1):35--65

    2011

  40. [48]

    Advances in neural information processing systems 30

    Lundberg SM, Lee SI (2017) A unified approach to interpreting model predictions. Advances in neural information processing systems 30

    2017

  41. [49]

    Management Science Letters 4(4):659--662

    Makrani K, Zamanian B (2014) Ranking mutual funds using sortino method. Management Science Letters 4(4):659--662

    2014

  42. [50]

    Ieee Access 8:203203--203223

    Mashrur A, Luo W, Zaidi NA, Robles-Kelly A (2020) Machine learning for financial risk management: a survey. Ieee Access 8:203203--203223

    2020

  43. [51]

    Intelligent Computing: Proceedings of the 2019 Computing Conference, Volume 2, 1269--1292 (Springer)

    Mathews SM (2019) Explainable artificial intelligence applications in nlp, biomedical, and malware classification: a literature review. Intelligent Computing: Proceedings of the 2019 Computing Conference, Volume 2, 1269--1292 (Springer)

    2019

  44. [52]

    Information Systems Management 39(1):53--63, ://dx.doi.org/10.1080/10580530.2020.1849465

    Meske C, Bunde E, Schneider J, Gersch M (2022) Explainable artificial intelligence: Objectives, stakeholders, and future research opportunities. Information Systems Management 39(1):53--63, ://dx.doi.org/10.1080/10580530.2020.1849465

    2022

  45. [53]

    Mondello E (2023) Other Investment Characteristics, 63--99 (Wiesbaden: Springer Fachmedien Wiesbaden), ISBN 978-3-658-41021-6, ://dx.doi.org/10.1007/978-3-658-41021-6_3

    2023 doi

  46. [54]

    Management Science 61(5):931--948

    Muslu V, Radhakrishnan S, Subramanyam K, Lim D (2015) Forward-looking md&a disclosures and the information environment. Management Science 61(5):931--948

    2015

  47. [55]

    Myers J, Well A, Lorch R (2013) Research Design and Statistical Analysis: Third Edition (Taylor & Francis), ISBN 9781135811631

    2013

  48. [56]

    Nopp C, Hanbury A (2015) Detecting risks in the banking system by sentiment analysis. Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing, 591--600 (Lisbon, Portugal: Association for Computational Linguistics), ://dx.doi.org/10.18653/v1/D15-1071

    2015 doi

  49. [57]

    Proceedings of the 2014 conference on empirical methods in natural language processing (EMNLP), 1532--1543

    Pennington J, Socher R, Manning CD (2014) Glove: Global vectors for word representation. Proceedings of the 2014 conference on empirical methods in natural language processing (EMNLP), 1532--1543

    2014

  50. [58]

    Rajaraman A, Ullman JD (2011) Mining of massive datasets (Autoedicion)

    2011

  51. [59]

    Rekabsaz N, Lupu M, Baklanov A, D \"u r A, Andersson L, Hanbury A (2017) Volatility prediction using financial disclosures sentiments with word embedding-based IR models. Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Pa...

    2017 doi

  52. [60]

    Thirty Fourth International Conference on Information Systems, Milan

    Ren J, Ge H, Wu X, Wang G, Wang W, Liao S (2013) Effective sentiment analysis of corporate financial reports. Thirty Fourth International Conference on Information Systems, Milan

    2013

  53. [61]

    Ribeiro M, Singh S, Guestrin C (2016) `` why should I trust you? '' : Explaining the predictions of any classifier. Proceedings of the 2016 Conference of the North A merican Chapter of the Association for Computational Linguistics: Demonstrations , 97--101 (San Diego, Californ...

    2016 doi

  54. [62]

    Journal of Risk and Financial Management 14(11):540

    Samunderu E, Murahwa YT (2021) Return based risk measures for non-normally distributed returns: An alternative modelling approach. Journal of Risk and Financial Management 14(11):540

    2021

  55. [63]

    International Conference on Computational Linguistics

    Santos (2014) Deep convolutional neural networks for sentiment analysis of short texts. International Conference on Computational Linguistics

    2014

  56. [64]

    Neural Computing and Applications 34(17):14327--14339, ://dx.doi.org/10.1007/s00521-022-07472-2

    Shi S, Tse R, Luo W, Zhang X (2022) Machine learning-driven credit risk: a systemic review. Neural Computing and Applications 34(17):14327--14339, ://dx.doi.org/10.1007/s00521-022-07472-2

    2022 doi

  57. [65]

    MIS Quarterly Executive 21(2)

    Someh I, Wixom BH, Beath CM, Zutavern A (2022) Building an artificial intelligence explanation capability. MIS Quarterly Executive 21(2)

    2022

  58. [66]

    Sun X, Lu W (2020) Understanding attention for text classification. Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, 3418--3428 (Online: Association for Computational Linguistics), ://dx.doi.org/10.18653/v1/2020.acl-main.312

    2020 doi

  59. [67]

    International journal of forecasting 16(4):437--450

    Tashman LJ (2000) Out-of-sample tests of forecasting accuracy: an analysis and review. International journal of forecasting 16(4):437--450

    2000

  60. [68]

    Management Science 62(6):1598--1609

    Theodossiou P, Savva CS (2016) Skewness and the relation between risk and return. Management Science 62(6):1598--1609

    2016

  61. [69]

    ://dx.doi.org/10.1109/TNNLS.2020.3027314

    Tjoa E, Guan C (2020) A survey on explainable artificial intelligence (xai) : toward medical xai. ://dx.doi.org/10.1109/TNNLS.2020.3027314

    2020

  62. [70]

    Toma A, Dedu S (2014) Quantitative techniques for financial risk assessment: A comparative approach using different risk measures and estimation methods. Procedia Economics and Finance 8:712--719, ISSN 2212-5671, ://dx.doi.org/https://doi.org/10.1016/S2212-5671(14)00149-X, 1st...

    2014 doi

  63. [71]

    European Journal of Operational Research 257(1):243--250, ISSN 0377-2217, ://dx.doi.org/https://doi.org/10.1016/j.ejor.2016.06.069

    Tsai MF, Wang CJ (2017) On the risk prediction and analysis of soft information in finance reports. European Journal of Operational Research 257(1):243--250, ISSN 0377-2217, ://dx.doi.org/https://doi.org/10.1016/j.ejor.2016.06.069

    2017 doi

  64. [72]

    Advances in neural information processing systems 30

    Vaswani A, Shazeer N, Parmar N, Uszkoreit J, Jones L, Gomez AN, Kaiser , Polosukhin I (2017) Attention is all you need. Advances in neural information processing systems 30

    2017

  65. [73]

    Proceedings of the IEEE conference on computer vision and pattern recognition, 1386--1393

    Wang J, Song Y, Leung T, Rosenberg C, Wang J, Philbin J, Chen B, Wu Y (2014) Learning fine-grained image similarity with deep ranking. Proceedings of the IEEE conference on computer vision and pattern recognition, 1386--1393

    2014

  66. [74]

    Wiegreffe S, Pinter Y (2019) Attention is not not explanation. Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP), 11--20

    2019

  67. [75]

    (2024) Qwen2

    Yang A, Yang B, Zhang B, Hui B, Zheng B, Yu B, Li C, Liu D, Huang F, Wei H, et al. (2024) Qwen2. 5 technical report. arXiv preprint arXiv:2412.15115

    2024 arXiv

  68. [76]

    INFORMS Journal on Computing 34(1):522--540

    Yang Y, Zhang K, Fan Y (2022) Analyzing firm reports for volatility prediction: A knowledge-driven text-embedding approach. INFORMS Journal on Computing 34(1):522--540

    2022

  69. [77]

    Proceedings of the 2016 conference of the North American chapter of the association for computational linguistics: human language technologies, 1480--1489

    Yang Z, Yang D, Dyer C, He X, Smola A, Hovy E (2016) Hierarchical attention networks for document classification. Proceedings of the 2016 conference of the North American chapter of the association for computational linguistics: human language technologies, 1480--1489

    2016

  70. [78]

    Applied mathematical sciences 8:3229--3242, ://api.semanticscholar.org/CorpusID:42092623

    Zeng G (2014) A necessary condition for a good binning algorithm in credit scoring. Applied mathematical sciences 8:3229--3242, ://api.semanticscholar.org/CorpusID:42092623

    2014

  71. [79]

    MIS Quarterly Executive 19(4)

    Zhang Z, Hummel JT, Nandhakumar J, Waardenburg L (2020) Addressing the key challenges of developing machine learning ai systems for knowledge-intensive work. MIS Quarterly Executive 19(4)

    2020

  72. [80]

    ArXiv abs/2008.11825

    Zhao W, Joshi T, Nair VN, Sudjianto A (2020) Shap values for explaining cnn-based text classification models. ArXiv abs/2008.11825

    2020 arXiv

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.