Pith. sign in

REVIEW 4 major objections 6 minor 23 references

Decision-informed Neural Networks with Large Language Model Integration for Portfolio Optimization

T0 review · 4 major / 6 minor · reviewed 2026-08-09 · deepseek-v4-flash

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

Pith's one-line read Predicting returns accurately is not enough: DINN trains portfolio weights end-to-end with a decision-regret loss and beats prediction-only deep learning baselines.

desk verdict A genuinely new LLM+DFL architecture, but the headline results rest on a survivor-only backtest and the proofs have load-bearing errors; worth a serious redo, not a citation yet. read the letter →

arxiv 2502.00828 v1 pith:SNY7ONKL submitted 2025-02-02 q-fin.PM cs.AIq-fin.CP

Yoontae Hwang , Yaxuan Kong , Stefan Zohren , Yongjae Lee This is my paper · ORCID
classification q-fin.PMcs.AIq-fin.CP
keywords portfoliooptimizationdecision-focusedlearninglargelanguagemodelscross-attentiondifferentiablemean-varianceestimationerrorsurvivorshipbias
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 tries to establish that portfolio selection should be trained on decision quality, not forecast accuracy alone. It proposes DINN, a decision-informed neural network in which a cross-attention mechanism fuses LLM-derived embeddings of inter-asset and macroeconomic information with decomposed return series, and a differentiable optimization layer converts the resulting return forecasts directly into long-only portfolio weights. Training minimizes a hybrid loss that combines mean squared forecast error with decision regret, defined as the gap between the true mean-variance objective evaluated at predicted weights and at hindsight-optimal weights. On S&P 100 and Dow 30 data from 2010 to 2023, the paper reports that DINN beats Transformer- and LLM-based baselines on annualized return, Sharpe and Sortino ratios, return-over-VaR, and terminal wealth, with much lower trial-to-trial variability. A sympathetic reader would care because this is a direct test of the claim that prediction-error minimization is the wrong objective for investment decisions.

What carries the argument

The load-bearing mechanism is the differentiable portfolio-optimization layer coupled with a hybrid loss. Returns are first decomposed into long-term trend and short-term residual components using multi-scale averages; these are fed through prob-sparse cross-attention against two LLM-produced embedding sets, one encoding pairwise stock and sector outperformance counts, the other encoding macroeconomic statistics from irregularly sampled indicators. A frozen pretrained LLM, GPT-2, chosen to limit data leakage, refines the fused representations before a linear projection produces normalized return forecasts, which are denormalized and passed to a second-order cone optimization layer that outputs portfolio weights. The hybrid objective $\mathcal{L} = \beta \mathcal{L}_{\mathrm{MSE}} + (1-\beta)\mathcal{L}_{\mathrm{Decision}}$ with $\beta=0.4$ balances forecast accuracy against decision regret, and Theorems 1 and 2 provide the weight sensitivities needed to backpropagate through the optimization layer.

What would settle it

Re-run DINN and the baselines on the full constituent lists of both indices as of each rebalance date, carrying delisted companies' final returns through the backtest rather than deleting them; if the Sharpe ratio and terminal wealth gaps over the best baseline shrink or reverse, the central empirical claim fails.

Watch

Extended reading notes

Core claim

The central claim is that the two-stage predict-then-optimize workflow is why improved forecasts do not reliably improve portfolios, and that the fix is to differentiate through the portfolio optimization itself. DINN's optimization layer solves a second-order cone formulation of mean-variance optimization with full-investment and long-only constraints; because the layer is differentiable, gradients of the decision regret $\Delta J_{t+h}=\hat J_{t+h}-J^\star_{t+h}$ flow back through the predicted return and covariance estimates. Theorems 1 and 2 give closed-form sensitivities $\partial \hat w_{t+h}/\partial \hat\mu_{t+h}$ and $\partial \hat w_{t+h}/\partial \hat L_{t+h}$ for this propagation. Proposition 1 supplies a two-asset example in which predicted returns converge to the truth in squared error while the induced optimal portfolio weights do not converge to the true optimum, which is the paper's theoretical reason for rejecting MSE-only training. Empirically, the paper attributes DINN's advantage to its attention mechanism concentrating on assets whose misprediction would be most costly: across four macroeconomic regimes, assets with the largest gradient sensitivities show lower MSE and MAE than assets with small sensitivities, and portfolios built from attention-selected assets beat random and uniform portfolios on Sharpe ratio and drawdown.

Load-bearing premise

The empirical claim assumes in Section 4.1.1 that a backtest containing only companies which stayed in each index for the full 2010-2023 period is representative; because firms that were dropped or delisted are removed, the reported Sharpe ratios and terminal wealth could be inflated relative to a live trading experience.

Editorial extensions

If this is right

  • If the results hold, any forecasting model used for portfolio selection can be upgraded by appending a differentiable optimization layer and a decision-regret term without changing the underlying return predictor.
  • The gradient sensitivities $\partial \hat w/\partial \hat\mu$ and $\partial \hat w/\partial \hat L$ give a practical asset-ranking rule: concentrate forecasting capacity on assets where estimation error would move portfolio weights most.
  • LLM embeddings of macroeconomic variables give a route to include irregularly sampled economic indicators directly in the differentiable pipeline rather than as separately estimated factors.
  • The reported reduction in trial-to-trial variability suggests decision-focused training stabilizes portfolio construction even when return forecasts remain noisy.

Reading between the lines

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

  • A testable extension the paper leaves implicit: the same decision-regret construction could be applied to non-convex or non-differentiable objectives such as expected shortfall or drawdown-constrained portfolios, which the paper lists as future work rather than a demonstrated result.
  • Swapping the frozen GPT-2 backbone for a more capable LLM would separate two explanations: if DINN's edge comes mainly from decision-focused training, performance should change little, whereas if the embeddings matter, the gap should widen.
  • Because the backtest keeps only companies that stayed in each index for the whole 2010-2023 period, the reported terminal wealth figures are best interpreted as upper bounds; including delisted companies would be the natural stress test of the practical claim.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 6 minor

Summary. The manuscript proposes DINN, a decision-informed neural network that combines pretrained LLM embeddings with prob-sparse cross-attention, a differentiable mean-variance optimization layer, and a hybrid training objective mixing mean squared error with a decision-regret term. The authors claim that minimizing prediction error alone is suboptimal and that DINN consistently outperforms state-of-the-art deep learning baselines on the S&P 100 and DOW 30, with reported annualized returns of 43.5% and 63.3%, Sharpe ratios of 1.04 and 1.29, and terminal wealth of 3.02x and 4.47x. They also provide sensitivity theorems for the optimization layer, a theoretical proposition on the prediction-decision gap, and a gradient-based interpretability analysis of asset importance.

Significance. The combination of large language model embeddings with decision-focused learning for portfolio construction is timely and potentially impactful. The paper has several strengths: the end-to-end differentiable optimization with a regret-based loss is well motivated, the attention mechanism provides an interpretability route, and the authors explicitly consider data leakage by choosing GPT-2. However, the reported empirical superiority is not credible because the test universe is conditioned on future index membership, and the theoretical results contain load-bearing errors. As presented, the evidence does not support the paper's central claims.

major comments (4)
  1. [Section 4.1.1] The sentence 'To address potential survivorship bias, we include only companies that maintained consistent index membership throughout the study period' describes the opposite of a bias correction. Restricting the S&P 100 to 50 firms and the DOW 30 to 30 firms that survived from 2010 to 2023 conditions the backtest on future index membership, which is textbook survivorship bias: firms that were dropped, merged, or delisted, disproportionately poor performers, are excluded. Consequently the absolute levels in Table 1 (annualized returns 43.53% and 63.25%, terminal wealth 3.0213 and 4.4715) are not realizable strategy performance, and the relative ranking against baselines is contaminated because all models are evaluated on the same survivor set. A point-in-time constituent backtest with delisting returns is required before any outperformance claim can be assessed.
  2. [Section 3.4 / Appendix A.1, Theorem 1] The proof of Theorem 1 derives Eq. (43) after explicitly setting 2λ=1, so the sensitivity formula (30) is not valid for the general problem stated in Eq. (29); the correct derivative contains a factor 1/(2λ). More importantly, the implemented optimization problem in Eq. (21) includes the box constraints 0≤w_i≤1 and an inequality constraint ‖L w‖_2≤s, whereas Theorem 1 is proved only for an unconstrained-in-sign problem with an equality risk constraint. The KKT conditions for Eq. (21) are different, so Eq. (30) is not established for the actual optimization layer. Since Section 4.5 and Table 3 interpret gradient magnitudes |∂ŵ/∂μ| as evidence of decision-aware learning, this gap is load-bearing.
  3. [Proposition 1, Section 4.3] The constructed sequence is ˜μ^(k) = (μ_1 − δ + 1/k, μ_2) with δ = (μ_1 − μ_2)/2 > 0. As k→∞ this sequence converges to (μ_1 − δ, μ_2), not to the true mean μ = (μ_1, μ_2); the example therefore does not show that MSE-convergent predictions can induce non-convergent optimal portfolios. The subsequent computation of w^(k)_1 confirms that the limit is 1/2 + (μ_1 − μ_2)/(8λ), which differs from w*_1 only because the wrong limit was used. Thus Proposition 1, as stated and proved, is invalid.
  4. [Section 4.2 / Table 1] The reported returns, Sharpe ratios, and terminal wealth are computed on gross returns without transaction costs or turnover constraints. DINN rebalances at each forecast horizon and its optimization layer can produce high-turnover allocations; with realistic transaction costs and market impact, the 43.5% and 63.3% annualized returns could be materially reduced, and the ranking against baselines could change. The paper should report net-of-cost performance or at minimum turnover and a break-even cost analysis.
minor comments (6)
  1. [Section 3.3.4, Eq. (23)] The notation for the return window switches between ˆr_{t:t+H} and ˆr_{t+1:t+H} in Eqs. (23)-(25); please make the indexing consistent throughout.
  2. [Section 2.2] The reference 'Waswani et al. 2017' is a typo for 'Vaswani et al. 2017'.
  3. [Figure 3] The caption describes a comparison of portfolio standard deviation across trials, but the panels display Sharpe ratio, terminal wealth, maximum drawdown, and VaR; the caption and the figure content are inconsistent.
  4. [Table A.3] The caption repeats the formula |∂ŵ_{t+h}/∂μ_{t+h}| although the table reports gradients with respect to the Cholesky factor L_{t+h}; update the caption to match the table.
  5. [Section 4.1] The statement that code is available at 'Anonymous Github' does not provide a URL or repository identifier; without an accessible repository, the reproducibility claim cannot be verified.
  6. [Appendix A.4] The opening sentence refers to 'Appendix K', which appears to be a leftover placeholder and should be corrected.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the hybrid loss supervises against realized returns and ex-post optimal portfolios; self-citations are contextual, not load-bearing.

full rationale

The DINN derivation is self-contained. The training objective (Eq. 26) combines an MSE term (Eq. 23) on realized future returns with a regret term (Eq. 25) equal to |J_hat_{t+h} - J*_{t+h}|, where J* is evaluated at the actual return/covariance pair (mu*, Sigma*) and the ex-post optimal weights w* (Eqs. 3 and 24); this is a standard supervised target, not an output of the model folded back into itself. Predicted weights come from solving the SOCP in Eq. 21, and the gradient sensitivities in Theorems 1-2 are derived from the KKT conditions of that optimization problem rather than assumed from the model's predictions. Frozen GPT-2 embeddings are used with an explicit data-leakage guard (Section 3.3.2). The self-citations (Hwang et al. 2024, Lee et al. 2024, Zhang et al. 2021) appear as contextual references and do not carry the central outperformance claim. Two genuine concerns exist but are not circularity: the Section 4.1.1 universe conditions the backtest on future index membership (survivorship bias), and Proposition 1's constructed sequence converges to (mu1 - delta, mu2), not to mu, so the proof as written does not establish the claim. Neither concern makes the derivation equivalent to its inputs by construction, so no circular step is flagged.

Assumptions & free parameters 4 free parameters · 4 assumptions · 0 invented entities

The central claim rests on the validity of differentiable optimization gradients, a survivorship-free test set, and informative LLM embeddings. The first is compromised by the lambda normalization error and ignored box constraints; the second is violated by the sample filter; the third is unablated. These are the main load-bearing assumptions beyond standard supervised learning assumptions.

free parameters (4)
  • risk-aversion lambda = 0.9545 (balanced), selected from {0.0145, 0.2656, 0.9545, 2.4305, 3.4623}
    Controls the mean-variance trade-off in the optimization layer; final value chosen based on out-of-sample comparison in Table A.1, making it a hand-tuned/fitted parameter.
  • loss weight beta = 0.4
    Balances MSE and decision regret in Eq. (26); set as default without reported sensitivity analysis.
  • lookback length L and covariance window K = not reported for L; K = 3 months
    The return lookback length L affects normalization and decomposition, and the covariance estimation window is fixed at three months; both are modeling choices not justified by data.
  • attention heads, encoder depth, LLM hidden dimension = heads 2 or 4; depth 1,2,4; hidden 12,24,36,72
    Selected by validation; the paper reports ranges but not the chosen values for the main experiments.
assumptions (4)
  • domain assumption The sensitivity formulas (30) and (32) describe the true gradient of the optimization layer for arbitrary risk-aversion lambda and with box constraints active.
    Section 3.4 and Theorems 1-2 use these formulas to justify end-to-end training, but the proof fixes lambda = 1/2 and the derivation ignores the 0 <= w <= 1 constraints in problem (21).
  • domain assumption Index constituents that survived 2010-2023 are a valid test universe with no look-ahead bias.
    Section 4.1.1 restricts to firms with continuous index membership, which is a survivorship filter; the paper treats this as addressing bias instead of introducing it.
  • domain assumption Frozen GPT-2 embeddings of prompt statistics provide useful signal for return forecasting without data leakage.
    Section 3.2.2 assumes the LLM tokenizer and embeddings map financial prompts into informative representations; no ablation isolates the marginal value of the LLM component.
  • ad hoc to paper The absolute value in L_Decision is replaced by a differentiable surrogate without changing the optimum.
    Section 3.4 states 'ignoring non-differentiability at zero for the absolute value and assuming a differentiable approximation if needed', but no surrogate is specified or implemented.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Decision-informed Neural Networks with Large Language Model Integration for Portfolio Optimization." pith.science (2026). https://pith.science/paper/SNY7ONKL

@misc{pith2026250200828,
  author       = {Pith},
  title        = {Pith review of: Decision-informed Neural Networks with Large Language Model Integration for Portfolio Optimization},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/SNY7ONKL}},
  note         = {Machine review of arXiv:2502.00828}
}
read the original abstract

This paper addresses the critical disconnect between prediction and decision quality in portfolio optimization by integrating Large Language Models (LLMs) with decision-focused learning. We demonstrate both theoretically and empirically that minimizing the prediction error alone leads to suboptimal portfolio decisions. We aim to exploit the representational power of LLMs for investment decisions. An attention mechanism processes asset relationships, temporal dependencies, and macro variables, which are then directly integrated into a portfolio optimization layer. This enables the model to capture complex market dynamics and align predictions with the decision objectives. Extensive experiments on S\&P100 and DOW30 datasets show that our model consistently outperforms state-of-the-art deep learning models. In addition, gradient-based analyses show that our model prioritizes the assets most crucial to decision making, thus mitigating the effects of prediction errors on portfolio performance. These findings underscore the value of integrating decision objectives into predictions for more robust and context-aware portfolio management.

Figures

Figures reproduced from arXiv: 2502.00828 by the authors.

Figure 1
Figure 1. Schematic of the proposed Decision-Informed Neural Network (DINN) architecture for unified return forecasting and [PITH_FULL_IMAGE:figures/full_fig_p006_1.png] view at source ↗
Figure 2
Figure 2. Illustration of LLM-based prompt generation from pairwise outperformance statistics and macroeconomic summaries. [PITH_FULL_IMAGE:figures/full_fig_p009_2.png] view at source ↗
Figure 3
Figure 3. Comparison of portfolio standard deviation across experimental trials for models trained with prediction-based loss [PITH_FULL_IMAGE:figures/full_fig_p019_3.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

23 extracted references · 19 canonical work pages

  1. [1]

    et al., Gpt-4 technical report

    Achiam, J., Adler, S., Agarwal, S., Ahmad, L., Akkaya, I., Aleman, F.L., Almeida, D., Altenschmidt, J., Altman, S., Anadkat, S. et al., Gpt-4 technical report. arXiv preprint arXiv:2303.08774 ,

    arXiv

  2. [2]

    In this formulation, ˆΣ−1 t+h depends on ˆLt+h through the relation ˆΣt+h = ˆLt+h ˆL⊤ t+h

    Proof: Starting from the expression derived in Theorem 1 under the normalization 2 λ = 1, the optimal portfolio weights can be written as ˆwt+h = ˆΣ−1 t+hˆµt+h−p ˆΣ−1 t+h1, where p = 1⊤ ˆΣ−1 t+hˆµt+h−1 1⊤ ˆΣ−1 t+h1 . In this formulation, ˆΣ−1 t+h depends on ˆLt+h through the relation ˆΣt+h = ˆLt+h ˆL⊤ t+h. Thus, the chain rule of differentiation implies th...

    work page 2008

  3. [4]

    and Timmermann, A., Do return prediction models add economic value?.Journal of Banking & Finance, 2012, 36, 2974–2987

    Cenesizoglu, T. and Timmermann, A., Do return prediction models add economic value?.Journal of Banking & Finance, 2012, 36, 2974–2987. Chan, L.K., Karceski, J. and Lakonishok, J., On portfolio optimization: Forecasting covariances and choosing the risk model. The review of Financial studies , 1999, 12, 937–974. Chen, L., Pelger, M. and Zhu, J., Deep learn...

    work page 2012

  4. [6]

    and Fabozzi, F.J., The effects of errors in means, variances, and correlations on the mean-variance framework

    Chung, M., Lee, Y., Kim, J.H., Kim, W.C. and Fabozzi, F.J., The effects of errors in means, variances, and correlations on the mean-variance framework. Quantitative Finance, 2022, 22, 1893–1903. Costa, G. and Iyengar, G.N., Distributionally robust end-to-end portfolio construction.Quantitative Finance, 2023, 23, 1465–1482. DeMiguel, V., Garlappi, L. and Up...

    arXiv 2022

  5. [7]

    predict, then optimize

    Elmachtoub, A.N. and Grigas, P., Smart “predict, then optimize”. Management Science, 2022, 68, 9–26. Fama, E.F. and French, K.R., Common risk factors in the returns on stocks and bonds. Journal of financial economics, 1993, 33, 3–56. Fama, E.F. and French, K.R., A five-factor asset pricing model. Journal of financial economics , 2015, 116, 1–22. Feng, G., Gi...

    work page 2022

  6. [8]

    and Ma, T., Risk reduction in large portfolios: Why imposing the wrong constraints helps

    Jagannathan, R. and Ma, T., Risk reduction in large portfolios: Why imposing the wrong constraints helps. The journal of finance , 2003, 58, 1651–1683. Jin, M., Wang, S., Ma, L., Chu, Z., Zhang, J.Y., Shi, X., Chen, P.Y., Liang, Y., Li, Y.F., Pan, S. and Wen, Q., Time-LLM: Time series forecasting by reprogramming large language models. In Proceedings of th...

    work page 2003

  7. [9]

    and Su, Y., Characteristics are covariances: A unified model of risk and return.Journal of Financial Economics , 2019, 134, 501–524

    Kelly, B.T., Pruitt, S. and Su, Y., Characteristics are covariances: A unified model of risk and return.Journal of Financial Economics , 2019, 134, 501–524. Kim, J.H., Lee, Y., Kim, W.C. and Fabozzi, F.J., Mean-variance optimization for asset allocation. Journal of Portfolio Management , 2021a, 47, 24–40. Kim, J.H., Lee, Y., Kim, W.C., Kang, T. and Fabozzi...

    work page 2019

  8. [11]

    In Stochastic optimization models in finance , pp

    Lintner, J., The valuation of risk assets and the selection of risky investments in stock portfolios and capital budgets. In Stochastic optimization models in finance , pp. 131–155, 1975, Elsevier. Liu, Y., Hu, T., Zhang, H., Wu, H., Wang, S., Ma, L. and Long, M., iTransformer: Inverted Transformers Are 25 February 4, 2025 arxiv main Effective for Time Seri...

    work page 1975

Show all 23 references
  1. [12]

    Mandi, J., Kotary, J., Berden, S., Mulamba, M., Bucarey, V., Guns, T

    L¨ offler, G., The effects of estimation error on measures of portfolio credit risk.Journal of Banking & Finance, 2003, 27, 1427–1453. Mandi, J., Kotary, J., Berden, S., Mulamba, M., Bucarey, V., Guns, T. and Fioretto, F., Decision-focused learning: Foundations, state of the art,...

    2003

  2. [13]

    The Journal of Finance , 1952, 7, 77–91

    Markowitz, H., Portfolio Selection. The Journal of Finance , 1952, 7, 77–91. Michaud, R.O., The Markowitz optimization enigma: Is ‘optimized’optimal?. Financial analysts journal , 1989, 45, 31–42. Nie, Y., H. Nguyen, N., Sinthong, P. and Kalagnanam, J., A Time Series is Worth ...

    1952

  3. [15]

    et al., The matrix cookbook

    Petersen, K.B., Pedersen, M.S. et al., The matrix cookbook. Technical University of Denmark, 2008, 7,

    2008

  4. [16]

    and Driessen, K.V., A fast algorithm for the minimum covariance determinant estimator

    Rousseeuw, P.J. and Driessen, K.V., A fast algorithm for the minimum covariance determinant estimator. Technometrics, 1999, 41, 212–223. Sharpe, W.F., Capital asset prices: A theory of market equilibrium under conditions of risk. The journal of finance, 1964, 19, 425–442. Tan, ...

    1999

  5. [17]

    and Zohren, S., Estimation of Large Financial Covariances: A Cross-Validation Approach

    Tan, V. and Zohren, S., Estimation of Large Financial Covariances: A Cross-Validation Approach. arXiv preprint arXiv:2012.05757,

    2012 arXiv

  6. [20]

    and Zhang, W., Informer: Beyond Efficient Transformer for Long Sequence Time-Series Forecasting

    Zhou, H., Zhang, S., Peng, J., Zhang, S., Li, J., Xiong, H. and Zhang, W., Informer: Beyond Efficient Transformer for Long Sequence Time-Series Forecasting. In Proceedings of the The Thirty-Fifth AAAI Conference on Artificial Intelligence, AAAI 2021, Virtual Conference , Vol. 35,...

    2021

  7. [23]

    balanced

    Below, we detail the key hyper-parameter ranges and selection criteria employed for model training and evaluation. The codes are available at Anonymous Github. Model Hyper-Parameters and training strategy’s • Attention Heads: We examined configurations with either 2 or 4 attent...

    2024

  8. [48]

    and Xiu, D., Factor models, machine learning, and asset pricing

    Giglio, S., Kelly, B. and Xiu, D., Factor models, machine learning, and asset pricing. Annual Review of Financial Economics, 2022, 14, 337–368. Gu, S., Kelly, B. and Xiu, D., Empirical asset pricing via machine learning.The Review of Financial Studies, 2020, 33, 2223–2273. Gui...

    2022 arXiv

  9. [51]

    and Choo, J., Reversible instance normalization for accu- rate time-series forecasting against distribution shift

    Kim, T., Kim, J., Tae, Y., Park, C., Choi, J.H. and Choo, J., Reversible instance normalization for accu- rate time-series forecasting against distribution shift. In Proceedings of the International Conference on Learning Representations, 2021b. Kim, W.C., Lee, Y. and Lee, Y.H...

    2014

  10. [2017]

    and Kwon, R.H., End-to-end, decision-based, cardinality-constrained portfolio optimization

    Anis, H.T. and Kwon, R.H., End-to-end, decision-based, cardinality-constrained portfolio optimization. Eu- ropean Journal of Operational Research, 2025, 320, 739–753. Ansari, A.F., Stella, L., Turkmen, C., Zhang, X., Mercado, P., Shen, H., Shchur, O., Rangapuram, S.S., Arango,...

    2025

  11. [2020]

    and Rousseeuw, P., Minimum volume ellipsoid.Wiley Interdisciplinary Reviews: Computational Statistics, 2009, 1, 71–82

    Van Aelst, S. and Rousseeuw, P., Minimum volume ellipsoid.Wiley Interdisciplinary Reviews: Computational Statistics, 2009, 1, 71–82. Waswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A., Kaiser, L. and Polosukhin, I., Attention is all you need. In Proceedi...

    2009

  12. [2021]

    and Zohren, S., A universal end-to-end approach to portfolio optimiza- tion via deep learning

    Zhang, C., Zhang, Z., Cucuringu, M. and Zohren, S., A universal end-to-end approach to portfolio optimiza- tion via deep learning. arXiv preprint arXiv:2111.09170 ,

    arXiv

  13. [2022]

    et al., One fits all: Power general time series analysis by pretrained lm

    Zhou, T., Niu, P., Sun, L., Jin, R. et al., One fits all: Power general time series analysis by pretrained lm. Advances in neural information processing systems , 2023a, 36, 43322–43355. Zhou, T., Niu, P., Sun, L., Jin, R. et al., One Fits All: Power General Time Series Analysi...

    2025

  14. [2023]

    and Zohren, S., A Survey of Large Language Models for Financial Applications: Progress, Prospects and Challenges

    Nie, Y., Kong, Y., Dong, X., Mulvey, J.M., Poor, H.V., Wen, Q. and Zohren, S., A Survey of Large Language Models for Financial Applications: Progress, Prospects and Challenges. arXiv preprint arXiv:2406.11903,

    arXiv

  15. [2024]

    and Lim, A.E., Machine learning and portfolio optimization

    Ban, G.Y., El Karoui, N. and Lim, A.E., Machine learning and portfolio optimization. Management Science, 2018, 64, 1136–1154. Bekaert, G., Harvey, C.R. and Lumsdaine, R.L., The dynamics of emerging market equity flows. Journal of International money and Finance , 2002, 21, 295–...

    2018

Pith tools

Reviewed August 9, 2026 · model on record in the stance chip above.