FinXplore: An Adaptive Deep Reinforcement Learning Framework for Balancing and Discovering Investment Opportunities

Arishi Orra; Himanshu Choudhary; Manoj Thakur

REVIEW 3 major objections 4 minor 18 references

FinXplore: An Adaptive Deep Reinforcement Learning Framework for Balancing and Discovering Investment Opportunities

T0 review · 3 major / 4 minor · reviewed 2026-08-05 · deepseek-v4-flash

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

Pith's one-line read FinXplore claims that a second reinforcement-learning agent, which scouts new assets and is accepted only when it improves a 60-day risk-adjusted return, beats fixed-universe benchmarks on NIFTY and DJIA backtests.

desk verdict The dual-agent exploration idea is a legitimate extension of Aquino et al., but Algorithm 1's acceptance rule makes the backtest results unreadable as tradeable performance. read the letter →

arxiv 2509.10531 v1 pith:2XZWA6AJ submitted 2025-09-05 cs.LG cs.AI

Himanshu Choudhary , Arishi Orra , Manoj Thakur This is my paper · ORCID

classification cs.LGcs.AI

keywords portfoliooptimizationdeepreinforcementlearningPPODQNexploration-exploitationtrade-offSharperatioinvestmentuniversebacktesting

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

The paper sets out to show that portfolio optimization should not be stuck with a fixed set of assets. FinXplore pairs one reinforcement-learning agent (a policy-gradient method called PPO) that allocates wealth across the familiar universe with a second agent (deep Q-learning) that proposes new assets from an extended universe of commodities. The second agent is rewarded by how much its suggestion raises the portfolio's 60-day Sharpe ratio, and the first agent accepts the suggestion only when that ratio improves. On backtests from January 2022 to November 2024, the paper reports cumulative returns of 127.91% on NIFTY stocks versus 37.61% for the index, and 49.56% versus 23.23% on DJIA, with the best Sharpe and Calmar ratios among all compared strategies. If those results hold, a guided exploration channel is a genuine addition to DRL portfolio management rather than a speculative extra.

What carries the argument

The central mechanism is Algorithm 1's accept/reject loop with Sharpe-ratio gating. At each rebalancing period, Agent 1 produces portfolio weights; Agent 2 proposes an asset; the framework re-optimizes weights with κ=10% of wealth reserved for the proposed asset, computes the 60-day Sharpe ratio before and after, and keeps the new asset only if the ratio rises. The same ΔSR is the reward that trains the DQN agent, while the PPO agent is trained on the Sharpe ratio of the chosen portfolio. The Sharpe-ratio gate is what converts open-ended exploration into targeted diversification, and the rolling 60-day window plus a 0.05% transaction cost defines how the environment evaluates each action.

What would settle it

Re-run the NIFTY and DJIA backtests with Algorithm 1's acceptance test computed on Sharpe ratios that end before the rebalance (for example, through t−1 only, or via a walk-forward procedure that retrains the agents without touching the test period). If the corrected FinXplore does not beat the without-exploration agent and the market index by the reported margins, the superior performance is an artifact of look-ahead rather than of exploration.

Watch

Extended reading notes

Core claim

The discovery claim is that exploitation and exploration can be cleanly separated into two cooperating deep-RL agents, and that the collaboration improves the portfolio on both raw and risk-adjusted terms. Agent 1 (PPO) learns how to allocate wealth among the existing n assets; Agent 2 (DQN) learns which asset from the extended universe to propose. The link between them is the reward difference ΔSR = SR_new − SR_current, so Agent 2 is trained to find assets that improve the risk-adjusted portfolio, and Agent 1's acceptance rule (keep the new asset only when SR_new > SR_current) makes the allocation adaptive to changing markets. The reported experiments are intended to demonstrate that this d

Load-bearing premise

The load-bearing premise is that the agent can know the realized returns of the period it is about to rebalance: Algorithm 1 accepts a newly proposed asset only when SR_new, computed from the latest 60 portfolio returns, exceeds SR_current, which means the decision uses the same period's outcome to choose that period's trade. If a backtest is instead forced to act only on information available before the trade, this acceptance rule may no longer support the reported returns.

Editorial extensions

If this is right

On the paper's reported results, FinXplore beats every benchmark on cumulative return, annualized return, Sharpe ratio, and Calmar ratio on both the NIFTY and DJIA datasets.
The gap between FinXplore and the DRL agent without exploration (127.91% vs 93.86% on NIFTY; 49.56% vs 37.15% on DJIA) indicates that the exploration channel, not the choice of PPO alone, drives most of the improvement.
The Sharpe-ratio acceptance rule keeps the framework from being forced into bad suggestions: it only allocates the 10% exploration budget when the proposed asset improves the portfolio's risk-adjusted return.
The quarterly-return plots show that exploration can produce worse quarters as well as better ones, so the framework's edge is statistical over the full period rather than uniform across regimes.

Reading between the lines

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

A corrected experiment that computes both Sharpe ratios only from data available before the rebalance would settle whether the reported edge survives without look-ahead; this is the first test to run before adopting the strategy.
The exploration action space could be widened from five commodities to a large pool of ETFs or sectors, with the DQN learning a ranking over candidates rather than a fixed choice; the ΔSR reward supports that generalization directly.
The fixed κ=10% exploration budget is an obvious place to make the framework adaptive: a meta-controller could raise the budget in regimes where accepted proposals consistently improve Sharpe and lower it during turbulent quarters.
A practical deployment would need rebalancing-frequency and transaction-cost sensitivity analysis; daily rebalancing at 0.05% cost may not be realistic once the exploration channel is added.

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.

Desk Editor's Note

read the letter

Quick take: this paper extends Aquino et al.'s extended-universe idea by using PPO for exploitation and DQN for exploration. That's a legitimate, if incremental, combination. But the backtest design in Algorithm 1 makes the reported returns unreliable, and the claim that exploration helps is not established.

What's good: The two-agent setup is clearly described, the state includes technical indicators, and the authors use standard DRL components. The results are reported with standard deviations from five runs, which is better than many finance papers. The writing is straightforward.

The problem: Algorithm 1 decides whether to include the DQN-suggested asset by comparing the trailing 60-day Sharpe ratios of the PPO allocation and the reoptimized allocation. As written, it is ambiguous whether the 60-day window includes the current period's return. If it does, that's textbook look-ahead. If it ends at the previous close, the rule is a momentum/selection rule that picks, ex post, the better of two candidate portfolios for each period. Under either reading, the reported cumulative returns (127.91% on NIFTY, 49.56% on DJIA) are not the returns of a single tradeable strategy; they are the results of a selection mechanism that uses realized performance to choose allocations. The gap over the 'Without Exploration' baseline is therefore partly a construction artifact, not a clean test of the exploration concept.

The paper also claims the framework is 'adaptive' but fixes kappa at 10%, and it never specifies the exact point-in-time protocol. That is a fixable omission, but until it's addressed, the empirical results cannot be taken as evidence.

One more subtle issue: the reward R_PPO = SR_current is the Sharpe ratio of the current proposed portfolio over the past 60 days, not the realized return of the portfolio the agent actually held. This makes the training signal a backtest of the current suggestion rather than a reflection of realized performance. That is a bit unusual and should be explained.

Bottom line: the architecture is reasonable and the paper is readable, but the evaluation has a load-bearing flaw. The authors should rerun the backtest with a strict point-in-time protocol, and compare against a baseline that also gets the same 'pick the better of two' selection rule without DRL, to isolate the effect of exploration. Without that, the superiority claim is unsupported.

For a reading group, this is a useful case study in backtest bias. I wouldn't cite it in my own work yet. If the protocol gets fixed, it could become a solid incremental contribution.

Referee Report

3 major / 4 minor

Summary. The paper proposes FinXplore, a dual-agent deep reinforcement learning framework for portfolio optimization. A PPO agent allocates wealth within a pre-defined investment universe, while a DQN agent recommends adding assets from an extended universe (commodities). The two agents are trained on NIFTY and DJIA data from 2011-2021 and backtested on 2022-2024, with reported cumulative returns of 127.91% (NIFTY) and 49.56% (DJIA), surpassing all benchmarks. The paper claims this demonstrates the benefit of dynamically balancing exploitation of existing assets with exploration of new opportunities.

Significance. The dual-agent architecture and the idea of expanding the investment universe via a learned exploration policy are potentially interesting and go beyond fixed-universe DRL portfolio methods. The use of two established RL algorithms (PPO and DQN), evaluation against several baselines, and reporting of multiple risk-adjusted metrics are positive features. However, the main empirical claim is not credible as presented: the acceptance rule in Algorithm 1 appears to use realized returns from the period being traded, making the reported backtest results unattainable in real time. Because the central result rests on this evaluation, the contribution is not currently established.

major comments (3)

[Algorithm 1 and Section II-A] Algorithm 1 (lines 2-10) computes SR_current = f(a_U) and SR_new = f(a_E) at each rebalancing step and selects a_E only if SR_new > SR_current. Section II-A defines the reward as SR_60, the Sharpe ratio over the last 60 portfolio returns, but does not specify whether the current period's return is included in that window. If the 60-day window includes the period about to be traded, the acceptance decision uses the same realized outcome that the policy is trying to predict, which is a look-ahead bias. If the window ends at the previous close, the rule is a mechanical ex post selection rule rather than evidence of learned predictive skill. In neither reading are the reported cumulative returns (Table II: 127.91%; Table III: 49.56%) established as tradeable out-of-sample results.
[Tables II and III, 'Without Exploration' baseline] The comparison against the 'Without Exploration' DRL baseline is biased by construction. Algorithm 1 always chooses the allocation with the higher realized 60-day Sharpe ratio, so FinXplore's realized Sharpe is at least that of the no-exploration portfolio in every period. This mechanically guarantees a nonnegative ΔSR for the DQN reward and systematically favors FinXplore. The roughly 34pp (NIFTY) and 12pp (DJIA) cumulative-return gaps therefore largely reflect the selection rule, not an improved learned policy.
[Table I and Sections III-V] The paper repeatedly claims the framework 'dynamically balances' exploitation and exploration, but the exploration weight is fixed at κ=10% (Table I) for all periods. Algorithm 1 only toggles between the full existing-universe allocation and the allocation that includes the suggested asset; there is no mechanism that adjusts the fraction of wealth devoted to exploration in response to market conditions. Thus the central 'adaptive balancing' claim is unsupported by the implemented method.

minor comments (4)

[Throughout] There are typographical errors: 'effciency' in the Abstract, 'to access the risk-adjusted portfolio return' should be 'to assess', and 'closed' should be 'close' in Section IV-A. A careful proofread is needed.
[Section IV-B] Hyperparameter tuning is described only by search ranges (Table I); the final selected hyperparameters are not reported. This hampers reproducibility, especially since no code repository is mentioned.
[Tables II and III] Markowitz and the market index are reported without standard deviations, while DRL methods have means ± std from five runs. It would be helpful to know whether the reported point estimates are stable or based on a single run.
[Figures 4 and 5] The quarterly-return figures are plotted only for FinXplore, 'Without Explore', and 'Markowitz Baseline', which is a subset of the models in Tables II and III. The caption for Figure 4 labels the third series 'Baseline' but does not identify which baseline it is.

Circularity Check

1 steps flagged · score 6.0 of 10

Algorithm 1's accept-if-higher-Sharpe rule makes the exploration advantage over 'Without Exploration' true by construction.

self definitional [Algorithm 1, lines 6–12; Section III (Proposed Methodology)]
"if SR_new > SRcurrent then w←a_E else w←a_U ... R_PPO ←SR_current; R_DQN ←∆SR=SR new −SR current"

At every rebalance, Algorithm 1 defines the final portfolio w to be a_E when the 60-day Sharpe ratio of the explored portfolio exceeds that of the current portfolio, and a_U otherwise. Hence f(w) = max(f(a_U), f(a_E)) ≥ f(a_U). The 'Without Exploration' baseline is exactly w = a_U at all times, so its Sharpe ratio is f(a_U). Therefore FinXplore's per-rebalance Sharpe ratio is, by construction, at least that of the baseline. The paper then reports FinXplore's superior cumulative returns and Sharpe ratios (Tables II–III) as evidence that the dual-agent design works, but the gap is guaranteed by the max-selection rule rather than by any learned predictive skill. The reward updates use the same f: R_PPO = SR_current and R_DQN = SR_new − SR_current, so the selection criterion, the training rewa

full rationale

The central empirical claim (Section IV-E) is that FinXplore outperforms all benchmarks, including 'Without Exploration.' That specific comparison is not independent: Algorithm 1 selects between a_U and a_E by comparing their 60-day Sharpe ratios, so the selected portfolio's Sharpe is the maximum of the two candidates by construction. The no-exploration baseline never receives this option, so the reported improvement over it is partly forced by the algorithm's definition, not by demonstrated out-of-sample predictive skill. No load-bearing self-citation was found; reference [16] is a minor self-citation for hyperparameters and transaction costs and does not support the main derivation. The paper also leaves ambiguous whether the 60-return window includes the period to be traded, which could entail look-ahead; that is a separate correctness issue, not the circularity scored here. Overall, the evaluation design creates a partial circularity because the metric used to accept/reject the exploration is the same metric used as the agents' reward and as the reported performance measure.

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

The central claim relies on the standard RL assumptions plus two paper-specific choices: the fixed kappa=10% exploration budget and the hindsight-based acceptance rule. The latter is the most serious because it makes the reported results unachievable in practice.

free parameters (5)

Exploration weight kappa = 10%
Fixed in Table I; not optimized or adjusted over time despite the claim of adaptive balancing.
Transaction cost delta = 0.05%
Assumed from reference [16]; results would change with different costs.
Sharpe window = 60 days
Reward uses SR over last 60 returns, chosen without justification.
Hyperparameters = not reported (ranges only)
Tuned with Hyperopt but selected values are not given; different choices could change results.
Stock universe = 18 unnamed stocks per market
Randomly chosen stocks are not listed, so the experiment cannot be replicated.

assumptions (5)

domain assumption Markov Decision Process is an appropriate model for portfolio optimization.
Section II-A models the portfolio problem as an MDP with p, r, gamma. This is standard but unproven for financial markets.
domain assumption The 60-day Sharpe ratio is an appropriate reward signal for learning.
Section II-A defines reward as SR_60; no evidence that this shapes policy in a risk-adjusted way better than alternatives.
domain assumption Historical OHLCV data and technical indicators are sufficient state for the environment.
Section II-A lists the state features; no analysis of whether these suffice.
ad hoc to paper The realized Sharpe ratio of the current period is available when deciding the current period's allocation.
Algorithm 1 lines 4-10 depend on SR_new and SR_current computed over the last 60 returns. This is only possible with hindsight, making the decision rule untradeable.
domain assumption The five selected commodities form a representative extended investment universe.
Section IV-A chooses gold, oil, silver, copper, natural gas without criteria or ablation.

how reviews work

0 comments

Cite this review

Pith. "Pith review of FinXplore: An Adaptive Deep Reinforcement Learning Framework for Balancing and Discovering Investment Opportunities." pith.science (2026). https://pith.science/paper/2XZWA6AJ

@misc{pith2026250910531,
  author       = {Pith},
  title        = {Pith review of: FinXplore: An Adaptive Deep Reinforcement Learning Framework for Balancing and Discovering Investment Opportunities},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/2XZWA6AJ}},
  note         = {Machine review of arXiv:2509.10531}
}

read the original abstract

Portfolio optimization is essential for balancing risk and return in financial decision-making. Deep Reinforcement Learning (DRL) has stood out as a cutting-edge tool for portfolio optimization that learns dynamic asset allocation using trial-and-error interactions. However, most DRL-based methods are restricted to allocating assets within a pre-defined investment universe and overlook exploring new opportunities. This study introduces an investment landscape that integrates exploiting existing assets with exploring new investment opportunities in an extended universe. The proposed approach leverages two DRL agents and dynamically balances these objectives to adapt to evolving markets while enhancing portfolio performance. One agent allocates assets within the existing universe, while another assists in exploring new opportunities in the extended universe. The effciency of the proposed methodology is determined using two real-world market data sets. The experiments demonstrate the superiority of the suggested approach against the state-of-the-art portfolio strategies and baseline methods.

Figures

Figures reproduced from arXiv: 2509.10531 by the authors.

**Figure 1.** The overall architecture of FinXplore Algorithm 1 Portfolio Optimization with Exploration Assistance Require: States st and ct for Existing and Extended Investment Universes respectively, Agent 1 (PPO), Agent 2 (DQN), Existing Universe U, Extended Universe E Ensure: Updated portfolio weights w 1: aU ← PPO(st, at) Portfolio weights from Agent 1 2: SRcurrent ← f(aU ) Sharpe Ratio for current portfolio 3: a∗ ← DQN(st… view at source ↗

**Figure 2.** Cumulative Return Plot of the proposed FinXplore approach and state-of-the-art portfolio strategies over the trading period on Nifty Dataset [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗

**Figure 3.** Cumulative Return Plot of the proposed FinXplore approach and state-of-the-art portfolio strategies over the trading period on DJIA Dataset [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗

Figures from the paper (2 more)

**Figure 4.** Figure 4: Quarterly Returns for Nifty dataset over the trading period [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]

**Figure 5.** Figure 5: Quarterly Returns for DJIA dataset over the trading period [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

18 extracted references · 15 canonical work pages

[1]

Portfolio selection,

H. Markowitz, “Portfolio selection,” The journal of finance, vol. 7, no. 1, pp. 77–91, 1952

work page 1952
[2]

HA Davis and A

M. HA Davis and A. R. Norman. Portfolio selection with transaction costs. Mathematics of operations research, 15(4):676–713, 1990

work page 1990
[3]

Ehrgott, K

M. Ehrgott, K. Klamroth, and C. Schwehm. An MCDM approach to portfolio optimization. European Journal of Operational Research, 155(3):752–770, 2004

work page 2004
[4]

A novel approach to incorporate investor’s preference in fuzzy multi-objective portfolio selection problem using credibility measure,

H. Jalota, P. K. Mandal, M. Thakur, and G. Mittal, “A novel approach to incorporate investor’s preference in fuzzy multi-objective portfolio selection problem using credibility measure,” Expert Systems with Applications, vol. 212, p. 118583, 2023

work page 2023
[5]

Mastering the game of Go with deep neural networks and tree search,

D. Silver, A. Huang, C. J. Maddison, A. Guez, L. Sifre, G. Van Den Driessche, J. Schrittwieser, I. Antonoglou, V . Panneershelvam, M. Lanctot, et al., “Mastering the game of Go with deep neural networks and tree search,” Nature, vol. 529, no. 7587, pp. 484–489, 2016

work page 2016
[7]

A deep reinforcement learning frame- work for the financial portfolio management problem

Z. Jiang, D. Xu, J. Liang. “A deep reinforcement learning frame- work for the financial portfolio management problem.” arXiv preprint arXiv:1706.10059. 2017

arXiv 2017
[8]

Deeptrader: A deep reinforcement learning approach for risk-return balanced portfolio management with market conditions embedding

Z. Wang, B. Huang, S. Tu, K. Zhang, and L. Xu. “Deeptrader: A deep reinforcement learning approach for risk-return balanced portfolio management with market conditions embedding.” In Proceedings of the AAAI conference on artificial intelligence, volume 35, pages 643–650, 2021

work page 2021
[9]

Deep reinforcement learning for stock portfo- lio optimization by connecting with modern portfolio theory

J. Jang and N. Seong. “Deep reinforcement learning for stock portfo- lio optimization by connecting with modern portfolio theory.” Expert Systems with Applications, 218:119556, 2023

work page 2023

Show all 18 references

[10]

Multi-period portfolio opti- mization using a deep reinforcement learning hyper-heuristic approach

T. Cui, N. Du, X. Yang, and S. Ding. “Multi-period portfolio opti- mization using a deep reinforcement learning hyper-heuristic approach.” Technological Forecasting and Social Change, 198:122944, 2024

2024
[11]

Y . Ye, H. Pei, B. Wang, P. Y . Chen, Y . Zhu, J. Xiao, and B. Li. ”Reinforcement-learning based portfolio management with augmented asset movement prediction states.” In Proceedings of the AAAI confer- ence on artificial intelligence (V ol. 34, No. 01, pp. 1112-1119), 2020, April

2020
[12]

Continuous-time mean–variance portfolio selection: A reinforcement learning framework

H. Wang and X. Y . Zhou. “Continuous-time mean–variance portfolio selection: A reinforcement learning framework.” Mathematical Finance 30, no. 4, 1273-1308. 2020

2020
[13]

Portfolio selection with exploration of new investment opportunities

L. D. G. Aquino, D. Sornette, and M. S. Strub. “Portfolio selection with exploration of new investment opportunities.” European Journal of Operational Research, V olume 310, Issue 2, Pages 773-792, 2023

2023
[14]

FinRL-Meta: Market environments and benchmarks for data-driven financial reinforcement learning

X.Y . Liu, Z. Xia, J. Rui, J. Gao, H. Yang, M. Zhu, C. Wang, Z. Wang, and J. Guo. “FinRL-Meta: Market environments and benchmarks for data-driven financial reinforcement learning.” Advances in Neural Information Processing Systems, 35, pp.1835-1849, 2022

2022
[15]

The sharpe ratio. Streetwise–the Best of the Journal of Portfolio Management

W. F. Sharpe. “The sharpe ratio. Streetwise–the Best of the Journal of Portfolio Management”. 3, 169–85, 1998

1998
[16]

Dynamic Rein- forced Ensemble using Bayesian Optimization for Stock Trading

A. Orra, A. Bhambu, H. Choudhary, and M. Thakur. “Dynamic Rein- forced Ensemble using Bayesian Optimization for Stock Trading.” In Proceedings of the 5th ACM International Conference on AI in Finance (pp. 361-369), November, 2024

2024
[17]

Schulman, F

J. Schulman, F. Wolski, P. Dhariwal, A. Radford, O. Klimov. “Proximal policy optimization algorithms.“ arXiv preprint arXiv:1707.06347, 2017

2017 arXiv
[18]

V . Mnih, K. Kavukcuoglu, D. Silver, A. Graves, I. Antonoglou, D. Wierstra, and M. Riedmiller. ”Playing atari with deep reinforcement learning.” arXiv preprint arXiv:1312.5602, 2013

2013 arXiv
[19]

Deep reinforcement learning for portfolio selection

Y . Jiang, J. Olmo, and M. Atwi. “Deep reinforcement learning for portfolio selection.” Global Finance Journal, 62, p.101016, 2024

2024

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.

[1] [1]

Portfolio selection,

H. Markowitz, “Portfolio selection,” The journal of finance, vol. 7, no. 1, pp. 77–91, 1952

work page 1952

[2] [2]

HA Davis and A

M. HA Davis and A. R. Norman. Portfolio selection with transaction costs. Mathematics of operations research, 15(4):676–713, 1990

work page 1990

[3] [3]

Ehrgott, K

M. Ehrgott, K. Klamroth, and C. Schwehm. An MCDM approach to portfolio optimization. European Journal of Operational Research, 155(3):752–770, 2004

work page 2004

[4] [4]

A novel approach to incorporate investor’s preference in fuzzy multi-objective portfolio selection problem using credibility measure,

H. Jalota, P. K. Mandal, M. Thakur, and G. Mittal, “A novel approach to incorporate investor’s preference in fuzzy multi-objective portfolio selection problem using credibility measure,” Expert Systems with Applications, vol. 212, p. 118583, 2023

work page 2023

[5] [5]

Mastering the game of Go with deep neural networks and tree search,

D. Silver, A. Huang, C. J. Maddison, A. Guez, L. Sifre, G. Van Den Driessche, J. Schrittwieser, I. Antonoglou, V . Panneershelvam, M. Lanctot, et al., “Mastering the game of Go with deep neural networks and tree search,” Nature, vol. 529, no. 7587, pp. 484–489, 2016

work page 2016

[6] [7]

A deep reinforcement learning frame- work for the financial portfolio management problem

Z. Jiang, D. Xu, J. Liang. “A deep reinforcement learning frame- work for the financial portfolio management problem.” arXiv preprint arXiv:1706.10059. 2017

arXiv 2017

[7] [8]

Deeptrader: A deep reinforcement learning approach for risk-return balanced portfolio management with market conditions embedding

Z. Wang, B. Huang, S. Tu, K. Zhang, and L. Xu. “Deeptrader: A deep reinforcement learning approach for risk-return balanced portfolio management with market conditions embedding.” In Proceedings of the AAAI conference on artificial intelligence, volume 35, pages 643–650, 2021

work page 2021

[8] [9]

Deep reinforcement learning for stock portfo- lio optimization by connecting with modern portfolio theory

J. Jang and N. Seong. “Deep reinforcement learning for stock portfo- lio optimization by connecting with modern portfolio theory.” Expert Systems with Applications, 218:119556, 2023

work page 2023