Papers

We study continuous-time multi-asset portfolio choice and consumption under smooth pointwise constraints, including state-dependent feasible sets. The method separates dynamic information acquisition from local constrained recovery. A pointwise-feasible neural actor generates reference rollouts; after training, its realized latent outputs are frozen and first- and second-order adjoints are harvested from a fixed-latent open-loop backpropagation-through-time graph. Feedback therefore generates the reference trajectory without restricting the adjoint formulation to Markov controls. Conditional on the harvested adjoints, deployment solves a local generalized Pontryagin-Hamiltonian problem: quadratic-affine portfolio blocks are recovered exactly by a quadratic program, while a log barrier approximates more general regular KKT branches. We establish local chart representations, an OL-BPTT-to-adjoint correspondence retaining orthogonal martingale residuals, and an end-to-end bound from reference value loss and numerical errors to recovered-policy and local QP-gap errors. Analytical constant- and predictable-opportunity benchmarks validate the adjoints. Common-input experiments show that recovery reduces residual PMP/KKT error left by finite-budget direct policy optimization, including under a state-dependent consumption cap and with up to 100 risky assets. Scalability concerns the constrained action block rather than dimension-free state-space complexity.

The authors present a rigorous empirical evaluation of three distinct optimization paradigms for institutional factor portfolio construction: an entropy-based photonic quantum annealer (Dirac-3, Quantum Computing Inc.), a commercial mixed-integer programming solver (Gurobi), and a model-free deep reinforcement learning agent (SAC). Evaluating these pipelines on the Jensen-Kelly-Pedersen 13-factor equity library across 164 months test window, we implement a full factorial penalty sweep comprising 48 hyperparameter configurations that govern return, volatility, and skewness trade-offs. Our findings demonstrate that while photonic hardware can locate superior risk-return topologies within a narrow operating range, classical mixed-integer programming remains superior for risk-constrained mandates requiring tight tail-risk control and cross-seed stability. Furthermore, we document structural failure modes in reinforcement learning factor allocators under unanchored higher-moment shaping. We translate these empirical results into actionable, mandate-specific guidelines for quantitative portfolio managers deploying advanced optimization engines.

Large language models can extract richer signals from financial news than fixed sentiment lexicons, and recent work has explored feeding such signals into portfolio construction. We study an uncertainty-aware construction that feeds model-predicted risk – decomposed into aleatoric and epistemic components – directly into the covariance matrix of portfolio allocators, rather than treating portfolio risk as fixed or adjusting only expected returns. We evaluate the pipeline on Russell 2000 equities under three stock-selection regimes: a pure-alpha trigger that isolates abnormal stock moves not explained by macro indicators, a pure-beta trigger that captures macro-indicator moves before the stock itself fires, and a beta trigger in which both channels agree. Across the full holding-period grid, the separated pure-alpha and pure-beta legs usually dominate the beta intersection on Sharpe and return. Two horizons are especially informative. At one day, pure beta can work under low and moderate transaction costs because it captures immediate lead-lag spillovers from liquid macro and sector indicators into exposed small-cap stocks, but this advantage disappears at 100 bps when turnover and microstructure noise dominate. At 40 days, pure beta works for a different reason: slower macro repricing overtakes the firm-specific pure-alpha channel. The strongest conservative row is pure beta with GPT-4o mini sentiment, a Student-t target, a 40-day holding period, and risk parity allocation, reaching Sharpe 2.33 at 100 bps. The results suggest that stock-selection regime and allocator choice matter at least as much as the sentiment model, and that separating firm-specific and macro-exposure triggers is more informative than requiring both to fire simultaneously.

Automated quantitative research has made striking progress, yet each system answers the same question: which strategy scores highest on a scalar metric? We argue this question is incomplete. Professional investors do not order "the highest return"; they order an identity–pure stock-selection alpha uncontaminated by style exposure, resilient in unilateral market declines, within turnover and capacity budgets. We call the incumbent paradigm result-oriented and propose Objective-Oriented Quantitative Investment (OOQI): a specification-driven framework in which (i) the full strategy pipeline is modeled as a typed design space of interchangeable modules with explicit interface contracts (8.85 x 10^8 assemblies in our reference instantiation); (ii) investor intent is formalized as a strategy profile specification–a composable language of measurable, falsifiable clauses from eight requirement families, with hard/soft semantics and an interaction algebra; and (iii) a compiler translates specifications into constrained assemblies and verifies satisfaction clause-by-clause. Because search over large assembly spaces inflates apparent satisfaction, we develop a verification protocol treating the satisfaction rate itself as a statistical object, subject to deflation for search width, temporal holdout, and random-assembly null models. A synthetic demonstration with 32 pipeline assemblies shows that result-oriented selection attains the top in-sample information ratio while satisfying only 25% of the specification, whereas specification-driven selection satisfies 100% of it at a 5.5% score cost. The accompanying theory shows satisfaction-driven synthesis is NP-hard in general yet constant-factor approximable in a conflict-free regime; specifications form a lattice dual to assemblies; each clause carries a Lagrangian shadow price; and rolling re-certification is anytime-valid via e-processes.

A new class of software systems is transforming investment analysis. Large language model agents assembled into collaborative team structures including analysts, researchers, and risk managers are increasingly deployed across financial markets. Yet current multi-agent frameworks share a critical limitation: they rely on the foundational assumption that companies can be valued through traditional cash flows. This paradigm fails in clinical-stage biotechnology, where enterprise value depends entirely on binary scientific and regulatory milestones. To bridge this gap, this paper introduces a specialized multi-agent framework. Its valuation layer translates qualitative scientific judgment into defensible valuations for pre-revenue assets; its cross-market coordination layer reconciles pricing across international venues simultaneously; and its conflict-fusion mechanism systematically arbitrates between bullish scientific conviction and cautious regulatory constraints in a domain-specific manner. Crucially, the architecture is not a speculative design: it encodes a method the author first executed by hand as sole portfolio manager of China's first dedicated cross-border biotechnology fund, a human practice that returned 127.17% against a 50.67% benchmark within sixteen months. That record is evidence for the underlying method rather than for any AI system; no implementation is evaluated here. This paper presents the framework at the architectural level, establishing foundational design principles for extending agentic investment systems into complex, event-driven asset classes they currently serve poorly.

How much capital a trading strategy can absorb before its edge disappears is a causal question about how much is deployed, but it is answered with observational proxies that rest on incompatible assumptions. We ask what experiment would answer it instead, and show that two features of the problem interact to constrain any answer. Deployed capital erodes the edge gradually, so a trial of fixed length measures less than the eventual effect; and parallel implementations of one strategy trade the same securities, so they are not independent units. Comparing implementations on the same date removes market-wide shocks, which is what makes the comparison credible. But the crowding created by the strategy's own accumulated position is common to those implementations too, and an arbitrary date effect absorbs it exactly: the comparison that makes the experiment robust is the one that prevents it from measuring the crowding capacity is about. A same-date design recovers one implementation's private response at the prevailing level of aggregate positioning, and reaching the aggregate effect requires either implementations with deliberately different exposure to that position or variation in it over time. We characterise what each route identifies and what it costs, establish how far a fixed holding period understates the eventual effect and how to correct for it, and show what a finite set of deployment levels can and cannot reveal. A calibration on a purpose-built panel illustrates the resulting design rules and prices a study that would follow them.

We develop a certified, scalable approximation for high-dimensional Wasserstein distributionally robust portfolio optimization. For expected-utility maximization under order-one Wasserstein ambiguity, standard duality yields a semi-infinite convex program. For long-only portfolios with box support under the one-norm ground metric, an exact sample-specific vertex reformulation provides an exponential-size computational benchmark. We then majorize the utility by supporting hyperplanes and dualize the support subproblems, obtaining a finite hyperplane–dual formulation over compact polyhedral supports. Under the one-norm ground metric and polyhedral portfolio constraints, this formulation is a polynomial-size linear program. The uniform utility-approximation error bounds both the robust-value error and the near-optimality gap for the original robust problem. Experiments validate the certified approximation and demonstrate monthly 476-asset rebalancing and computational scalability to 1,000 assets.

Current portfolio construction methods are either agnostic to the effects of idiosyncratic shocks (standard factor models) or to the latent data structure driving systematic returns (recent graph-based approaches). This presents an opportunity to combine the complementary market aspects captured by the factor and graph domains, allowing asset allocations to operate directly on the underlying market structure, rather than on its observed co-movement or its finite-sample artefacts. In this work, we introduce the Mutually-INformed Graph-Locality and Exposures framework (MINGLE), which mutually regularises the factor and graph domains by redefining graph locality through systematic factor exposure profiles, rather than via observed co-movements. This is formalised through a unified Alternating Direction Method of Multipliers (ADMM) framework that jointly learns a latent factor representation and its induced graph topology directly from market returns. The resulting exposure-similarity graph aligns more closely with established economic sectors than conventional correlation-based graphs. Portfolios constructed from this representation are shown to consistently outperform their correlation-based counterparts across a range of volatility regimes and transaction cost levels. For rigour, paired statistical testing confirms that these gains stem from the reconciliation of the graph and factor domains.

The enormous growth in datasets, both in number and size, has prompted investors to adapt to new ways for assimilating information. Normatively, the approach has been to integrate such datasets into pricing formulations and assess the performance of portfolios created thereafter. However, such approaches underestimate their influence in portfolio investments by limiting their impact to pricing only. While being theoretically valid, this results in a potential sub-optimal performance in the presence of real-life decision constraints, and a blind spot for performance attribution. We start by analysing investment decisions from a knowledge perspective, which unfurls a new structure. We then propose a FinTech process termed Knowledge Optimisation that aims to integrate the influence of knowledge components that could be related to data, models, or business units that extract information. A 3-stage process, namely, decision structure, portfolio selection, and performance assessment is designed. We present an alternative to the ex-ante Sharpe Ratio, integrating a term for knowledge units. Through scenario analysis involving portfolio investment situations, we illustrate the utility. By design, the process improves the importance of knowledge in investment decisions.

Cross-correlations between financial signals are neither scale-free nor amplitude-independent: they vary with the time scale over which they are measured and with the magnitude of the fluctuations that dominate the average. We exploit this structure to construct a portfolio allocation model in which the risk functional is the signed fluctuation function of multifractal cross-correlation analysis (MFCCA), indexed by a scale $s$ and a fluctuation order $q$. Unlike MFDCCA-type criteria, which rectify local detrended covariances before aggregation, MFCCA retains their sign, so that co-moving and counter-moving components contribute to risk with opposite signs; for $q=2$ the resulting quadratic form coincides with the detrended fluctuation function of the portfolio series itself, recovering the mean–variance criterion as a scale-dependent limit. Using two-component ARFIMA and Markov-switching multifractal processes, we show that prescribed multiscale and multifractal dependence is transmitted into the optimal weights, and that sign preservation contributes more to the reduction of tail risk than aggregation over fluctuation orders. Applied to financial multi-assets, the criterion lowers drawdown, Value-at-Risk, and expected shortfall relative to the mean–variance benchmark at every required return, in and out of sample, without any loss in realized portfolio return. The construction maps signed multiscale interaction structures onto resource-allocation decisions, and applies to any complex system whose components interact across heterogeneous scales with amplitude-dependent coupling.

This paper studies the investment and insurance strategies of defined-contribution (DC) pension plans under the mean-variance framework. We consider a stochastic environment with time-varying interest rates, contributions, and mortality risk. The DC plan members are allowed to decide their bond and stock allocations, as well as their life insurance coverage. Adopting the martingale approach, we derive the closed-form optimal strategies and the mean-variance efficient frontier. Further numerical analysis investigates how mortality improvements affect investment and insurance decisions, as well as the sensitivity of the optimal decision to market parameters. Our analysis suggests that longevity raises expectations of future contributions, allowing pension members to adopt a less risky investment strategy. Meanwhile, insurance strategy shifts toward early adulthood to protect the high value of future income and decreases significantly at later ages. Moreover, we conduct sensitivity analyses on the target expected wealth, market price of risk, and contribution growth. These findings provide practical guidance for pension members on investment and offer insights for the design of DC pension plans.

Language models can propose many plausible trading factors, but an autonomous research system must also allocate its evaluation budget, verify its own evidence, and preserve how each candidate was produced. We present AgonAlpha, an architecture that searches over frozen research artifacts—hypotheses, executable expressions, platform evidence, rationales, and review status—rather than formulas alone. To our knowledge, AgonAlpha is the first alpha-mining system to combine verified artifact search, a fresh-context adversarial reviewer with re-execution and veto authority, and pending-aware parallel budget allocation, together with a complete public evidence trail. Independent deployments on WorldQuant BRAIN produced SPECTACULAR-grade alphas across five users and six model backends, with Fitness reaching 9.50 and Sharpe reaching 3.48, while retaining prompt-to-expression provenance for every submission.

Automated market makers (AMMs) are typically interpreted and evaluated as decentralized exchanges. Herein, we take the perspective envisioned by Balancer that an AMM can also be viewed as a portfolio technology that programmatically enforces an economic mandate. In particular, we follow the geometric mean market maker (G3M) invariant employed by that protocol in order to enforce a target-weighted portfolio. We introduce a multi-asset fee structure to the G3M under which competitive arbitrage implements a band-rebalancing strategy with mis-weighting bounded ex ante, allowing compliance with the mandate to be verified directly from the pool's observable holdings. We then compare simulated G3M portfolios against the realized performance of VBIAX, EQL, and EDOW on annualized returns and tracking error against the portfolio mandate. Across these historical case studies, and using arbitrage-only order flow, the G3M is found to outperform the incumbent funds in both metrics for certain fee ranges.

This paper builds Path Portfolio Optimization: portfolio theory on a path-first framework in which the signature is the universal coordinate of the price path, and asks whether it survives estimation. A portfolio is a linear functional of the signature, so the control lives in a truncated tensor algebra, the covariance of signature coordinates is the non-group-like part of the expected signature — a defect form — and the whole mean–variance problem becomes a linear system in one tensor. Two structural results follow. The lift is the execution convention: the gap between the Marcus and forward lifts, contracted with portfolio weights, is Fernholz's excess growth rate exactly, so excess growth is the geometricity defect of the portfolio map. And the antisymmetric block at level two is lift-invariant pathwise, so directional signals are convention-free while variance signals and ruin are not. The empirical finding is a dimensional trade-off. With the expected signature known, quadratic path functionals raise the certainty equivalent elevenfold for a pair of assets and sixtyfold for a cross section of twenty; with it estimated, the unregularized policy is severely negative until the sample exceeds roughly six observations per parameter, and shrinkage flips from harmful in the pair to indispensable in the cross section. The entire gain sits in the symmetric block, which is convexity in the terminal increment rather than path-dependence; the path-dependent antisymmetric block earns nothing when the driver has no expected area. And the sample-size floor belongs to unstructured estimation rather than to path complexity: an estimator that fits only the generator of the driver and rebuilds the expected signature recovers almost all of the attainable value at barely one observation per parameter

We ask a representative sample to write prompts seeking spending and investing advice from LLMs, then simulate the lifetime effects of following the advice under realistic asset and labor market conditions. Applying this method to GPT-5.2, we find following the advice would move respondents toward life cycle theory: broader participation in diversified equity funds, age-declining equity shares, and larger savings buffers. Recommendations vary systematically by gender, prior AI experience, and financial literacy. For gender, two-thirds of recommended equity-share differences arise from men and women writing different prompts (demand), while one-third arise from gender labels attached to otherwise identical prompts (supply).