Search arXiv⌕ Search
Home/Archive/arXiv

arXiv · 2312.14203

Shai: A large language model for asset management

Zhongyang Guo·Guanran Jiang·Zhongdan Zhang·Peng Li·Zhefeng Wang·Yinchun Wang
View original source ↗Read source document ↗Export metadata ↓

Abstract

This paper introduces "Shai" a 10B level large language model specifically designed for the asset management industry, built upon an open-source foundational model. With continuous pre-training and fine-tuning using a targeted corpus, Shai demonstrates enhanced performance in tasks relevant to its domain, outperforming baseline models. Our research includes the development of an innovative evaluation framework, which integrates professional qualification exams, tailored tasks, open-ended question answering, and safety assessments, to comprehensively assess Shai's capabilities. Furthermore, we discuss the challenges and implications of utilizing large language models like GPT-4 for performance assessment in asset management, suggesting a combination of automated evaluation and human judgment. Shai's development, showcasing the potential and versatility of 10B-level large language models in the financial sector with significant performance and modest computational requirements, hopes to provide practical insights and methodologies to assist industry peers in their similar endeavors.

Explore related subjects

q-fin.PMcs.CLcs.LG

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Zhongyang Guo, Guanran Jiang, Zhongdan Zhang, Peng Li, Zhefeng Wang, Yinchun Wang. 2023-12-21. Shai: A large language model for asset management. https://arxiv.org/abs/2312.14203

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

arXiv2026

A Declining CVaR Glidepath Framework for Target-Date Fund Design with an Application to the Chilean Pension System

We propose a framework for designing Target-Date Funds (TDFs) around an explicit return objective while controlling risk directly at the portfolio level through a declining Conditional Value-at-Risk (CVaR) constraint. In this approach, the regulator or sponsor specifies a CVaR glidepath that gives the portfolio manager enough flexibility to reach a target return with a reasonably high probability. The target return is determined exogenously from pension-design inputs such as retirement age, contribution rate, working years, life expectancy, and replacement-rate goals. This differs from conventional TDF design, where age-dependent asset-class limits are set without an explicit link to a required return. A key feature of the method is that it does not assume the manager selects an optimal portfolio each period. Instead, each month the manager draws an allocation from the set of portfolios satisfying the CVaR constraint. This yields a conservative evaluation of each glidepath: success probabilities are averages over admissible allocations, rather than best-case outcomes. We introduce two figures of merit: the probability of meeting the target return and the cumulative risk assumed over the life of the TDF. As a proof of concept, we apply the framework to Chile's 2025 pension reform using nine Chilean and global asset classes and a 40-year accumulation horizon. The results show that the transition age at which risk starts to decline is the most consequential design parameter, and that contribution density acts as a hard constraint: below a critical threshold, portfolio design alone cannot compensate for structurally low contributions. The framework is general and can be applied to any TDF designed around an explicit return objective.

q-fin.PM↗
arXiv2026

Active Portfolio Management in Concentrated Equity Markets

The equal-weighted portfolio is a passive, rule-based strategy that has historically been difficult to outperform, delivering higher returns than the capitalization-weighted "market" benchmark across many markets and periods. Stochastic portfolio theory (SPT) reveals that this relative performance is regime dependent, with the equal-weighted portfolio underperforming during periods of increasing market concentration and high correlations, particularly market bubbles. These observations have motivated us to formulate and solve a stochastic control problem in which an investor actively allocates between the equal-weighted and market portfolios. The investor bases their allocation decisions on forecasts made under a flexible stochastic diversity--dispersion (SDD) model. Using a quadratic surrogate for implementation frictions, we characterize the optimal allocation through a linear forward--backward SDE and obtain an explicit "aiming in front of a moving target'' representation of the optimal trading rate, in the spirit of Gârleanu and Pedersen. The penalty parameters are calibrated in sample to match the cumulative wealth effect of proportional transaction costs, while out-of-sample performance is evaluated with those costs deducted directly from portfolio wealth. Using historical S&P 500 data, we show that a mean-reverting SDD specification reproduces several empirical features of market diversity and dispersion. In out-of-sample backtests from 1995 to 2024, the resulting strategies deliver higher cumulative net returns than both the equal-weighted and market portfolios, and higher information ratios than the equal-weighted portfolio after 15-basis-point proportional transaction costs.

q-fin.PM↗
arXiv2026

Financially Guided Deep Portfolio Optimization

Portfolio optimization in real-world financial markets is notoriously difficult due to non-stationarity, noisy data, and high transaction costs. Standard predict-then-optimize methods first forecast returns and then solve for weights, compounding prediction errors and often failing under regime shifts. We propose an end-to-end framework that directly optimizes differentiable surrogates of key financial metrics (Sharpe ratio, Omega ratio, Conditional Value-at-Risk, and risk parity), allowing neural networks to learn portfolio weights via backpropagation. Our expanding-window walk-forward procedure, applied to 50 S&P 500 stocks from 2007 to 2023, incorporates realistic bid-ask spread costs and rebalances quarterly. On the challenging out-of-sample test period (2022-2023), the best model, an AttentionLSTM with the Omega-CVaR-RiskParity loss, achieves an annualized Sharpe of 0.29 and a total compounded return of +7.86%, while the S&P 500 delivers -4.52% total compounded return and an annualized Sharpe of -0.02. This outperforms the S&P 500 by 12.38 percentage points, while keeping tail risk (CVaR) nearly unchanged. The framework outperforms the equal-weight portfolio, S&P 500, and traditional methods (MVP, HRP, NCO, ERC), demonstrating that embedding financial objectives directly into model training yields robust, economically meaningful outperformance even in adverse market conditions.

q-fin.PM↗