Context-Aware Language Models for Forecasting Market Impact from Sequences of Financial News ArXiv ID: 2509.12519 “View on arXiv” Authors: Ross Koval, Nicholas Andrews, Xifeng Yan Abstract Financial news plays a critical role in the information diffusion process in financial markets and is a known driver of stock prices. However, the information in each news article is not necessarily self-contained, often requiring a broader understanding of the historical news coverage for accurate interpretation. Further, identifying and incorporating the most relevant contextual information presents significant challenges. In this work, we explore the value of historical context in the ability of large language models to understand the market impact of financial news. We find that historical context provides a consistent and significant improvement in performance across methods and time horizons. To this end, we propose an efficient and effective contextualization method that uses a large LM to process the main article, while a small LM encodes the historical context into concise summary embeddings that are then aligned with the large model’s representation space. We explore the behavior of the model through multiple qualitative and quantitative interpretability tests and reveal insights into the value of contextualization. Finally, we demonstrate that the value of historical context in model predictions has real-world applications, translating to substantial improvements in simulated investment performance. ...

Dynamic Factor Models with Forward-Looking Views ArXiv ID: 2509.11528 “View on arXiv” Authors: Anas Abdelhakmi, Andrew E. B. Lim Abstract Prediction models calibrated using historical data may forecast poorly if the dynamics of the present and future differ from observations in the past. For this reason, predictions can be improved if information like forward looking views about the state of the system are used to refine the forecast. We develop an approach for combining a dynamic factor model for risky asset prices calibrated on historical data, and noisy expert views of future values of the factors/covariates in the model, and study the implications for dynamic portfolio choice. By exploiting the graphical structure linking factors, asset prices, and views, we derive closed-form expressions for the dynamics of the factor and price processes after conditioning on the views. For linear factor models, the price process becomes a time-inhomogeneous affine process with a new covariate formed from the views. We establish a novel theoretical connection between the conditional factor process and a process we call a Mean-Reverting Bridge (MrB), an extension of the classical Brownian bridge. We derive the investor’s optimal portfolio strategy and show that views influence both the myopic mean-variance term and the intertemporal hedge. The optimal dynamic portfolio when the long-run mean of the expected return is uncertain and learned online from data is also derived. More generally, our framework offers a generalizable approach for embedding forward-looking information about covariates in a dynamic factor model. ...

Meta-Learning Neural Process for Implied Volatility Surfaces with SABR-induced Priors ArXiv ID: 2509.11928 “View on arXiv” Authors: Jirong Zhuang, Xuan Wu Abstract We treat implied volatility surface (IVS) reconstruction as a learning problem guided by two principles. First, we adopt a meta-learning view that trains across trading days to learn a procedure that maps sparse option quotes to a full IVS via conditional prediction, avoiding per-day calibration at test time. Second, we impose a structural prior via transfer learning: pre-train on SABR-generated dataset to encode geometric prior, then fine-tune on historical market dataset to align with empirical patterns. We implement both principles in a single attention-based Neural Process (Volatility Neural Process, VolNP) that produces a complete IVS from a sparse context set in one forward pass. On SPX options, the VolNP outperforms SABR, SSVI, and Gaussian process. Relative to an ablation trained only on market data, the SABR-induced prior reduces RMSE by about 40% and suppresses large errors, with pronounced gains at long maturities where quotes are sparse. The resulting model is fast (single pass), stable (no daily recalibration), and practical for deployment at scale. ...

Reinforcement Learning-Based Market Making as a Stochastic Control on Non-Stationary Limit Order Book Dynamics ArXiv ID: 2509.12456 “View on arXiv” Authors: Rafael Zimmer, Oswaldo Luiz do Valle Costa Abstract Reinforcement Learning has emerged as a promising framework for developing adaptive and data-driven strategies, enabling market makers to optimize decision-making policies based on interactions with the limit order book environment. This paper explores the integration of a reinforcement learning agent in a market-making context, where the underlying market dynamics have been explicitly modeled to capture observed stylized facts of real markets, including clustered order arrival times, non-stationary spreads and return drifts, stochastic order quantities and price volatility. These mechanisms aim to enhance stability of the resulting control agent, and serve to incorporate domain-specific knowledge into the agent policy learning process. Our contributions include a practical implementation of a market making agent based on the Proximal-Policy Optimization (PPO) algorithm, alongside a comparative evaluation of the agent’s performance under varying market conditions via a simulator-based environment. As evidenced by our analysis of the financial return and risk metrics when compared to a closed-form optimal solution, our results suggest that the reinforcement learning agent can effectively be used under non-stationary market conditions, and that the proposed simulator-based environment can serve as a valuable tool for training and pre-training reinforcement learning agents in market-making scenarios. ...

Sentiment Feedback in Equity Markets: Asymmetries, Retail Heterogeneity, and Structural Calibration ArXiv ID: 2509.11970 “View on arXiv” Authors: Lucas Marques Sneller Abstract We study how sentiment shocks propagate through equity returns and investor clientele using four independent proxies with sign-aligned kappa-rho parameters. A structural calibration links a one standard deviation innovation in sentiment to a pricing impact of 1.06 basis points with persistence parameter rho = 0.940, yielding a half-life of 11.2 months. The impulse response peaks around the 12-month horizon, indicating slow-moving amplification. Cross-sectionally, a simple D10-D1 portfolio earns 4.0 basis points per month with Sharpe ratios of 0.18-0.85, consistent with tradable exposure to the sentiment factor. Three regularities emerge: (i) positive sentiment innovations transmit more strongly than negative shocks (amplification asymmetry); (ii) effects are concentrated in retail-tilted and non-optionable stocks (clientele heterogeneity); and (iii) responses are state-dependent across volatility regimes - larger on impact in high-VIX months but more persistent in low-VIX months. Baseline time-series fits are parsimonious (R2 ~ 0.001; 420 monthly observations), yet the calibrated dynamics reconcile modest impact estimates with sizable long-short payoffs. Consistent with Miller (1977), a one standard deviation sentiment shock has 1.72-8.69 basis points larger effects in low-breadth stocks across horizons of 1-12 months, is robust to institutional flows, and exhibits volatility state dependence - larger on impact but less persistent in high-VIX months, smaller on impact but more persistent in low-VIX months. ...

Mamba Outpaces Reformer in Stock Prediction with Sentiments from Top Ten LLMs ArXiv ID: 2510.01203 “View on arXiv” Authors: Lokesh Antony Kadiyala, Amir Mirzaeinia Abstract The stock market is extremely difficult to predict in the short term due to high market volatility, changes caused by news, and the non-linear nature of the financial time series. This research proposes a novel framework for improving minute-level prediction accuracy using semantic sentiment scores from top ten different large language models (LLMs) combined with minute interval intraday stock price data. We systematically constructed a time-aligned dataset of AAPL news articles and 1-minute Apple Inc. (AAPL) stock prices for the dates of April 4 to May 2, 2025. The sentiment analysis was achieved using the DeepSeek-V3, GPT variants, LLaMA, Claude, Gemini, Qwen, and Mistral models through their APIs. Each article obtained sentiment scores from all ten LLMs, which were scaled to a [“0, 1”] range and combined with prices and technical indicators like RSI, ROC, and Bollinger Band Width. Two state-of-the-art such as Reformer and Mamba were trained separately on the dataset using the sentiment scores produced by each LLM as input. Hyper parameters were optimized by means of Optuna and were evaluated through a 3-day evaluation period. Reformer had mean squared error (MSE) or the evaluation metrics, and it should be noted that Mamba performed not only faster but also better than Reformer for every LLM across the 10 LLMs tested. Mamba performed best with LLaMA 3.3–70B, with the lowest error of 0.137. While Reformer could capture broader trends within the data, the model appeared to over smooth sudden changes by the LLMs. This study highlights the potential of integrating LLM-based semantic analysis paired with efficient temporal modeling to enhance real-time financial forecasting. ...

Quantum and Classical Machine Learning in Decentralized Finance: Comparative Evidence from Multi-Asset Backtesting of Automated Market Makers ArXiv ID: 2510.15903 “View on arXiv” Authors: Chi-Sheng Chen, Aidan Hung-Wen Tsai Abstract This study presents a comprehensive empirical comparison between quantum machine learning (QML) and classical machine learning (CML) approaches in Automated Market Makers (AMM) and Decentralized Finance (DeFi) trading strategies through extensive backtesting on 10 models across multiple cryptocurrency assets. Our analysis encompasses classical ML models (Random Forest, Gradient Boosting, Logistic Regression), pure quantum models (VQE Classifier, QNN, QSVM), hybrid quantum-classical models (QASA Hybrid, QASA Sequence, QuantumRWKV), and transformer models. The results demonstrate that hybrid quantum models achieve superior overall performance with 11.2% average return and 1.42 average Sharpe ratio, while classical ML models show 9.8% average return and 1.47 average Sharpe ratio. The QASA Sequence hybrid model achieves the highest individual return of 13.99% with the best Sharpe ratio of 1.76, demonstrating the potential of quantum-classical hybrid approaches in AMM and DeFi trading strategies. ...

RegimeFolio: A Regime Aware ML System for Sectoral Portfolio Optimization in Dynamic Markets ArXiv ID: 2510.14986 “View on arXiv” Authors: Yiyao Zhang, Diksha Goel, Hussain Ahmad, Claudia Szabo Abstract Financial markets are inherently non-stationary, with shifting volatility regimes that alter asset co-movements and return distributions. Standard portfolio optimization methods, typically built on stationarity or regime-agnostic assumptions, struggle to adapt to such changes. To address these challenges, we propose RegimeFolio, a novel regime-aware and sector-specialized framework that, unlike existing regime-agnostic models such as DeepVol and DRL optimizers, integrates explicit volatility regime segmentation with sector-specific ensemble forecasting and adaptive mean-variance allocation. This modular architecture ensures forecasts and portfolio decisions remain aligned with current market conditions, enhancing robustness and interpretability in dynamic markets. RegimeFolio combines three components: (i) an interpretable VIX-based classifier for market regime detection; (ii) regime and sector-specific ensemble learners (Random Forest, Gradient Boosting) to capture conditional return structures; and (iii) a dynamic mean-variance optimizer with shrinkage-regularized covariance estimates for regime-aware allocation. We evaluate RegimeFolio on 34 large cap U.S. equities from 2020 to 2024. The framework achieves a cumulative return of 137 percent, a Sharpe ratio of 1.17, a 12 percent lower maximum drawdown, and a 15 to 20 percent improvement in forecast accuracy compared to conventional and advanced machine learning benchmarks. These results show that explicitly modeling volatility regimes in predictive learning and portfolio allocation enhances robustness and leads to more dependable decision-making in real markets. ...

Trading-R1: Financial Trading with LLM Reasoning via Reinforcement Learning ArXiv ID: 2509.11420 “View on arXiv” Authors: Yijia Xiao, Edward Sun, Tong Chen, Fang Wu, Di Luo, Wei Wang Abstract Developing professional, structured reasoning on par with human financial analysts and traders remains a central challenge in AI for finance, where markets demand interpretability and trust. Traditional time-series models lack explainability, while LLMs face challenges in turning natural-language analysis into disciplined, executable trades. Although reasoning LLMs have advanced in step-by-step planning and verification, their application to risk-sensitive financial decisions is underexplored. We present Trading-R1, a financially-aware model that incorporates strategic thinking and planning for comprehensive thesis composition, facts-grounded analysis, and volatility-adjusted decision making. Trading-R1 aligns reasoning with trading principles through supervised fine-tuning and reinforcement learning with a three-stage easy-to-hard curriculum. Training uses Tauric-TR1-DB, a 100k-sample corpus spanning 18 months, 14 equities, and five heterogeneous financial data sources. Evaluated on six major equities and ETFs, Trading-R1 demonstrates improved risk-adjusted returns and lower drawdowns compared to both open-source and proprietary instruction-following models as well as reasoning models. The system generates structured, evidence-based investment theses that support disciplined and interpretable trading decisions. Trading-R1 Terminal will be released at https://github.com/TauricResearch/Trading-R1. ...

Why Bonds Fail Differently? Explainable Multimodal Learning for Multi-Class Default Prediction ArXiv ID: 2509.10802 “View on arXiv” Authors: Yi Lu, Aifan Ling, Chaoqun Wang, Yaxin Xu Abstract In recent years, China’s bond market has seen a surge in defaults amid regulatory reforms and macroeconomic volatility. Traditional machine learning models struggle to capture financial data’s irregularity and temporal dependencies, while most deep learning models lack interpretability-critical for financial decision-making. To tackle these issues, we propose EMDLOT (Explainable Multimodal Deep Learning for Time-series), a novel framework for multi-class bond default prediction. EMDLOT integrates numerical time-series (financial/macroeconomic indicators) and unstructured textual data (bond prospectuses), uses Time-Aware LSTM to handle irregular sequences, and adopts soft clustering and multi-level attention to boost interpretability. Experiments on 1994 Chinese firms (2015-2024) show EMDLOT outperforms traditional (e.g., XGBoost) and deep learning (e.g., LSTM) benchmarks in recall, F1-score, and mAP, especially in identifying default/extended firms. Ablation studies validate each component’s value, and attention analyses reveal economically intuitive default drivers. This work provides a practical tool and a trustworthy framework for transparent financial risk modeling. ...