State of the Field
Recent advancements in financial AI are focusing on enhancing model robustness and decision-making accuracy in dynamic markets. Researchers are developing adaptive dataflow systems that address the limitations of static historical data, enabling models to evolve alongside market conditions. This shift is crucial for improving risk-adjusted returns and mitigating overfitting issues. Concurrently, probabilistic foundation models are being refined to provide better uncertainty quantification, allowing for more informed forecasting in volatile environments. The introduction of benchmarks that evaluate financial recommendations beyond mere behavioral imitation is also gaining traction, emphasizing the importance of aligning model outputs with long-term investment goals. Furthermore, innovations in multi-agent systems are optimizing trading strategies by breaking down complex tasks into manageable components, leading to improved performance in real-world scenarios. Collectively, these efforts are poised to tackle significant commercial challenges in finance, such as enhancing predictive accuracy and aligning automated systems with investor preferences.
Papers
1–8 of 8History Is Not Enough: An Adaptive Dataflow System for Financial Time-Series Synthesis
In quantitative finance, the gap between training and real-world performance-driven by concept drift and distributional non-stationarity-remains a critical obstacle for building reliable data-driven s...
ProbFM: Probabilistic Time Series Foundation Model with Uncertainty Decomposition
Time Series Foundation Models (TSFMs) have emerged as a promising approach for zero-shot financial forecasting, demonstrating strong transferability and data efficiency gains. However, their adoption ...
Conv-FinRe: A Conversational and Longitudinal Benchmark for Utility-Grounded Financial Recommendation
Most recommendation benchmarks evaluate how well a model imitates user behavior. In financial advisory, however, observed actions can be noisy or short-sighted under market volatility and may conflict...
Bridging the Arithmetic Gap: The Cognitive Complexity Benchmark and Financial-PoT for Robust Financial Reasoning
While Large Language Models excel at semantic tasks, they face a critical bottleneck in financial quantitative reasoning, frequently suffering from "Arithmetic Hallucinations" and a systemic failure m...
Cross-Sectional Asset Retrieval via Future-Aligned Soft Contrastive Learning
Asset retrieval--finding similar assets in a financial universe--is central to quantitative investment decision-making. Existing approaches define similarity through historical price patterns or secto...
TradeFM: A Generative Foundation Model for Trade-flow and Market Microstructure
Foundation models have transformed domains from language to genomics by learning general-purpose representations from large-scale, heterogeneous data. We introduce TradeFM, a 524M-parameter generative...
Kinematic Tokenization: Optimization-Based Continuous-Time Tokens for Learnable Decision Policies in Noisy Time Series
Transformers are designed for discrete tokens, yet many real-world signals are continuous processes observed through noisy sampling. Discrete tokenizations (raw values, patches, finite differences) ca...
Portfolio Reinforcement Learning with Scenario-Context Rollout
Market regime shifts induce distribution shifts that can degrade the performance of portfolio rebalancing policies. We propose macro-conditioned scenario-context rollout (SCR) that generates plausible...