Paper: arXiv 2402.09746

Abstract

Recently, we introduced a new paradigm for alpha mining in the realm of quantitative investment, developing a new interactive alpha mining system framework, Alpha-GPT. This system is centered on iterative Human-AI interaction based on large language models, introducing a Human-in-the-Loop approach to alpha discovery. In this paper, we present the next-generation Alpha-GPT 2.0 \footnote{“Draft. Work in progress”}, a quantitative investment framework that further encompasses crucial modeling and analysis phases in quantitative investment. This framework emphasizes the iterative, interactive research between humans and AI, embodying a Human-in-the-Loop strategy throughout the entire quantitative investment pipeline. By assimilating the insights of human researchers into the systematic alpha research process, we effectively leverage the Human-in-the-Loop approach, enhancing the efficiency and precision of quantitative investment research.

Complexity vs Empirical Score

  • Math Complexity: 2.5/10
  • Empirical Rigor: 3.0/10
  • Quadrant: Philosophers — conceptual discussion, limited math and data

Why this score: The paper focuses on a framework and system architecture for Human-AI interaction using LLMs, which involves more conceptual and workflow design than advanced mathematical modeling. While it mentions backtesting and benchmarking, the excerpt lacks specific statistical metrics, code, or implementation details, suggesting it is more conceptual than data-heavy.

Research Flowchart

  flowchart TD
  A["Research Goal:<br>Enhance Alpha Mining &<br>Quantitative Pipeline"] --> B["Methodology:<br>Alpha-GPT 2.0 Framework<br>Human-in-the-Loop AI"]
  
  B --> C["Data Inputs:<br>Market Data &<br>Financial Reports"]
  
  C --> D["Computational Process:<br>Iterative Human-AI Interaction<br>LLM Reasoning & Backtesting"]
  
  D --> E["Key Outcomes:<br>Improved Alpha Discovery<br>Efficient Modeling & Analysis<br>Enhanced Research Precision"]

Cited References (78)

  1. 101 Formulaic Alphas (2016) — link
  2. Alpha- GPT : Human - AI Interactive Alpha Mining for Quantitative Investment (2023) — link
  3. A Survey on Gaps between Mean - Variance Approach and Exponential Growth Rate Approach for Portfolio Optimization (2022) — link
  4. Introduction to Alpha Design (2019) — link
  5. The arbitrage theory of capital asset pricing (1976) — link
  6. The Cross - Section of Expected Stock Returns (1992) — link
  7. Alpha Discovery Neural Network based on Prior Knowledge (2020) — link
  8. Bayesian Symbolic Regression (2020) — link
and 70 more
  1. AutoAlpha : an Efficient Hierarchical Evolutionary Algorithm for Mining Alpha Factors in Quantitative Investment (2020) — link
  2. Neural network-based automatic factor construction (2020) — link
  3. Deep symbolic regression: Recovering mathematical expressions from data via risk-seeking policy gradients (2021) — link
  4. Grammar Variational Autoencoder (2017) — link
  5. Learning a Formula of Interpretability to Learn Interpretable Formulas (2020) — link
  6. Discovering Symbolic Models from Deep Learning with Inductive Biases (2020) — link
  7. Fast Neural Models for Symbolic Regression at Scale (2021) — link
  8. Contemporary Symbolic Regression Methods and their Relative Performance (2021) — link
  9. SymbolicGPT : A Generative Transformer Model for Symbolic Regression (2021) — link
  10. AlphaEvolve : A Learning Framework to Discover Novel Alphas in Quantitative Investment (2021) — link
  11. Symbolic Regression via Deep Reinforcement Learning Enhanced Genetic Programming Seeding (2022) — link
  12. A Probabilistic and Multi - Objective Analysis of Lexicase Selection and ε- Lexicase Selection (2019) — link
  13. Solving the exponential growth of symbolic regression trees in geometric semantic genetic programming (2018) — link
  14. Neural- Guided Symbolic Regression with Asymptotic Constraints (2019) — link
  15. Interaction– Transformation Evolutionary Algorithm for Symbolic Regression (2021) — link
  16. AI Feynman : A physics-inspired method for symbolic regression (2020) — link
  17. Symbolic regression by uniform random global search (2019) — link
  18. AI Feynman 2.0: Pareto -optimal symbolic regression exploiting graph modularity (2020) — link
  19. A Seq2Seq approach to Symbolic Regression (2022) — link
  20. A Reinforcement Learning Approach to Domain - Knowledge Inclusion Using Grammar Guided Symbolic Regression (2022) — link
  21. Integration of Neural Network - Based Symbolic Regression in Deep Learning for Scientific Discovery (2021) — link
  22. Symbolic Learning to Optimize : Towards Interpretability and Scalability (2022) — link
  23. End-to-end Symbolic Regression with Transformers (2022) — link
  24. AI - Assisted Discovery of Quantitative and Formal Models in Social Science (2022) — link
  25. SymFormer : End -to-end symbolic regression using transformer-based architecture (2022) — link
  26. Dimensionless machine learning: Imposing exact units equivariance (2022) — link
  27. Probabilistic grammars for equation discovery (2021) — link
  28. Exhaustive Symbolic Regression (2022) — link
  29. Neural Symbolic Regression that scales (2021) — link
  30. Rethinking Symbolic Regression : Morphology and Adaptability in the Context of Evolutionary Algorithms (2023) — link
  31. Symbolic Physics Learner : Discovering governing equations via Monte Carlo tree search (2023) — link
  32. Transformer-based model for symbolic regression via joint supervised learning (2023) — link
  33. Forecasting the Crude Oil Spot Price with Bayesian Symbolic Regression (2023) — link
  34. A Unified Framework for Deep Symbolic Regression (2022) — link
  35. Discovering governing equations from data by sparse identification of nonlinear dynamical systems (2016) — link
  36. Controllable Neural Symbolic Regression (2023) — link
  37. Combining data and theory for derivable scientific discovery with AI - Descartes (2023) — link
  38. Learning where and when to reason in neuro-symbolic inference (2023) — link
  39. Deep Generative Symbolic Regression (2023) — link
  40. Deep Generative Symbolic Regression with Monte - Carlo - Tree - Search (2023) — link
  41. Controllable Neural Symbolic Regression (2023) — link
  42. Generating Synergistic Formulaic Alpha Collections via Reinforcement Learning — → 2306.12964
  43. HXPY : A High - Performance Data Processing Package for Financial Time - Series Data (2023) — link
  44. Taming the Factor Zoo : A Test of New Factors (2020) — link
  45. A Unified Approach to Interpreting Model Predictions (2017) — link
  46. " Why Should I Trust You ?": Explaining the Predictions of Any Classifier (2016) — link
  47. Foundations of Technical Analysis : Computational Algorithms , Statistical Inference , and Empirical Implementation (2000) — link
  48. A Survey of Large Language Models (2023) — link
  49. Rethinking the Role of Demonstrations : What Makes In - Context Learning Work ? (2022) — link
  50. Pre-train, Prompt , and Predict : A Systematic Survey of Prompting Methods in Natural Language Processing (2021) — link
  51. Deep Reinforcement Learning from Human Preferences (2017) — link
  52. Quant 4.0: Engineering Quantitative Investment with Automated , Explainable and Knowledge -driven Artificial Intelligence (2022) — link
  53. Similarity Search in High Dimensions via Hashing (1999)
  54. R. N . Elliott ’s Masterworks : The Definitive Collection (2005) — link
  55. Emergent Abilities of Large Language Models (2022) — link
  56. ReAct : Synergizing Reasoning and Acting in Language Models (2022) — link
  57. Tree of Thoughts : Deliberate Problem Solving with Large Language Models (2023) — link
  58. Chain-of- Thought Prompting Elicits Reasoning in Large Language Models (2022) — link
  59. Toolformer: Language Models Can Teach Themselves to Use Tools (2023) — link
  60. LLM Powered Autonomous Agents (2023) — link
  61. Language Models are Few - Shot Learners (2020) — link
  62. ZeRO : Memory Optimizations Toward Training Trillion Parameter Models (2020) — link
  63. Deep Compression : Compressing Deep Neural Network with Pruning , Trained Quantization and Huffman Coding (2016) — link
  64. Computer Architecture , Fifth Edition : A Quantitative Approach (2011)
  65. NVIDIA CUDA C Programming Guide (2010)
  66. Attention is all you need (2017)
  67. ABIDES - Gym : Gym Environments for Multi - Agent Discrete Event Simulation and Application to Financial Markets (2021) — link
  68. slundberg/shap: A game theoretic approach to explain the output of any machine learning model. — link
  69. How Robust are Limit Order Book Representations under Data Perturbation ? (2021)
  70. Fully Sharded Data Parallel : faster AI training with fewer GPUs (2021) — link

1 of 78 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2518 linked papers.