همایش بین المللی هوش مصنوعی

صفحه اصلی / دومین همایش بین المللی هوش مصنوعی

Enhanced Diffusion Policy for Robust Multimodal Stock Trading through the Integration of Technical Indicators and Fundamental News

نویسندگان :

Amirali Vakili¹ Mahdi Shahbazi Khojasteh² Mahan Veisi³ Armin Salimi-Badr⁴

1- دانشگاه شهید بهشتی 2- دانشگاه شهید بهشتی 3- دانشگاه شهید بهشتی 4- دانشگاه شهید بهشتی

کلمات کلیدی :

Automated Trading،Deep Reinforcement Learning،Diffusion Policy,،Stock Market,،Q-Learning

چکیده :

The pursuit of automated trading strategies that can consist-ently outperform market benchmarks remains a significant challenge in financial markets. Although Deep Reinforcement Learning (DRL) has shown promise for stock trading, conven-tional DRL approaches often rely on restrictive policy represen-tations that fail to capture the multimodal nature of optimal trading decisions. To address this limitation, this paper applies the Enhanced Diffusion-QL (EDiffusion-QL) framework to automated stock trading on U.S. equity markets, integrating it with a Bidirectional Long Short-Term Memory (Bi-LSTM) state encoder that processes both technical indicators and fi-nancial news embeddings. With EDiffusion-QL, trading ac-tions are generated by using a diffusion-based policy and twin Q-learning critics for robust value estimation. We evaluate our proposed method on a comprehensive dataset comprising 30 U.S. stocks spanning from 2009 to 2020, comparing perfor-mance against state-of-the-art DRL baselines. The results demonstrate that our model achieves superior performance, attaining the highest cumulative return, Sharpe ratio, and av-erage profitability per trade. By modeling multi-modal action distributions, EDiffusion-QL establishes a more robust and profitable approach to portfolio optimization and automated trading.

لیست مقالات

لیست مقالات بایگانی شده

Examining the Role of Artificial Intelligence in Enhancing Educational Equity: A Systematic Review

Ali Rahmanipur - Shakila Mohammadi

Attention-Based Noise Reduction for Surface-Electromyography: A Novel Method for Enhanced Signal Quality in Clinical Diagnostics

Seyyed Ali Zendehbad - Abdollah PourMottaghi - Marzieh Allami Sanjani

Application of machine learning algorithms in the prediction of the reliability of post-tensioned concrete members

Pooria Poorahad A. - Mahmoud R. Shiravand - Mahtab Ebadati

Title Generation for the Qur'anic chapters by summarizing them

Masoume Maleki - Alireza Talebpour - Mostafa Moradi

Fibroglandular Tissue Classification in Breast MRI: A Comparative Study of Automated Decision Strategies

Meysam Khalaj - Arvin Arian - Ala Torabi - Nasrin Ahmadinejad - Masoumeh Gity - Seyedeh Nooshin Miratashi Yazdi - Mohammad Pooya Afshari - Melika Sadeghi Tabrizi - Hamid Soltanian-Zadeh

Creating a Foundation for Dynamic Difficulty Adjustment within PCG of games using Imitation Learning

Navid Siamakmanesh - Arian Ganji - Monireh Abdoos - Mojtaba Vahidi-Asl

Vision-based Static Detection of Windows Malware with Rule-Augmented Feature Maps

Mahdi Seyfipoor - Mohammad Mahdi Eskandari - Siamak Mohammadi

AI Hugs the Theology of Peace, but friendly or backstabbing?

Aliasghar Ahmadi

Applications of Artificial Intelligence in Upstream Oil and Gas Industries

Parisa Salimi - Zahra Nargessi - Fatemeh Alamolhoda

Early Detection of Congestive Heart Failure in Coronary Artery Disease Patients Using ECG Based Hybrid CNN-LSTM Model

Seyyed Ali Zendehbad - Farinaz Azari - Hadi Dehbovid

بیشتر