Hi! Here's Iván with this week's exciting newsletter, packed with insights and discoveries in the world of AI and finance. In this edition, we're featuring:

• 🕹️ 2 Academic Articles: Dive into groundbreaking research with actionable ideas that are reshaping our understanding of AI in finance.

• ♒ 3 FinTech Tools: Discover the top tools driving innovation in the financial sector.

• 🙄 A Bundle of Headlines: Catch up with the latest and most significant AI news in our concise headline section.

• 💊 Easy-to-Follow Video Tutorial: Enhance your skills with our step-by-step guide on using AI for stock predictions.

• 🥐 Food for Thoughts: Wrap up with some thought-provoking insights and perspectives in AI and finance.

Academic Insights

“ChatGPT Informed Graph Neural Network for Stock Movement Prediction” 

by Zhian Chen, Lei Zheng, Cheng Lu, Jialu Yuan and Di Zhu

Introduction: ☀️The paper presents a novel approach that leverages ChatGPT's graph inference capabilities to enhance Graph Neural Networks (GNN). This framework adeptly extracts evolving network structures from financial news and incorporates them into GNN for predictive tasks. This innovative approach aims to explore ChatGPT's potential for inferring dynamic network structures from temporal textual data, particularly in the financial sector, marking a significant advancement in natural language processing applications​​.

Main Results: 🍭 The model was evaluated using real-world data from the DOW 30 companies, covering the period from September 2020 to December 2022. The model consistently outperformed baseline models in weighted F1, Micro F1, and Macro F1 metrics, showing a minimum improvement of 1.8%. Additionally, portfolios constructed based on the model's outputs demonstrated higher annualized cumulative returns, reduced volatility, and maximum drawdown. These results underscore the effectiveness of the ChatGPT-informed GNN model, highlighting the promising implications of Large Language Models (LLMs) for financial data processing​​.

Actionable Practical Ideas: 🎾 The study's findings suggest actionable strategies for leveraging ChatGPT's capabilities in financial engineering. The approach of distilling evolving network structures from daily financial news and integrating them with GNN for stock movement prediction can be practically implemented. Portfolios based on this model's outputs show promise in delivering superior returns with lower volatility and drawdowns. This method offers new strategies and perspectives in the financial sector, emphasizing the practical implications of using modern Language Learning Models to infer network structures from text for enhanced financial forecasting​

“Pairs Trading Using Clustering And Deep Reinforcement Learning”

by Raktim Roychoudhury, Rahul Bhagtani, and Aditya Daftari.

Introduction: ☀️ The paper addresses the limitations of conventional pairs trading strategies, which often assume linearity in stock returns. It proposes a novel approach using unsupervised learning (agglomerative clustering) to select pairs based on fundamental and technical indicators. This method diverges from linearity, allowing for trading strategies with varying market exposure and employs Reinforcement Learning (RL) with Proximal Policy Optimization for trading the selected indices

Main Results: 🍭 methodology involves clustering equity indices using Convolutional AutoEncoders (CAE) to identify 10 latent risk factors and then applying RL for trading. This non-linear approach to pairs trading demonstrates superior performance, with 12 out of 13 pairs outperforming the benchmark S&P 500 Index in terms of annualized returns and maximum drawdown. The top three pairs showed annualized returns of 21.86%, 18.61%, and 17.61%.

Actionable Practical Ideas: 🎾 The paper's findings suggest a more dynamic and adaptable trading strategy, suitable for an ever-changing stock market. Unlike traditional pairs trading, the RL-based strategy doesn't require retraining even in volatile market conditions. The approach is market-neutral, generating alpha consistently. For practical application, it's recommended to redefine the cumulative portfolio returns as the reward structure in the RL framework to tailor trading strategies according to specific risk profiles. Future work could include robustness tests, exploring ensembles of RL algorithms, and expanding the strategy to more clustered pairs.

Week’s AI FinTech Picks

• 📢 Zest.ai: ZestFinance offers a machine learning underwriting platform to enhance approval rates, reduce credit losses, and improve underwriting. Their credit-decisioning platform is designed to help lenders predict credit risk, increase revenues, and ensure compliance.

• 📢 Qraft: QraftTechnologies specializes in AI investment solutions, providing AI solutions customized for various uses, including the finance sector.

• 📢 Cleo: Cleo develops an app and chatbot that function as a financial assistant, offering users personalized financial management assistance.

AI-Essentials: Step-by-Step Tutorial

🚀 How to apply Random Forest for stock prediction step-by-step in this youtube tutorial, “Machine Learning Stock Prediction Using Random Forest Regressor”.

AI Buzz: The Week’s Top Stories

• ☀️ Microsoft's AI Expansion: Microsoft just unveiled three new AI tools – Copilot Azure, Copilot for Service, and Copilot Studio – aiming to make big waves in AI tech​​.

• ☄ Stability AI Shake-Up: Ed Newton-Rex, the head of audio at Stability AI, resigns over concerns about using copyrighted works in AI training​​.

• 🦄 Microsoft's AI Chips: Meet Maia and Cobalt, Microsoft's latest custom AI chips, set to bolster their AI strategy and data center operations​​.

• 🅿 ChatGPT Plus on Pause: OpenAI's ChatGPT Plus signups take a brief break due to overwhelming demand following its latest feature rollout​​.

• 💪 China's AI Leap: China's new open-source AI model is turning heads with its ability to process up to 200,000 tokens, outdoing popular models like ChatGPT​​.

AI-Finance: Food for Thought

The integration of artificial intelligence (AI) in the finance industry is leading to significant transformations across various sectors, including stock investing, decision-making, and risk management.

From the Institute of Analytics (IoA), we learn that AI is revolutionizing the finance industry by enhancing computational capabilities through hybrid cloud solutions, enabling financial institutions to make data-driven decisions and gain valuable market insights. The diverse applications of AI in finance include fraud detection, customer experience improvements, and policy enforcement automation. Notably, AI also contributes to creating proxy/synthetic data, stimulating growth and developing robust AI algorithms​​.

In the realm of stock investing, UK-based 3AI is leveraging AI to enhance the analysis of big data, aiming to achieve Alpha in investments. Their products, Smart Alpha Insights and the US Smart Alpha Index, use Deep Factor AI to assess hundreds of data factors per stock, thereby enabling investors to make informed decisions​​​​​​​​​​​​​​​​.

Lastly, the report from SiliconANGLE highlights how financial institutions are utilizing the vast amounts of customer data they collect for AI-powered decision-making. Companies like FICO and Infosys Ltd. are employing AI tools to improve customer service, risk management, and credit decisions. AI is also being used to analyze sentiment and provide hyper-personalized services, significantly enhancing customer engagement and marketing strategies​​.

If you're enjoying our newsletter and want to support us, please recommend it to anyone you know who's interested in AI and Finance. Your referrals are the biggest compliment and help us grow! 🌟🤖💼