tianhaofu Goto Github PK

followers: 167.0 following: 1.0K repos: 782.0 gists: 1.0

Name: Tianhao Fu

Type: User

Company: Peking University

Bio: Finance AGI is coming

We continue to see exciting breakthroughs in AI. The CHATGPT in text, the PIKA in video, and the Gemini in multimodal. I'm sure every AI practitioner can feel the waves of our time building up. And I want to do my part in this era. There is a hidden voice in my heart telling me that I can also realize AGI in some small field. Now I have found my goal: AGI in quantitative finance.

My goal is to build an AGI that can replace a Quant Alpha team of about 20 people, and I'm working towards that goal. Inputting historical stages of data, our AGI can output factors, factor combination strategies, trading strategies, and fulfill the functions brought by the whole Quant Alpha team.

If you are also interested, please feel free to contact me. (Or if your team is doing the same thing, I'd love to join you!)

👋 Hi, I’m TianhaoFu(付天豪)@PKU.
👀 I’m interested in data-centric AI and Quantitative Finance.
🌱 I’m currently learning data-centric AI. Check my 2022's data-centric AI work: Tiny3D: a data-centric AI based 3D object detection service production system. Also, I'm doing research on High-frequency Strategy, Mid-frequency Strategy, and Alternative Low-frequency Strategy in US Equity, China Equity, and Crypto.
💞️ I’m looking to collaborate on data-centric AI(data privacy, data programming, data simulation, data compression, data duration, data centric ai system) and Quantitative Finance(job or other opportunities).
📫 How to reach me [email protected].

Tianhao Fu's Projects

stemgnn

Spectral Temporal Graph Neural Network (StemGNN in short) for Multivariate Time-series Forecasting

stock-price-prediction-model

Stock Price prediction using news data. The datasets used consists news and stock price data from 2008 to 2016. The polarity(Subjectivity, Objectivity, Positive, Negative, Neutral) data is gathered from the news data and further used to predict stock prices. Achieved an accuracy of 94% using XGBoost.

stocknet-code

Code for stock movement prediction from tweets and historical stock prices.

style-people

Inference code for "StylePeople: A Generative Model of Fullbody Human Avatars" paper

styleformer

A Neural Language Style Transfer framework to transfer natural language text smoothly between fine-grained language styles like formal/casual, active/passive, and many more. Created by Prithiviraj Damodaran. Open to pull requests and other forms of collaboration.