Horse Race Tracking - TAS Alpha
Summary
TAS Alpha is an international company that builds quantitative models and proprietary strategies to support the trading industry. They needed an AI-powered solution that could support them in gathering and extracting various insights to drive race predictions based on factual data.
About the Customer
The Advantage Strategy (TAS) is the Hong Kong technology arm of an international company that operates in a unique niche market that has a worldwide turnover of more than USD $50 billion per annum. The main function of the company is to build quantitative models and proprietary strategies to support the trading system.
-
Team composition
4 members
-
Client name
TAS Alpha
-
Expertise used
Machine Learning, Computer Vision, and Deep Learning
-
Duration
12 Weeks
-
Services provided
Model training, On-Premise Deployment
-
Region
North America
-
Industry
Financial Services
Solution
Folio3 AI developed a Deep Learning solution for TAS Alpha that would take video inputs from the TAS Alpha platform that had video feeds of live races and process them on the algorithms trained on data provided by Tas Alpha to extract insights and return outputs in JSON format to the platform for further analysis.
AI Model Customization
Deep learning algorithms are customized and trained on client data sets for specific data outputs. The model processes a video clip and generates the output data file in .json file format with all the data extracted through processing the video clip.
Model Optimization
The AI model was fine-tuned and trained on the data set provided by the client for accurate analysis and continuous improvement.
Horse ID Assignment and Position Tracking
The model was trained to assign IDs to each horse in the race course individually and track individual coordinates with latitude and longitudinal values.
On-Site Deployment
The ML model was deployed onsite on the client’s existing platform as a standalone component.
Result
Folio3 successfully deployed an end-to-end ML pipeline that was integrated with the customers’ platform. The model was fine-tuned to deliver the required insights at scale and optimized for the ground truth for outputs. The entire solution enabled them to gather the facts and findings from the races and use that data for in-depth analysis.
