Alpha Quantum has developed a large number of artificial intelligence technologies and solutions, including:
- Satellite and Aerial Imagery Object Detection and Analysis, using convolutional neural networks
- Semantic Segmentation, using Fully Convolutional Networks
- Specialised Digital Assistants, using Face Recognition, Voice Recognition and TTS technologies
- RNNs (LSTM), Reinforcement learning applied on quantitative investment strategies
- Deep learning nets applied on news analytics
Object detection in Satellite and Aerial Imagery
Our convolutional neural nets enable us detection and object counting in satellite and aerial images. Speed: on a cluster of four GPU (NVIDIA Titan X) machines we can process 100.000 km^2 of images in around 1 hour. Examples of car detections:
We use semantic segmentation based on Fully Convolutional Networks for vehicles detection in satellites and aerial imagery, automated assessment of car damages, detection of car parts for repairs and for analysing growth stage and quality of crops in fields. Example of semantic segmentation of a scenery image:
Use of deep learning nets for quantitative investment strategies
We use artificial Intelligence to run automatised backtesting of strategies and to search for optimal strategies in parameter manifolds. Example of using deep learning nets for dynamic fund of funds allocation. Grey points are a group of backtesting results with the 20% highest Sharpe ratios. Picture is a 2D view of a particular plane from 3D structure with train and test results merged in final view.
Specialised Digital Assistants
Our AI solutions and technologies (face recognition, voice recognition, etc.) enable us to build digital assistants that can be specialised to individual industries and workplaces. We are currently building such specialised digital assistants for the asset management industry and insurance industry.
Alpha Quantum AiVA - Specialised Digital Assistant for Asset Management
Alpha Quantum AiVA - Specialised Digital Assistant for Financial Advisory