Qindom Extends Boundaries of Quantum Machine Learning to Application

Qindom, the Toronto-based quantum machine learning startup, announced its launch of the cloud-based quantum machine learning platform, Quantum Intelligence Toolbox ("QIT"). The platform connects Qindom's proprietary technologies and algorithms with industrial demands and powers up quantum machine learning further to application levels.

"QIT is the first quantum machine learning cloud platform following the quantum-hybridization principle. It integrates the state-of-the-art quantum intelligence (QI) algorithms with world-leading quantum computing hardware and combines them with classical computing infrastructure. By addressing practical problems in machine learning and AI, this will lead us closer to quantum supremacy in the NISQ era," said Jimmy Tang, product manager from Qindom, who previously served as the product manager for Tmall, Alibaba Group.

Qindom locates in the middleware of AI and optimization ecosphere. It is driven by proprietary quantum machine learning algorithms and methodologies and provides technology services to address complex AI optimization problems. Based on the Quantum Intelligence as a Service (QIaaS) hybridized platform architecture, QIT supports quantum computers, classical computers, and digital quantum simulators on the hardware layer. Current providers include D-Wave, AWS and AliCloud, etc.

With QIT in place, business users will be able to conduct R&D and effectively build applications online through QIaaS. Inside QIT, they can interact directly with API layers for authentication, modeling and prediction, and evoke machine learning algorithms from classical models and Qindom's QI-enhanced ones. "Users and programmers do not need to have a deep understanding of quantum computing and model specifics. We simplify the process by using Python as the only language and SK Learn-style interface to build up a consistent and user-friendly programming environment," explained by Steven Wu, engineering architect at Qindom. "We expect to engage with more quantum hardware companies and cluster providers to enhance the processing power for QIT, and to together share the vision of quantum machine learning."

QIT engineering plan was scheduled in June this year and its prototype was available for in-house testing since the late August. The current version, QIT 1.0, stays full-connect with available machine learning algorithms and methodologies, as well as QI algorithms for combinatorial optimization from Qindom R&D. It supports the optimization of individual machine learning algorithms and the auto-optimization of the designated algorithm set. With Qindom's business network and connections in Canada and China, QIT has been introduced to industry clients who have demands for combinatorial optimization in areas including market price estimation and prediction, personal credit and risk analysis, and customer churn and retention, etc. Future QIT development plans include connecting with upgraded QI-enhanced algorithms and methodologies, supporting more machine learning and AI application scenarios, and filling the gap between user demands and realization by partnerships with independent software vendors (ISVs) and service provider (SPs).

QIT aims to connect current computing technologies and future quantum prospects and to achieve the machine learning speed-up through quantum hybridization. "It is a big step for us. QIT will unlock practical applications of quantum intelligence and advance quantum computing capacities to industry-level advantages," said Jimmy. "We welcome all stakeholders, including hardware providers, business users, ISVs and SPs, to together take a leap in real-world quantum machine learning application."

Around the web

The Department of Defense is gifting Case Western researchers a grant to study the use of AI in determining whether patients require surgery. 

This study brings light to the prospect of a "fully automated solution" for echocardiogram analysis, experts reported.

Experts noted a "significant" reduction in false positives and false negatives using their modified machine learning model.