Quantum Machine Learning Algorithm Research, Development and Application in Logistics Planning - ON-214

Preferred Disciplines: Machine Learning, Computer Science, Computer Engineering, Operations Research, Data Science (Masters, PhD and Post-Doc)
Company: Qindom
Project Length: 1 year
Desired start date: As soon as possible
Location: Toronto, ON
No. of Positions: 2-3 (1-2 PhD/Post-Doc and 1-2 master-level intern)
Preferences: We would prefer candidates based in Ontario, especially GTA. We also prefer students from University of Toronto and University of Waterloo.

About the Company: 

Qindom is a premier Quantum Intelligence (QI) research and application service provider. Born as the game-changer in the present AI world, we focus on developing Quantum Machine Learning (QML) algorithms and addressing complex AI optimization problems. Qindom has gathered the most brilliant minds of our times from different academics and industries. We do not just believe that the QI era is here, we practice realizing that. With our proprietary QML algorithms, we apply quantum-inspired and quantum-classic hybridization principles to fill the gap between users’ AI demands and realization via Quantum Intelligence as a Service provided by our Quantum Intelligence Toolbox.

Project Description:

Project 1

The project will be focused on developing a new quantum machine learning (QML) algorithm. We expect to first design and implement the architecture that links quantum optimization with classical machine learning methods (including deep learning and neural network), then build our novel models on top of the new architecture.

Project 2

The project is set to address sequencial location-based combinatorial optimization in an accurate and effective manner. Quantum machine learning and quantum optimiztion alogorithms will be applied through an API built for Qindom’s ISV partners in the logistics planning market. 

Research Objectives:

Project 1

Objective: The new algorithm is projected to boost the training and learning capabilities of classical machine learning (including deep learning and neural network) algorithms both in terms of speed and accuracy.

Sub-objectives:

  • Formalize the theories and proofs for the new quantum machine learning algorithm and implement it in our hybrid architecture;
  • Performance evaluation against current state-of-the-art algorithms;
  • The algorithm will be applied for perdictive analysis for multiple usercase scenarios, viz. logistics planning, market price estimation (real estate and FinTech), and recommendation system;
  • Results presentation on premier academic conferences, publications on top-notch peer-reviewed journals, and/or media outlet coverages.

Project 2

Objective: Apply Qindom’s proprietary technology and QML algorithms in logistics planning to generate real-time optimized solution sets for route forecasting and planning in complex and dynamic traffic conditions.

Sub-objectives:

  • Follow the quantum-inspired principle to apply QML algotithm;
  • Apply (simulated) annealing and parallel computing to address high-dimensional MSR problem;
  • Realize real-time forecasting by reducing computational time and removing pre-searching steps through online learning;
  • Beat industry best results with a relatively significant level in benchmarking;
  • Revenue generation via commercialization with ISV partnerships.

Methodology:

Project 1

  • Investigate existing research in this field and extract valuable information to enhance our algorithms;
  • Enhance ensemble methods with the theories of classical combinatorial optimization, machine learning, and quantum computing;
  • Formalize the theories and proofs following the quantum-inspired principle
  • Follow the quantum-classic hybridization principle and implement on the hybrid architecture;
  • Benchmark testing against current state-of-the-art algorithms on standard benchmark instances;
  • Integrate with Quantum Intelligence Toolbox (QIT) QIaaS (Quantum Intelligence as a Service) platform to connect to industry applications.

Project 2

  • Design and develop API specialized for logistics planning built on the QIT QIaaS platform;
  • Implement big data and cloud service development strategies;
  • Conduct feature engineering with logistics industry data and reconstruct training data;
  • Present optimized results via multiple platforms to our industry ISV partners, including S3 for massive data, restful API for small-scale data, and open-refine GUI;
  • Productionization and commercialization of the QML logistics planning system.

Expertise and Skills Needed:

Project 1 (PhD/Post-doc Level)

  • Expertise in statistical modelling, optimization, AI, and applications of AI with demonstrated success in ML application to practical domains;
  • Deep understanding of ML theory and recent advances (ex. Neural Network, Deep Learning);
  • Pragmatic orientation to build ML solutions that work in the field;
  • Experience in design, training, validation and evaluation of ML models;
  • Understand at least one of the ML frameworks – TensorFlow, Scikit, MLPack, SparkMLib, Caffe, etc.
  • Self-motivated, proactive and flexible with excellent problem-solving skills;
  • Preferable have published in top-notch peer-reviewed journals.

Project 2 (Mater Levels)

  • Hands-on experience using Java and Python;
  • Familiarity with knowledge of data structure, distributed system;
  • Experience with creating and consuming REST Services;
  • Proficient working with relational databases as well as NoSQL technologies;
  • Strong understanding of statistics;
  • Knowledge in Machine Learning algorithm, implementation, evaluation, maintenance and application;
  • Madarin speakers are preferred.

For more info or to apply to this applied research position, please

  1. Check your eligibility and find more information about open projects
  2. Interested students need to get the approval from their supervisor and send their CV along with a link to their supervisor’s university webpage by applying through the webform.
Program: