Innovative Projects Realized

Imitating Artist’s Brushstroke Style by Machine Learning Methods

Acrylic Robotics is a robotics startup based in Montreal, on a mission to use technology to make fine art accessible to the general public. To make fine art accessible, this project focuses on the implementation of a larger robotic arm to recreate artists’ artworks at scale. The intern will be responsible for the development of painting functionality for a new robotic arm, refinement of the robotic painting output, and pilot testing of the robotic system with both internal and external clients. To increase the quality of the paintings, brush strokes should be similar to humans. For this reason, we want to integrate the original artists’ style by collecting brushstrokes (hand brush motion) samples and training a generative model. In other words, after collecting data from artists, a stylized brushstroke generator model will be trained to improve the quality of the paintings. The intern will use their expertise to implement complex control algorithms to recreate the delicate movements of a painter with large robotic arms. The intern will work with both clientele and company personnel to ensure that all desired needs of the client are understood and further translate these needs into control functionalities within the robotic capabilities.

Développement d’un prototype de masque de photolithographie à micro-pixels actionnables (II)

Développement d’un photo-masque pour lithographie digitale dans les procédés de micro-fabrication de semiconducteur.

AI/ML-Driven Optimization of Perioperative Program Workflow: A Data-Driven Approach to Scheduling and Resource Management

We are unlocking existing surgical capacity using technology. Our solutions equip hospitals and clinicians with the insights to optimize surgical workflow and manage future performance with foresight. Behind Sifio’s empowering platform, we’re healthcare partners who combine advanced business science and operational analytics with technology to make a difference. We know that the ultimate goals are to deliver better outcomes, improve the clinical experience, increase sustainability, and enhance patient accessibility. But the chronic obstacles we see in the perioperative workflow (like staff burnout, patient backlogs, and increased costs) are connected to a lack of consistent and reliable information to guide actionable improvements. Our decades of leadership in workflow optimization for healthcare across North America and Europe showed us it is possible to enable collaborative ownership of better healthcare delivery through predictive analytics.

Service Delivery Process Review & Enhancement

Conduct a comprehensive review of Operating Processes related to the provision of the Company’s monthly inventory optimization SaaS program, identifying ways in which timelines can be compressed, process steps eliminated and level of effort reduced.

Identify ways in which alternative technological solutions can be employed to automate repetitive steps and/or error proof recurring activities.

Implement a Business Intelligence approach that provides clients with a “Self Serve” option for Data Visualization that reduces the need for time consuming report generation activities.

Supporting company growth initiatives through data reporting, data quality monitoring, and proactive issue identification. Designing and writing SQL queries to help define core business metrics and evaluate company performance.

Supporting company growth initiatives through data reporting, data quality monitoring, and proactive issue identification. Designing and writing SQL queries to help define core business metrics and evaluate company performance.

Price Optimization and Recommendations Engine

This project aims to help clients achieve their growth goals by leveraging AI and associated techniques to address their business challenges.

These challenges such as increased customer churn across all segments. We aim to develop a tool and an algorithm that can predict churn based on identifying the leading indicators for churn in each segment. By doing so, we will equip our clients with tools for identifying and proactively acting to reduce churn, ultimately helping them retain customers and achieve their growth goals.

To achieve this, we will analyze the data available on the client’s platform and use AI and associated techniques to identify patterns and leading indicators for churn. We will then develop an algorithm that can predict churn in each customer segment and provide our client with tools for identifying and addressing these issues.

This will involve working closely with the client to understand their needs, providing regular updates on our progress, and ensuring that the final product meets their requirements. Through this project, we aim to gain valuable experience in using AI and associated techniques to develop solutions for real-world business problems.

AI/ML-Driven Optimization of Surgical Clinician Schedule

We are unlocking existing surgical capacity using technology. Our solutions equip hospitals and clinicians with the insights to optimize surgical workflow and manage future performance with foresight. Behind Sifio’s empowering platform, we’re healthcare partners who combine advanced business science and operational analytics with technology to make a difference. We know that the ultimate goals are to deliver better outcomes, improve the clinical experience, increase sustainability, and enhance patient accessibility. But the chronic obstacles we see in the perioperative workflow (like staff burnout, patient backlogs, and increased costs) are connected to a lack of consistent and reliable information to guide actionable improvements. Our decades of leadership in workflow optimization for healthcare across North America and Europe showed us it is possible to enable collaborative ownership of better healthcare delivery through predictive analytics.

Automating interaction between Grain Management Software

VeriGrain™ Sampling Inc. (VeriGrain) is an AgTech firm who has created a farm inventory data management system. They solve a substantial problem in the food chain which results in grains being undervalued, under- utilized, and untraceable. VeriGrain requires the ability to connect to third party agricultural software so that their customers do not need to enter the same data into multiple different software packages. Integrating with third party software can be a very time-consuming process. The goal of this project is to develop a semi-automated means of establishing integrations with various agricultural software products and VeriGrain’s grain inventory management system.

Dynamics and control for integration renewable resources

Our research is related to developing mathematical tools for analyzing power systems with massive penetration of renewable power resources, mainly wind, solar and small hydropower: that convex optimization, dynamical systems, model predictive control, optimal control, and passivity-based control. The objective of this small project is to develop model-predictive controls for the integration of renewable energies, especially wind and solar. The student is required to develop a mathematical model, implement that model in Matlab-Simulink, analyze the result and present a final report.

Understanding the evolution of epidemic models during the COVID-19 pandemic

This research project aims to review the different epidemiological modelling approaches used to predict the spread of COVID-19. The project will conduct a comprehensive search of literature and relevant sources, evaluate the quality of the identified studies, extract relevant data, and synthesize the data to provide a narrative summary of the effectiveness of different modelling approaches. The project’s goal is to identify gaps in the literature and suggest areas for future research that can enhance the accuracy and reliability of epidemiological modelling. Ultimately, this research could help in the development and implementation of effective control measures to mitigate the effects of the pandemic.

Belt Conveyor Operating parameters Optimization for Eco-Efficient Ship-Unloading using machine learning techniques

One of the most important pieces of equipments used in maritime transportation for Self-Unloading (SUL) bulk carriers is the “conveyor belt,” which allows the ship to be a very effective and competitive solution. However, the conveyor belts use much energy, which means huge consumption of fuel materials that lead to the emission of polluting gases in harbor territory, calling for immediate actions to sustain the future green seaport vision. The operation parameters determination and their optimization are important to achieve high productivity during the unloading process, energy efficiency, and less pollution. Therefore, this project aims to use data science techniques involving machine learning to understand better conveyor belt operational parameters relationships, manipulating and processing sensors data to find a solution for energy consumption while maintaining the same efficiency and speed. The accuracy of the machine-learning techniques will be evaluated and validated.

Towards certification of a new green fan abradable material

The abradable rub strip (ARS) material’s main function is to minimize the clearance between blade tips and the fan case, thereby increasing the overall engine efficiency. Under tensile and compressive loading, the abradable material behaves different, and has a significant post-failure compaction under compressive loading. Therefore, its characterization is complex and requires a wide range of testing. The intern will continue their work from the previous Mitacs project on analyzing new experimental data, performing statistical analysis, comparing equivalency with incumbent material, and writing knowledge document reports on abradable rub strip (ARS) materials. Furthermore, the characterization test reports will be used to develop material models for ARS that have a significant impact on the engine simulation results of extreme cases like fan blade-off (FBO).