Skip to content
AI with Roche

Collaborative for Healthcare Transformation

AIR is a centre of excellence with a core purpose to deliver to people in Canada and beyond better health outcomes through the discovery and application of artificial intelligence research underpinned by an open and collaborative exchange uniquely supported by Roche expertise.

As we challenge ourselves to explore new methods and solutions to improve health outcomes for our communities, we recognize the great potential AI may bring to every part of the patient journey. As a tool and discipline, it is revolutionizing how we diagnose and treat diseases and discover new drugs while making health care more efficient, effective, and affordable in ways that are both responsible and transformative.

AIR is accelerating early data insights generation to bring tangible benefits to patients now and for the future.

A researcher looks at cells on a computer screen.

Leading Edge AI Research

The genesis of AIR lies in the belief that by bringing the most innovative research in AI to the forefront, we can make significant improvements in the health outcomes of our communities.

A father and his daughter talk to a doctor over video call.

Facilitating Partnerships

AIR exists to foster a highly collaborative environment and methodology. To that end, we facilitate meaningful public and private partnerships with an eye on impactful innovation.

A doctor wearing a mask and scrubs looks at vitals on a computer monitor.

Mobilizing AI into Practice

While scholarship may ground our work, practical application gives it its reason for being. We look to deliver real world solutions for real world challenges.

A doctor smiles at a patient who is sitting in his examination room.

Safety and Ethics

We act responsibly in everything we do with a particular focus on safety and privacy. We believe success is reliant on us being transparent, ethical and accountable to all our stakeholders.

Part of a Vibrant AI Community

The AI community is a multi-disciplinary one which we are proud to be a part of. The diversity of people; of thought; of approach helps our work and we believe in tapping into this robust and vibrant community, we can best deliver on our core purpose to improve health outcomes. A diagram showing all the different organizations in the AIR community: AIR, Roche, Amii, Mila, Vector Institute, Hospital/Healthcare Delivery Organizations, Patient Organizations, Academic Research Organizations, Government, Start-ups, Multi-nationals, and other organizations outside of healthcare

Community-Driven Science

AIR is a part of a Learning Health System comprised of patients, healthcare providers, and the AI community. The ecosystem relies on all parties working together to generate meaningful and higher-quality data, deploy tools to assist in transforming those insights into action, and build safe spaces for experimentation with new innovations.

  • Impactful solutions start with the patient.

    Our work aims to be truly patient-centric and approached from the patient’s perspective, ensuring we are listening to the patient’s needs and drawing meaningful conclusions. We seek to bring our expertise to partnerships and solutions that put the person first, and that facilitate the best experience and outcome for that person throughout their healthcare journey.

  • What is Data Governance?

    Data Governance is a best practices framework for managing data assets. It defines roles, responsibilities, and processes for ensuring accountability for and ownership of data assets across an organization or community of stakeholders, and ensures the interests of the community members are protected when it comes to managing their health data.

  • What does infrastructure refer to in this context?

    A system that is able to store, capture, and share health data in a safe, secure and effective and compliant manner across the community. Infrastructure connects existing systems and utilizes methods and solutions to combine data points and harness bigger data.

  • What is Solutions Development?

    At AIR, we are developing health solutions for citizens — designed and informed through the effective use of health data and insights — and encompassing perspectives from all community members.

  • What is Data Synthesis?

    Drawing from real-world experiences and generating data to continually provide real-time feedback to the system, healthcare providers, and patients. This feedback can be used to enable quality improvement in health solutions. Effective data synthesis involves transparent research plans ahead of data collection.

Xcelerate RARE: A Rare Disease Open Science Data Challenge

Rare Diseases face significant challenges in diagnosis and treatment and are associated with an unprecedented burden on individuals living with the disease and their caregivers ~ only 5% of the known 10,000+ rare diseases have an FDA-approved treatment.

Roche has formed a collaborative partnership with RARE-X Global Genes with an intent to resolve barriers for patients to responsibly share their data and engage in research.

AI with Roche (aiR) ( is happy to announce the launch of #RareDisease Open Science Data Challenge which is bringing together researchers, clinicians and data scientists alongside patients in an open environment to accelerate scientific discovery.

Learn more here: Join the challenge at and let’s make a difference together!

Dynamic Teams

At AIR, we create forums for public, private and academic discussion and use open science to engage people and organizations that would not normally come together.


Roche Data Science Coalition

RDSC is a group of like-minded public and private organizations committed to working with the global community to develop solutions to the challenges of the COVID-19 pandemic. The collaborators in the Coalition include: Alberta Machine Intelligence Institute (Amii),, NVIDIA, Self Care Catalysts, ThinkData Works Inc and Vector Institute.

Find out more

COVID Long-hauler Project

The industry coalition and Vector Institute will try to better understand the COVID-19 “Long Haulers” phenomenon by leveraging NLP methodologies and subject matter expertise on unstructured online data (social media, news articles, etc.). The intent is to develop a capability to recognize early signals of symptoms related to a condition that previously has been poorly defined and for which there is not yet consensus in the medical community, in order to help focus research questions and develop treatment strategies.

Find out more


The End ALS Challenge, administered by Kaggle, an online community of data scientists and machine learners, presents 150 rich datasets contributed from 1,000 Answer ALS patient research participants. The goal is to surface insights through an open data competition that connects the global AI and neuroscience communities to better understand the overall biology of amyotrophic lateral sclerosis (ALS), and improve diagnosis and drug discovery for ALS patients.

Find out more

AI-based Precision Oncology Platform for Development of RNA-based Predictive Biomarkers of Drug Response

This project aims to develop a new AI-based platform based on RNA-seq data to assist in matching cancer patients to therapies. The main innovation of the system is its ability to leverage both large-scale preclinical pharmacogenomic data and clinical trial datasets to enrich the portfolio of candidate drug response biomarkers.

The main focus is on developing univariate and multivariate predictors of drug response based on pharmacogenomics profiles of preclinical models using RNA-seq as input, and predict therapy response as output. The AIR team in particular will provide internal clinical trial datasets to support a clinical validation of developed models.

Find out more

Synthetic Data Bootcamp

Data researchers from various industries have joined forces to participate in the Vector Institute’s Synthetic Data Bootcamp – an intensive three-day study of current trends in synthetic data topics. Data scientists and computational biologists expanded their knowledge base and applied it to the exploration and benchmarking of generative models, such as DPC-GAN, PATE-GAN and fastGAN. The AIR team has put in practice the new skills to work with tabular datasets, specifically rare disease patient data made available thanks to the EndALS initiative. The major outcome of the Bootcamp is an improved understanding of the advantages and drawbacks of synthetic data approaches, which will complement the ongoing exploratory work on synthetic patient data.

A discussion on Artificial Intelligence in Health AI with Roche - Presented by EFFERVESCENCE

Canada adopted a national artificial intelligence (AI) strategy in 2017 and is now considered one of the most advanced countries in this field, as it continues to invest in and attract researchers, students, and experts to build a stronger community.

Considering the importance of AI innovation in healthcare in Canada and the impact that AI discovery and application can have for patients and the healthcare system, Roche Canada created a center of excellence in artificial intelligence called AI with Roche (AIR), in collaboration with the three national AI institutes under the CIFAR Pan-Canadian AI Strategy - Amii, Mila and the Vector Institute.

Joining forces with experts of the research community aims to improve those outcomes from AI applied in healthcare.

Join us for a virtual discussion about the potential of artificial intelligence in health, public-private partnerships and collaboration, and possibilities to improve health outcomes for populations and patients across Canada by leveraging AI.

Meet the AIR collaborators and hear institutional and political perspectives about the future of artificial intelligence in the healthcare system.

Let’s Collaborate

Let us know about your goals and challenges for AI adoption and we'll be in touch shortly!

[email protected]

What to Include

  • Some background on your organization
  • Would you like to participate in an existing initiative or do you have an idea for a new initiative?
  • Are you interested in implementing AI into medical practice, educating your organization on AI in healthcare, or contributing to the development of AI studies?
  • Anything else you think we should know

What Not to Include

  • Confidential or proprietary information