Why AI for the Real World?
AI as we know it has, until now, lived in the digital world. Today, a new era is beginning: one where AI acts directly on the physical world.
This action in the real world changes everything. AI must now contend with gravity, the vibrations of a bridge, the turbulence of an aircraft, or the weight of a decision regarding a person's life. It must therefore prove that it is reliable, that it integrates well with existing systems, that it is safe, that it does not consume too much energy, and that its uses are properly governed.
A collective response to an engineering question
For Polytechnique Montréal, this question is first and foremost an engineering question, and this is precisely where our engineering university plays a distinctive role. It approaches AI as a complete system, one to be designed, integrated, validated, and deployed.
This is how AI for the Real World was born: an initiative that calls on an entire ecosystem, bringing together Polytechnique's researchers, teams and projects, as well as companies, governments, industry partners, students, and engineers who, each in their own way, are bringing AI into the real world.
Bringing artificial intelligence into the real world requires three indissociable revolutions.
AI no longer simply understands or represents the world: it acts on it. It enters robots, drones, autonomous vehicles, smart infrastructure, energy systems, industrial processes, and healthcare environments. This axis focuses on these systems' ability to perceive their environment, make decisions, adapt, and act reliably and safely.
AI no longer simply assists engineers in their tasks: it gives them new ways to design, simulate, explore, and innovate. It is transforming engineering methods and opening up new possibilities, such as exploring a greater number of solutions, accelerating design cycles, or imagining materials, systems, and processes that were previously difficult to conceive.
The more AI acts, designs, and is deployed, the greater its computing needs become. It is becoming necessary to rethink how it computes, as well as the hardware and technological infrastructure it relies on. This axis focuses on developing higher-performing, more efficient computing approaches, and on our ability to master the infrastructure essential to its development.
Tomorrow's AI will rest on an entire ecosystem: algorithms, physical systems, computing infrastructure, data, talent, and governance.
Our commitments ensue from this.
Push the state of the art and develop new systems, methods, and architectures.
Test, evaluate, secure, and strengthen the robustness of AI systems.
Support the integration, scaling, and adoption of AI within organizations.
Develop the necessary skills, culture, and literacy.
