Important Mathematics Institutions | Centre de recherches mathématiques (CRM), Montréal

The simple answer: what mathematical job does CRM perform?

CRM acts as a mathematical network concentrator.

Its permanent laboratories gather much of Québec’s mathematical and statistical research capability into a coordinated system. Its thematic programmes then temporarily concentrate a frontier around a chosen subject, bringing researchers from Montréal into sustained contact with mathematicians from elsewhere in Canada and around the world.

This creates a loop:

local research depth → international thematic programme → new collaboration → postdoctoral and student training → stronger local capability → next international programme.

The institution therefore does not have to choose between being a local research centre and an international visitor institute. Its architecture makes each side more useful because the other side exists.

Why CRM is different from the institutions already mapped

The Important Mathematics Institutions series is now large enough to show that frontier Mathematics has several viable institutional designs.

• The Institute for Advanced Study combines permanent faculty with rotating Members and exceptional freedom for individual research.
• IHES combines a very small permanent faculty with international visitors and a strong Mathematics–physics interface.
• The Max Planck Institute for Mathematics makes its Guest Program one of its defining institutional mechanisms.
• SLMath creates semester-long thematic communities.
• The Isaac Newton Institute builds long programmes across pure, applied and interdisciplinary Mathematics.
• RIMS Kyoto combines permanent faculty, graduate education and international joint-use research.
• The Fields Institute combines thematic research with advanced training and industry-facing routes.
• Oberwolfach specialises in dense one-week workshops and smaller research stays.
• The Clay Mathematics Institute uses fellowships, awards, problems and global partnerships.
• The Institut Henri Poincaré combines thematic research, library/archive functions and a public Mathematics museum.
• BIRS uses focused workshops, small research teams and a distributed international network.
• CIRM Marseille turns residence, conferences, Jean Morlet Chairs and research schools into a high-density mathematical meeting system.

CRM Montréal adds another architecture: thematic visitor programmes sitting on top of a permanent network of research laboratories that already covers a large fraction of the regional mathematical ecosystem.

1968: a research centre built to gather mathematical strength

CRM was founded at the Université de Montréal in 1968.

Its history spans a period in which Canadian Mathematics and statistics expanded rapidly in scale, international connectivity and subject breadth. The centre’s directors over the decades have included researchers connected to graph theory, probability, mathematical biology, optimisation, symplectic geometry, dynamical systems and other areas.

That diversity of leadership reflects the breadth of the institution itself.

The current official history lists Franco Saliola as Director from 2025 onward, following Octav Cornea, Luc Vinet, François Lalonde and many earlier directors.

Official overview: CRM at a Glance and About the Centre de recherches mathématiques.

The dual structure is the key institutional idea

CRM describes its structure as unusual because it combines two systems.

System 1: permanent laboratories

Thirteen laboratories and research groups organise long-term research capability across analysis, applied Mathematics, mathematical biology, number theory, geometry and topology, combinatorics and discrete Mathematics, mathematical physics, probability, statistics, actuarial and financial Mathematics, numerical physics and AI-related research.

System 2: thematic and international programming

Each semester, CRM selects a frontier topic in pure or applied Mathematics and builds workshops, schools, conferences, research visits and postdoctoral opportunities around it.

The two systems interact.

A thematic programme does not arrive in an empty building. It encounters researchers who already know the local strengths. The visitors bring different methods and networks. Local students and postdoctoral fellows gain access to the temporary international community. Collaborations can then persist after the programme ends because the permanent laboratory structure remains.

Temporary frontier + permanent memory is more powerful than either one alone.

Thirteen laboratories make CRM a network rather than a single department

The current CRM structure includes thirteen named laboratories or research groupings.

• Mathematical Analysis;
• CAMBAM, the Centre for Applied Mathematics in Bioscience and Medicine;
• CICMA, the Centre Interuniversitaire de Calcul Mathématique Algébrique;
• CIRGET, the Interuniversity Research Centre in Geometry and Topology;
• GIREF, the Interdisciplinary Research Group in Finite Elements;
• LACIM, the Laboratory of Combinatorics and Mathematical Informatics;
• Applied Mathematics;
• Mila;
• Mathematical Physics;
• PhysNum;
• Probability;
• Quantact;
• Statistics.

This list is institutionally important because the laboratories are not all housed inside one conventional disciplinary vocabulary.

Mila connects Mathematics to machine learning. Quantact connects Mathematics to actuarial science and quantitative finance. CAMBAM connects Mathematics to biology and medicine. GIREF connects analysis and computation to numerical engineering. Mathematical physics creates another route between abstract structure and physical theory.

CRM therefore treats the mathematical sciences as a graph rather than a row of sealed departments.

More than 250 regular members changes the scale

CRM’s current official overview says it has more than 250 regular members, primarily professors at Québec universities.

This makes the centre much larger than an institute defined by a small permanent faculty.

The members remain based in their universities. CRM overlays a collaborative structure across them.

This creates a federation model.

Université de Montréal, McGill, Concordia, UQAM, Université Laval, Université de Sherbrooke and other partners retain their own departments and institutional identities. CRM connects selected research activity across those boundaries.

That allows Montréal and Québec to behave mathematically like a larger integrated research environment without erasing the universities that created the depth.

More than 2,000 participants a year create the international layer

The official CRM overview says its thematic programmes, summer schools and other scientific meetings attract more than 2,000 participants annually.

The number matters because it shows how much mathematical circulation sits on top of the local network.

But volume alone is not quality.

The higher-value question is whether the activities generate:

• new collaborations;
• new research directions;
• better access for early-career researchers;
• transfer between Mathematics and statistics;
• connections to AI, health, finance and physical sciences;
• international circulation of Québec-based researchers; and
• lasting mathematical outputs after the programme ends.

The thematic architecture is designed around these returns.

Thematic programmes are selected frontiers

CRM’s public guidance says that each semester it chooses a cutting-edge topic in pure or applied Mathematics.

The programme can contain workshops, conferences, schools, short-term and long-term visitors, postdoctoral support and Aisenstadt Chair holders.

Proposals are reviewed through CRM scientific committees, including local and international components.

This is a forecasting problem.

A good frontier programme should be mature enough that researchers share a useful language, but open enough that major questions remain unresolved. It should connect fields for a structural reason rather than because interdisciplinarity sounds fashionable. It should provide entry routes for younger researchers rather than becoming a closed seminar for an already-established elite.

Official programme information: CRM Thematic Programs and Proposing an Activity.

2026: Mathematics for Health becomes the current frontier

The current thematic programme is Mathematics for Health, running from 28 July to 13 November 2026.

This is an unusually important modern interface because health problems contain several mathematical layers simultaneously.

• Epidemiology uses dynamical systems, probability and statistical inference.
• Medical imaging uses inverse problems, geometry and numerical analysis.
• Genomics uses statistics, optimisation and high-dimensional data methods.
• Population health uses causal inference and heterogeneous data.
• Drug modelling uses differential equations and pharmacokinetics.
• Public-health forecasting uses uncertainty quantification and computational modelling.
• Health AI uses statistical learning while confronting bias, calibration and distribution shift.

The frontier is therefore not “apply Mathematics to medicine.” The harder task is deciding which mathematical abstraction preserves the biological or clinical structure relevant to the decision.

Public-health modelling shows why prediction is not the same as understanding

During 14–18 September 2026, CRM hosted a workshop on mathematical and data-driven modelling for public-health threats.

That event has just concluded at the time of this article’s 19 September verification.

Public-health modelling illustrates several important mathematical distinctions.

• A model can fit past data and still fail under a new intervention.
• A parameter can be statistically identifiable in one dataset and unidentifiable in another.
• A mechanistic differential-equation model and a machine-learning predictor may answer different questions.
• Confidence intervals do not automatically capture model misspecification.
• Behaviour changes in response to policy, creating feedback between prediction and the world being predicted.

The closed loop becomes:

world → data → model → forecast → intervention → changed behaviour → new data → revised model.

Health Mathematics is difficult precisely because the model participates in the system it is trying to describe.

The Aisenstadt Chair creates a high-impact visiting layer

The Aisenstadt Chair allows CRM to invite a world-leading mathematician or statistician into a thematic programme for a stay ranging from one week to a semester.

The Chair holder gives a lecture series whose first lecture is designed to be accessible to a wider audience, reflecting donor André Aisenstadt’s wish that advanced Mathematics should not remain completely sealed from broader intellectual life.

Chair holders are also invited to produce a monograph in the CRM Monographs series published by the American Mathematical Society.

This creates several returns from one visit:

• frontier research interaction;
• graduate and postdoctoral training;
• a public-facing lecture;
• a durable mathematical text; and
• new institutional links.

Abba Gumel connects epidemic modelling to the thematic programme

From 14–18 September 2026, Abba B. Gumel held Aisenstadt Chair lectures connected to Mathematics for Health.

Gumel is known for mathematical modelling of infectious diseases and public-health systems.

His presence is institutionally appropriate because epidemic modelling sits at the intersection of differential equations, stability theory, probability, parameter estimation and public-health decision-making.

The value of the chair is not simply a famous visitor. It is the temporary embedding of that visitor into a wider thematic system containing workshops, students, local researchers and related health-modelling activity.

Daniela Witten brings modern statistics and high-dimensional data next

The next scheduled Aisenstadt Chair lectures are by Daniela M. Witten on 28–29 September 2026, with events linked to UQAM, McGill and Mila.

Witten’s research is connected to statistical machine learning and high-dimensional inference, including methods used for complex biomedical data.

This placement inside Mathematics for Health is revealing.

Modern health data can contain far more variables than traditional statistical methods were designed to handle. Genomics, imaging and longitudinal digital-health records generate high-dimensional systems in which variable selection, regularisation, multiple testing, latent structure and out-of-sample validation become central.

The frontier is not simply “more data.” It is:

how much reliable structure can be inferred from data whose dimension, dependence and selection process may be as complicated as the phenomenon being studied?

Mila places frontier AI inside the CRM laboratory network

One of CRM’s thirteen laboratories is Mila, the Québec AI institute.

This is a major structural link because it means AI is not external to the CRM ecosystem.

Machine learning depends on Mathematics at several levels:

• linear algebra for representation and computation;
• probability for uncertainty and stochastic optimisation;
• statistics for inference and generalisation;
• optimisation for training;
• geometry for high-dimensional representation spaces;
• information theory for compression and learning limits;
• dynamical systems for iterative training behaviour;
• logic and formal methods for verification.

The institutional advantage is that AI researchers can interact with mathematicians who do not identify primarily as AI specialists but may possess the exact structure needed for a theoretical problem.

Yoshua Bengio is therefore part of a wider mathematical network

The Université de Montréal research directory lists Yoshua Bengio among CRM’s regular members.

This is a useful illustration of how individual, laboratory and institutional nodes should be represented separately in the eventual frontier graph.

Bengio’s individual research contributions belong to a person node. Mila belongs to an institute/laboratory node. CRM belongs to the federation-and-programme node. Université de Montréal belongs to the university node.

The edges between them explain more than collapsing them into one article.

Number theory remains another major CRM strength

CRM’s CICMA network and wider Montréal number-theory community have long been important internationally.

The centre’s September 2026 news provides a particularly strong example: the Simons Foundation announced a new collaboration on Universal Statistics in Number Theory, and CRM highlighted deep links between that collaboration and its own research history.

The CRM report notes that number theorists Kaisa Matomäki and Maksym Radziwiłł developed work during the CRM’s Winter 2014 thematic semester that later became part of research recognised with the 2023 Frank Nelson Cole Prize in Number Theory.

This is exactly the kind of institutional return thematic programmes aim to create.

A programme does not need to produce a theorem before everybody leaves. It needs to create relationships and mathematical pressure capable of continuing after the programme ends.

The centre also notes connections to mathematicians such as James Maynard, Peter Sarnak and Melanie Matchett Wood through fellowships, chairs and programmes.

Universal statistics in number theory shows why randomness appears inside arithmetic

Number theory studies deterministic objects: integers, primes, arithmetic functions and algebraic structures.

Yet many number-theoretic sequences display statistical patterns that resemble random systems.

Prime gaps fluctuate. Values of multiplicative functions display cancellation. Zeros of L-functions show spacing statistics related to random matrices. Arithmetic objects can behave pseudorandomly even though no random coin is being tossed.

This creates a rich interface among:

• analytic number theory;
• probability;
• random matrix theory;
• harmonic analysis;
• combinatorics;
• ergodic methods; and
• computation.

The Bukit Timah Tutor route Automorphic Representations, Harmonic Analysis, L-Functions and the Langlands Program provides an advanced entry into one neighbouring part of this frontier.

Andrew Granville is another bridge between number theory and the CRM ecosystem

The Université de Montréal directory lists Andrew Granville among CRM regular members.

Granville is known for analytic number theory and for work connecting prime numbers, arithmetic functions and probabilistic heuristics.

His presence illustrates the value of CRM’s permanent-member layer. International visitors do not arrive into an anonymous events venue. They enter an existing local research network with deep subject expertise.

Probability is not simply a support tool for statistics

CRM maintains a dedicated probability laboratory and a long tradition of probability research.

Probability today reaches far beyond elementary chance calculations.

• Random matrices model large interacting spectra.
• Stochastic processes model time-evolving uncertainty.
• Random graphs model networks.
• Percolation studies phase transitions in random media.
• Concentration inequalities explain why high-dimensional random variables can behave predictably.
• Stochastic differential equations model systems driven by noise.

These objects now interact with machine learning, statistical physics, finance, geometry and number theory.

The laboratory structure gives probability enough permanence to develop as a field while thematic programmes create temporary intersections with neighbouring areas.

Quantact connects actuarial Mathematics and finance to the research system

Quantact is another CRM laboratory, connecting quantitative actuarial science, risk, insurance and financial Mathematics.

This matters because insurance and finance force Mathematics to confront real decision constraints.

• Tail probabilities matter because catastrophic events dominate risk.
• Dependence matters because losses are rarely independent.
• Stochastic processes matter because prices and claims evolve through time.
• Optimisation matters because capital is limited.
• Statistical estimation matters because model parameters are uncertain.
• Regulation matters because the model is embedded in an institution.

CRM’s October 2026 calendar includes another workshop on fairness and discrimination in insurance, showing how quantitative models now meet ethical and regulatory questions as well as technical ones.

Fairness in insurance is a mathematical problem and a policy problem

An insurance pricing model may be statistically predictive and still raise fairness questions.

Which variables should be allowed? Are some variables proxies for protected characteristics? Does calibration guarantee fairness? Should similar risk imply similar price? How should social pooling interact with individualised prediction?

These questions cannot be solved by one metric.

The Mathematics can clarify trade-offs, impossibility results and sensitivity. Law and policy determine which trade-offs society accepts.

A research centre is useful here because actuaries, statisticians, mathematicians and policy researchers need a shared language before disagreement can become precise.

The postdoctoral layer keeps the network young

CRM’s official overview says it welcomes more than 80 postdoctoral researchers each year.

This is a large early-career population.

Postdoctoral years are mathematically important because a researcher is transitioning from supervised doctoral work toward independent research identity.

A strong postdoctoral environment provides:

• access to several potential mentors rather than one supervisor;
• exposure to seminars across neighbouring fields;
• international visitors;
• time to complete difficult work;
• opportunities to organise and speak;
• career connections across universities and institutes.

CRM and the Institut des sciences mathématiques currently run a joint CRM–ISM postdoctoral competition for researchers seeking to work with Mathematics teams across Québec.

Graduate and postdoctoral training is infrastructure, not an afterthought

A research institute can host brilliant programmes and still weaken over time if no new researchers learn the field.

CRM explicitly integrates students and postdoctoral researchers into its scientific programme.

Schools are one major mechanism. CRM has been a principal partner of the Séminaire de Mathématiques Supérieures since 2011, while the school itself has a history extending over half a century.

In October 2026, the Mathematics for Health programme continues with an autumn school on advanced methods in epidemiological modelling and computational biology.

The sequence is important:

research frontier → school → common language → younger researcher → next research frontier.

The CRM–Fields–PIMS Prize is a network-level signal

CRM jointly administers the CRM–Fields–PIMS Prize with the Fields Institute and the Pacific Institute for the Mathematical Sciences.

The prize recognises exceptional contributions in the mathematical sciences and is described by CRM as a premier Canadian award.

The important institutional point is not the ranking of mathematicians. It is that three different institutes cooperate to recognise research across Canada.

This reinforces the idea that national mathematical capability is a network rather than a winner-takes-all competition among institutes.

The Nirenberg Lectures preserve another mathematical lineage

CRM’s distinguished lecture programme includes the Nirenberg Lectures in Geometric Analysis, with the next series scheduled for 16–20 November 2026.

The series honours Louis Nirenberg, one of the great analysts of the twentieth century, whose work profoundly influenced partial differential equations and geometric analysis.

Geometric analysis is itself a bridge field. It uses analytic techniques—especially PDE—to understand geometric structures.

The Poincaré conjecture provides a famous historical example of this broad strategy: geometric evolution equations transformed a topological classification problem.

Relevant Bukit Timah Tutor routes include Riemannian Geometry and Geodesics and Connections and Curvature.

CRM and CNRS create a formal France–Canada research bridge

CRM operates an International Research Laboratory with the French CNRS.

This formalises another kind of mathematical edge: not one researcher visiting another, but two research systems building an institutional channel for recurring exchange.

Long-term bilateral structures can make collaboration more robust than ad hoc travel because researchers know the route will still exist next year.

The network effect grows when the same centre also maintains exchange agreements with Barcelona, CIRM Marseille and the Institute for Basic Science in South Korea.

CRM and CIRM show how meeting centres and research networks can reinforce one another

CRM maintains a collaboration agreement with CIRM.

The institutional jobs are complementary.

CRM has a permanent Québec-wide laboratory network plus thematic programmes. CIRM specialises in residential meetings, schools, research-in-residence and Jean Morlet Chairs.

A researcher can therefore move through both environments at different stages of a project.

long-term local research → international residential meeting → return with collaborator → thematic programme → paper or new project.

CRM, Fields, PIMS, AARMS and CANSSI form a Canadian institute network

CRM’s official overview lists partnerships with other Canadian mathematical institutes including AARMS, the Fields Institute, PIMS and CANSSI.

This is an important national architecture.

Canada is geographically enormous. No single city can physically contain all its mathematical capability.

A network of institutes allows:

• joint prizes;
• shared summer schools;
• regional training;
• research programmes;
• visitor circulation;
• national advocacy; and
• specialisation by institutional strength.

The national system becomes resilient because capability is distributed rather than concentrated in one organisation.

CRM and the Simons Foundation create another funding edge

The Simons Foundation is one of CRM’s important funders and partners.

This is another example of how philanthropic foundations enter the frontier graph.

Universities supply faculty and students. Governments supply public research funding. Foundations can add targeted support for visits, collaborations, programmes or fields that benefit from long-horizon investment.

The strongest ecosystem does not depend entirely on one funding source because different sources can fail or change priorities at different times.

Public outreach extends the return path beyond professional mathematicians

CRM’s mission also includes public outreach and numeracy programmes.

Its En Avant MATH! initiatives include material designed to help parents and educators support very young children’s mathematical learning.

This may look far removed from L-functions or geometric analysis.

Institutionally, the connection is direct.

A civilisation that funds frontier Mathematics also needs a broad population capable of basic quantitative reasoning and a pipeline of students who can eventually choose deeper mathematical study.

The frontier depends on the base.

Research dissemination creates a durable layer after the programme ends

CRM is associated with several publication series, including CRM Proceedings, CRM Monographs, mathematical-physics series and short courses.

This creates a durable return from temporary mathematical concentration.

A workshop conversation is ephemeral. A monograph, proceedings volume or recorded lecture can preserve a field long enough for researchers elsewhere to enter it.

The best research institutions therefore create multiple memory layers:

• conversation;
• seminar;
• recording;
• preprint;
• journal article;
• monograph;
• teaching material.

Why Montréal itself matters

Montréal contains an unusually dense mathematical and computational ecosystem.

Université de Montréal, McGill, UQAM, Concordia and nearby institutions create strong local depth. Mila adds a globally visible AI research community. Québec’s statistical, actuarial and applied Mathematics groups add industry and policy connections.

CRM gives these separate institutions a shared mathematical layer.

The city therefore behaves less like several isolated university departments and more like a connected mathematical region.

What a Secondary or JC student can learn from CRM

1. Mathematics is a network, not a ladder

School Mathematics often looks sequential: algebra, geometry, calculus, probability. Research Mathematics becomes a web in which several subjects can be needed at once.

2. Statistics is Mathematics at the frontier

High-dimensional inference, health modelling, AI and probability show that statistics is not merely classroom data handling.

3. Applications can generate new theoretical questions

Health, insurance and AI do not merely consume existing Mathematics. They expose new questions about identifiability, robustness, fairness, optimisation and uncertainty.

4. A research career is institutional as well as intellectual

Postdoctoral fellowships, thematic programmes, laboratories and visiting chairs create access to people and problems that no textbook can provide.

5. The frontier needs teaching and public understanding

Schools, public lectures and numeracy outreach are part of the same ecosystem because research capability must be renewed across generations.

From school Mathematics toward CRM frontiers

• Probability → stochastic processes → high-dimensional probability → statistics, finance, AI and health.
• Functions and calculus → analysis and differential equations → epidemiological models, geometry and mathematical physics.
• Number theory → arithmetic functions → L-functions and universal statistics.
• Algebra → groups and representations → combinatorics, geometry and mathematical physics.
• Data handling → statistical inference → high-dimensional modelling and biomedical data science.
• Linear algebra → numerical methods → machine learning, inverse problems and scientific computing.

The advanced frontier is not separate from the foundations. It is what happens when foundational ideas are generalised, connected and subjected to much stronger proof and uncertainty requirements.

CRM institutional map

Entity	Centre de recherches mathématiques (CRM)
Type	Mathematical-sciences research institute and Québec-wide research network
Founded	1968
Host	Université de Montréal
Location	Montréal, Québec, Canada
Current Director checked	Franco Saliola, from 2025
Permanent network	13 research laboratories
Membership scale	More than 250 regular members
Early-career scale	More than 80 postdoctoral researchers welcomed annually
Scientific-activity scale	More than 2,000 participants annually in thematic programmes, schools and scientific meetings
Current thematic programme	Mathematics for Health, 28 July–13 November 2026
Upcoming distinguished activity	Daniela Witten Aisenstadt Chair Lectures, 28–29 September 2026
Major partner institutes	Fields, PIMS, AARMS, CANSSI, CNRS, SLMath, CIRM, CRM Barcelona and others
Verification date	19 September 2026

Connections into the Bukit Timah Tutor Mathematics estate

This page owns the CRM Montréal institutional node. Detailed Mathematics remains with the specialist learning routes.

• Automorphic Representations, Harmonic Analysis, L-Functions and the Langlands Program
• Galois Representations | Frobenius Traces, Determinants and Arithmetic Symmetry
• Representation Theory in Mathematics
• Geometric Representation Theory
• Riemannian Geometry and Geodesics
• Connections and Curvature
• Inverse Problems, Hidden Quantities and Reconstruction
• Conditioning, Ill-Posedness, Sensitivity and Stable Answers
• Prime Sieves, Probable Primes and Primality Testing
• Quantum Entropy, Purity, Mutual Information and Correlations

Return to the Singapore Mathematics Hub for the wider school-to-frontier Mathematics estate.

Official CRM sources and current-status routes

• Centre de recherches mathématiques — Homepage
• About CRM
• CRM at a Glance
• Thematic Programs
• Scientific Meetings
• Distinguished Lecture Series
• Schools
• Proposing an Activity
• Aisenstadt Chairs
• CRM Activities Calendar

The larger lesson

The Centre de recherches mathématiques demonstrates that frontier Mathematics can be strengthened when a region’s permanent research capability and the world’s changing research frontier are placed inside the same institutional loop.

The permanent laboratories give CRM memory. Thematic programmes give it movement. More than 250 members give it regional depth. Thousands of annual participants give it international reach. Postdoctoral fellowships renew the research population. The Aisenstadt Chairs bring extraordinary specialists into temporary residence. Mila, CAMBAM, Quantact and the other laboratories keep Mathematics connected to AI, health, finance, biology and physical science without abandoning pure subjects such as number theory, geometry, analysis and probability.

The result is not one giant department.

It is a mathematical federation capable of changing shape around the frontier while preserving long-term local strength.

CRM matters because it turns Montréal and Québec into a connected mathematical region, then repeatedly opens that region to the world.

That makes the Centre de recherches mathématiques a central Canadian node in any serious global map of frontier Mathematics.