Representation theory is a branch of mathematics that studies abstract algebraic structures—such as groups, rings, Lie algebras, and associative algebras—by mapping them to linear transformations of vector spaces. This mapping, called a representation, allows complex algebraic objects to be analyzed using the tools of linear algebra and matrix theory. In the context of knowledge representation, representation theory provides a rigorous framework for encoding symbolic or structural information (e.g., symmetries, transformations, underlying invariants) into vector spaces, enabling computational manipulation and pattern extraction. The field is foundational in areas including quantum mechanics, crystallography, number theory, and data science.
1.1 Basic definitions and examples
A representation of a group \( G \) on a vector space \( V \) over a field \( F \) is a group homomorphism \( \rho : G \to \mathrm{GL}(V) \), where \( \mathrm{GL}(V) \) denotes the general linear group of invertible linear transformations of \( V \). The dimension of \( V \) is the degree of the representation. For example, the trivial representation maps every element of \( G \) to the identity transformation on a one‑dimensional vector space. For a finite cyclic group \( C_n \), representations can be constructed by sending a generator to a rotation matrix in \( \mathbb{R}^2 \) or to a complex root of unity on \( \mathbb{C} \). Representations of algebras and Lie algebras are defined analogously, with the homomorphism preserving the appropriate algebraic operations.
1.2 Invariant subspaces and irreducibility
A subspace \( W \subset V \) is invariant under a representation \( \rho \) if \( \rho(g)W \subseteq W \) for all \( g \in G \). A representation is irreducible if its only invariant subspaces are \(\{0\}\) and \( V \) itself. Reducible representations can be decomposed into a direct sum of irreducible components (for finite groups over fields of characteristic zero, this decomposition is unique up to isomorphism). Irreducible representations serve as the elementary building blocks for all representations.
1.3 Schur's lemma and its consequences
Schur’s lemma states that any linear map between two irreducible representations that commutes with the group action is either zero or an isomorphism. Over an algebraically closed field, any such map is a scalar multiple of the identity. This lemma has far‑reaching consequences: it implies that irreducible representations of abelian groups are one‑dimensional, and it provides the foundation for the orthogonality relations of characters.
1.4 Characters and character tables
The character \( \chi \) of a representation \( \rho \) is the function \( \chi(g) = \mathrm{tr}(\rho(g)) \). Characters are constant on conjugacy classes and determine the representation up to isomorphism (for finite groups over fields of characteristic zero). A character table lists the values of the irreducible characters on each conjugacy class; it encodes much of the group’s structure and is a central tool in classification problems.
2.1 Group representations
2.1.1 Finite group representations
| For a finite group \( G \), the regular representation (acting on the group algebra \( F[G] \)) contains every irreducible representation as a direct summand. The number of irreducible representations equals the number of conjugacy classes of \( G \). Classical results include Maschke’s theorem (complete reducibility when the characteristic of \( F \) does not divide \( | G | \)) and the orthogonality relations for characters. |
|---|
2.1.1.1 Ordinary and modular representations
| Ordinary representation theory studies representations over fields of characteristic zero (e.g., \( \mathbb{C} \)), where all representations are completely reducible. Modular representation theory considers fields whose characteristic divides the group order (e.g., \( \overline{\mathbb{F}}_p \) with \( p \mid | G | \)). In the modular case, indecomposable but reducible representations appear; the structure is governed by Brauer characters and decomposition matrices. |
|---|
2.1.2 Continuous group representations (Lie groups)
For a Lie group \( G \) (a smooth manifold with continuous group operations), representations are required to be continuous (or smooth) maps \( G \to \mathrm{GL}(V) \) where \( V \) is a finite‑ or infinite‑dimensional vector space. The theory involves the study of unitary representations (e.g., for compact Lie groups all irreducible representations are finite‑dimensional) and the classification via highest weights.
2.2 Module representations (over rings and algebras)
A representation of an associative algebra \( A \) is an \( A \)-module: a vector space \( V \) together with a bilinear map \( A \times V \to V \) satisfying the algebra’s multiplication. This viewpoint unifies group algebras, Lie algebras, and other structures. Modules over a ring can be studied using homological algebra, and the category of modules captures the representation theory of the ring.
2.3 Lie algebra representations
2.3.1 Highest weight theory
For a semisimple Lie algebra (e.g., \( \mathfrak{sl}_n(\mathbb{C}) \)), irreducible finite‑dimensional representations are classified by their highest weight—a dominant integral linear functional on a Cartan subalgebra. The theory uses root systems and the PBW (Poincaré–Birkhoff–Witt) theorem to construct representations via raising and lowering operators.
2.3.2 Verma modules
Verma modules are universal highest weight modules for a semisimple Lie algebra, defined as induced modules from a one‑dimensional representation of a Borel subalgebra. They are infinite‑dimensional but have a unique irreducible quotient, which is the finite‑dimensional representation corresponding to a dominant integral weight. The study of Verma modules is central to the BGG (Bernstein–Gelfand–Gelfand) resolution and understanding of category \( \mathcal{O} \).
3.1 Induced representations and Frobenius reciprocity
Given a representation of a subgroup \( H \subseteq G \), one can construct a representation of \( G \) by inducing it from \( H \). Frobenius reciprocity states a natural bijection between \( \mathrm{Hom}_G(\mathrm{Ind}_H^G W, V) \) and \( \mathrm{Hom}_H(W, \mathrm{Res}_H^G V) \). This technique is fundamental for constructing representations of finite groups and Lie groups.
3.2 Tensor products and Clebsch–Gordan coefficients
The tensor product of two representations yields a new representation. Decomposing a tensor product into irreducible components is a key problem; the coefficients in this decomposition are the Clebsch–Gordan coefficients. For compact Lie groups (e.g., \( \mathrm{SU}(2) \)), these coefficients describe addition of angular momentum in quantum mechanics.
3.3 Harmonic analysis and Fourier transforms on groups
Harmonic analysis generalizes the classical Fourier transform to functions on groups. The Fourier transform on a finite abelian group is given by the characters; for non‑abelian groups it uses unitary irreducible representations. The Peter–Weyl theorem for compact Lie groups states that the matrix coefficients of irreducible representations form an orthonormal basis for square‑integrable functions.
3.4 Categorical viewpoints (Tannaka–Krein duality)
Tannaka–Krein duality recovers a compact group from its category of finite‑dimensional representations, endowed with the tensor product, duals, and associativity constraints. More generally, the representation category of an algebraic structure can be studied via tensor categories, and reconstruction theorems allow one to define a “quantum symmetry” from a given category.
4.1 Physics
4.1.1 Particle physics and gauge theories
Elementary particles are classified by irreducible representations of symmetry groups (e.g., the gauge group \( \mathrm{SU}(3) \times \mathrm{SU}(2) \times \mathrm{U}(1) \)). The Standard Model uses representation theory to describe quarks, leptons, and gauge bosons. The Lie algebra \( \mathfrak{su}(3) \) governs the strong interaction; its eight‑dimensional adjoint representation corresponds to gluons.
4.1.2 Quantum mechanics and spin
In quantum mechanics, the spin of a particle is described by irreducible representations of \( \mathrm{SU}(2) \) (or equivalently of the Lie algebra \( \mathfrak{su}(2) \)). The spin‑\( j \) representation has dimension \( 2j+1 \). The Pauli matrices generate the spin‑\( 1/2 \) representation, and Clebsch–Gordan coefficients govern addition of angular momenta.
4.2 Combinatorics and symmetric functions
4.2.1 Representation theory of the symmetric group
The irreducible representations of the symmetric group \( S_n \) over \( \mathbb{C} \) are indexed by partitions of \( n \). Their characters are given by the Frobenius formula, and the representation theory is intimately related to the combinatorics of standard Young tableaux and the Robinson–Schensted correspondence.
4.2.2 Young tableaux
Young diagrams and tableaux provide a combinatorial tool to construct and label irreducible representations of \( S_n \) and \( \mathrm{GL}_n(\mathbb{C}) \). The decomposition of tensor products (e.g., via the Littlewood–Richardson rule) and the computation of characters rely heavily on these combinatorial objects.
4.3 Data science and machine learning
4.3.1 Group-equivariant neural networks
Group‑equivariant neural networks incorporate symmetry by requiring that learned functions commute with the action of a symmetry group (e.g., rotations, permutations). Steerable CNNs use representations of the rotation group to convolve feature maps; equivariance improves sample efficiency and generalization.
4.3.2 Invariant learning
Invariant learning aims to extract features that are unchanged under the action of a group, often by averaging over the group (group integration) or by employing invariant polynomials. Representation theory provides a systematic way to construct invariant and equivariant mappings in kernel methods and deep learning architectures.
5.1 Representation theory of affine Lie algebras
Affine Lie algebras are infinite‑dimensional Kac–Moody algebras constructed from a finite‑dimensional simple Lie algebra plus a central extension. Their representations are classified by highest weight and integrability, and they play a central role in conformal field theory and the theory of modular forms.
5.2 Quantum groups and deformed representations
Quantum groups are deformations of the universal enveloping algebra of a Lie algebra, often denoted \( U_q(\mathfrak{g}) \). They arise in the theory of exactly solvable models and knot invariants (e.g., the Jones polynomial). Representations of quantum groups are typically finite‑dimensional and exhibit \( q \)-analogues of classical weight theory.
5.3 Geometric representation theory and the orbit method
The orbit method, pioneered by Kirillov, constructs unitary representations of nilpotent and solvable Lie groups from coadjoint orbits in the dual of the Lie algebra. Geometric representation theory uses techniques from algebraic geometry (e.g., sheaves, D‑modules) to study representations; examples include the Beilinson–Bernstein localization relating representations of a Lie algebra to twisted D‑modules on the flag variety.