Tensor Networks & Matrix Product States
When entanglement is low, the exponential wall disappears.
A generic 100-qubit state needs amplitudes, yet DMRG routinely finds ground states of long spin chains to many digits on an ordinary workstation. The reason is entanglement: when every cut through a state is only weakly entangled, the exponential wall collapses into a chain of small matrices.
What you’ll learn
- Define matrix product states and bond dimension, and count their parameters.
- Prove that every state is an MPS with bond dimension equal to the Schmidt rank across each cut, via successive SVDs.
- State and prove the truncation bound .
- Explain area laws and why they make gapped 1D ground states efficiently representable, while volume-law states are not.
Matrix product states
Label the sites from left to right, each with local dimension , and write every amplitude as a product of matrices, one per site:
Here is a matrix for each value of site , and , so the product is a number. The are the bond dimensions and . Storage drops from to numbers.
| State | Bond dimension | Matrices |
|---|---|---|
| Product | , just numbers | |
| GHZ, | bulk , ; ends , | |
| W, | bulk , ; ends , | |
| AKLT spin-1 chain ground state | each spin 1 is a projected pair of virtual spins ½ in singlets with its neighbors | |
| Generic state | up to | no structure to exploit |
Two facts make MPS practical. The norm and the expectation value of any local operator can be computed by contracting the chain from one end in time. And there is a gauge freedom, , that leaves every amplitude unchanged; fixing it, for example to the left-canonical form , turns the bond indices into orthonormal bases of the left blocks.
A generic 20-qubit state is written exactly as an MPS. What is of the bond dimension needed at the middle cut?
Show the answer
10
The middle cut separates 10 qubits from 10 qubits, so the Schmidt rank can be as large as , and for a generic state it is. Exact MPS buy nothing for generic states: the savings come only from low entanglement.
Every state is an MPS
Reshape the amplitudes into a matrix (site 1 against the rest) and take its singular value decomposition:
Set , a row. Regroup the remainder as an matrix and decompose again:
Set and continue to the right end. Each has orthonormal columns, so the result is in left-canonical form:
Bond has the right size: left-canonical form makes the left-block vectors orthonormal. So the matrix decomposed at step is the coefficient matrix of across cut in an orthonormal basis, its singular values are the Schmidt coefficients, and its rank is .
Minimality: an MPS with bond dimension writes across cut (contract everything to the left and to the right of the bond), so the Schmidt rank is at most . A state of Schmidt rank has entanglement entropy at most , which gives the entropy bound.
Truncation: what a small bond costs
Let be the eigenvalues of the reduced state of sites , that is, the squared Schmidt coefficients at cut , and define the discarded weight . For a single cut, the Eckart–Young theorem says truncating the SVD is optimal: the closest vector of Schmidt rank at most is at squared distance exactly . For all cuts at once:
Sweep left to right. At cut , let project sites onto the top- left Schmidt vectors of the current state, and set , with :
Nesting: lies in , so its left Schmidt vectors at cut lie in , and . The projectors therefore commute, , and the errors are mutually orthogonal:
Each term is the weight that cut of discards. Since and is a contraction acting only on sites left of cut , the reduced state of sites can only decrease in operator order; by Weyl's monotonicity theorem each of its eigenvalues decreases, so its tail beyond is at most :
Every bond has dimension at most : cut is truncated at step , and each later truncation can equally be applied as a projection on the sites to the right of its own cut (a truncated Schmidt sum can be produced from either side), which cannot raise the Schmidt rank at cut .
Normalize. is a projection of , so with , and for :
The bound turns entanglement spectra into cost: if the Schmidt spectrum at every cut decays fast, a modest suffices for all cuts at once. How fast is fast enough? Schuch, Wolf, Verstraete and Cirac (2008) showed that even a logarithmically growing von Neumann entropy does not by itself guarantee an efficient MPS approximation, whereas a similar bound on a Rényi entropy with does.
A circuit on a chain starts from a product state and applies CNOTs across the cut between sites and , plus any gates that act entirely on one side of that cut. Ignoring the cap , what is the largest possible Schmidt rank across the cut?
Show the answer
B.
is a sum of two products across the cut (operator Schmidt rank 2), so each application at most doubles the Schmidt rank, and gates on one side do not change it. Hence , attained by building Bell pairs across the cut. A general two-qubit gate has operator Schmidt rank up to 4, giving . This is why shallow 1D circuits are cheap to simulate with MPS.
Area laws: why low entanglement is common
For which states are the small? Ground states of local Hamiltonians typically obey an area law: the entanglement entropy of a region scales with the size of its boundary, not its volume.
| Setting | Block entropy | Consequence |
|---|---|---|
| Gapped 1D local Hamiltonian | , independent of and (Hastings 2007) | MPS with bond dimension suffice; a provably polynomial-time algorithm exists (Landau, Vazirani & Vidick 2015) |
| Critical 1D chain (CFT with central charge ) | for a block in an infinite chain (Calabrese & Cardy 2004) | bond dimension must grow polynomially with ; DMRG still works well |
| Gapped 2D, region | (expected; proved only in special cases) | MPS need ; PEPS are natural, but exact contraction is #P-hard in general |
| Random states, long-time quench dynamics | (volume law; Page 1993) | no efficient MPS: bond dimension exponential in , or in time |
Area laws explain why DMRG (White 1992), which variationally optimizes over MPS, is so accurate in 1D. Circuits fit the same picture: a depth- circuit of nearest-neighbor gates on a chain crosses each cut with at most gates, so it outputs an MPS with , efficiently simulable at shallow depth (Vidal 2003). Deep generic circuits produce volume-law states and defeat MPS.
Key takeaways
- An MPS writes amplitudes as products of matrices: parameters instead of .
- Successive SVDs give an exact MPS whose bond dimensions equal the Schmidt ranks, which is optimal; the entanglement entropy across a bond is at most .
- Truncation error is controlled by discarded Schmidt weight: .
- Area laws ( for gapped 1D ground states) explain the success of MPS and DMRG in 1D; volume-law states from random states, long-time dynamics or deep circuits remain hard.
Check yourself
In the interactive lesson these are scored with your confidence, and what you miss comes back for review.
An MPS on qubits with bond dimension stores how many numbers, up to constant factors?
Show the answer
C.
sites, each with matrices of entries: , linear in .
Across one cut, a state has Schmidt rank . What is the minimal MPS bond dimension on that bond?
Show the answer
B.
An MPS bond of dimension writes the state as a sum of products across the cut, so ; successive SVDs achieve .
For the ground state of a gapped local Hamiltonian on a long chain, the entanglement entropy of a block of consecutive sites (with ):
- is bounded by a constant independent of
- grows like
- grows like
- grows like
Show the answer
A. is bounded by a constant independent of
This is Hastings' 1D area law: the boundary of a 1D block is two points, so the entropy is . Logarithmic growth is the signature of a gapless (critical) chain.
Truncating every bond to discards Schmidt weight at each of the cuts. Which bound holds for the best normalized MPS approximation of bond dimension ?
- No bound is possible without knowing the Hamiltonian
Show the answer
D.
The Verstraete–Cirac bound: errors from different cuts add (nested truncations make them orthogonal), and normalizing costs at most a factor 2.
Further reading
- G. Vidal, "Efficient classical simulation of slightly entangled quantum computations", Phys. Rev. Lett. 91, 147902 (2003).
- F. Verstraete & J. I. Cirac, "Matrix product states represent ground states faithfully", Phys. Rev. B 73, 094423 (2006).
- M. B. Hastings, "An area law for one-dimensional quantum systems", J. Stat. Mech. P08024 (2007).
- U. Schollwöck, "The density-matrix renormalization group in the age of matrix product states", Ann. Phys. 326, 96 (2011).
- J. Eisert, M. Cramer & M. B. Plenio, "Colloquium: Area laws for the entanglement entropy", Rev. Mod. Phys. 82, 277 (2010).