Holography and Quantum Information

Holography relates certain gravitational theories to nongravitational quantum theories, providing a setting in which quantum-information structures acquire geometric meaning. For a quantum-mechanics researcher, encoding, error correction, and recovery offer concrete entry points. Reviewed 11 October 2026, this assessment uses the established bulk-locality/error-correction program and Harlow’s lectures as its selected baseline.

The cited evidence extends through the 2019 universal-recovery treatment. Exact finite-code results below are mathematical benchmarks. Their interpretation as bulk reconstruction is conditional on a gravitational dictionary; the proposed extensions require a fresh topic-specific novelty search, not an inference from the age of these foundational references.

Helpful background. Relationship to Quantum Field Theory distinguishes single-particle descriptions from local-field structure. Knowledge of quantum channels, entanglement, and error correction is useful; a full AdS/CFT calculation additionally requires quantum field theory and general relativity.

Bulk reconstruction as an encoding question

Let an isometry encode a code subspace:

$$V:\mathcal H_{\mathrm{code}}\longrightarrow\mathcal H_{\mathrm{boundary}}.$$

A code operator $O$ is reconstructible on boundary region $A$ when an operator $O_A$ satisfies

$$O_A V=VO$$

on the intended code subspace. This equality can be approximate and state-space dependent. The word “reconstruction” therefore demands a norm, error tolerance, and code-subspace specification.

Almheiri, Dong, and Harlow explain how quantum error correction resolves aspects of bulk locality in AdS/CFT. Harlow’s lectures develop the dictionary and the role of entanglement wedges. A finite quantum code can illuminate the encoding mechanism without becoming a complete gravitational theory.

Exact erasure recovery supplies a falsifiable baseline

For an accessible subsystem $A$ and erased subsystem $E$, exact correction of the full logical system requires that the erased output contain no logical information. In finite dimensions this can be checked on an operator basis:

$$V^\dagger(I_A\otimes X_E)V=c_X I_{\mathrm{code}} \quad\text{for every }X_E.$$

Testing only logical basis populations misses leaked coherence; the condition includes off-diagonal logical matrix elements. Recovering a selected observable algebra is a different, weaker task.

As a small established example, the three-qutrit secret-sharing code of Cleve, Gottesman, and Lo can be written

$$|j_L\rangle=\frac1{\sqrt3}\sum_{k=0}^2 |k,k+j,k+2j\rangle,\qquad j=0,1,2,$$

where additions are modulo three. Every single share is maximally mixed for every logical input. If the third share is erased, apply the basis permutation $|a,b\rangle\mapsto|b-a,2b-a\rangle$ to the first two. It maps the encoded state to

$$|j\rangle\otimes\frac1{\sqrt3}\sum_{t=0}^2|t,t\rangle,$$

where the final pair contains the decoder’s unused qutrit and the erased qutrit. By linearity the first output recovers any logical superposition and its entanglement with a reference. Different authorized pairs overlap, so reconstruction does not create disjoint clones.

This example realizes redundant encoding and complementary ignorance. It has no continuum boundary theory, gravitational dynamics, or geometric area operator. Those missing properties are the reason a correct toy decoder is a starting calculation rather than evidence for AdS/CFT itself.

Open directions suitable for quantum-information methods

Approximate recovery. Physical code subspaces and finite-size models are rarely exact. How does reconstruction error depend on region size, code dimension, and perturbations? Cotler et al. use universal recovery to handle approximate bulk/boundary relative-entropy relations. The active task is to obtain controlled errors under a specified access and code-space restriction, rather than replacing an approximate equality with an exact theorem.

Algebraic encoding. Reconstructible observables can form an algebra rather than the full matrix algebra of a tensor factor. Centers and constraints matter; assuming every region corresponds to an independent factor can erase the relevant structure.

Complexity. Existence of a boundary representative does not imply efficient construction or laboratory implementation. Recovery cost and access restrictions are scientific questions.

Model discrimination. Different toy networks can share an error-correction property while disagreeing on spectra, dynamics, or locality. A research claim should say which gravitational feature the model reproduces and what it leaves absent.

These directions connect to entanglement and spacetime and black-hole information, but their deliverables differ: geometric reconstruction, evaporation entropy, and code recovery should not become interchangeable evidence.

An entry project: recovery under a controlled perturbation

Baseline. Implement the three-qutrit code and decoder above, then check recovery on all nine logical matrix units and on a logical qutrit maximally entangled with an untouched reference. Verify the erased output is $I/3$ for arbitrary logical inputs. This fits in exact matrices of dimension 27 before adding the reference.

Resources. A laptop, finite-dimensional linear algebra, quantum channels, and basic convex optimization. Use a small code where exact matrices remain available.

Deliverable. Perturb the encoding as $V_\epsilon=e^{-i\epsilon K}V$, with a fixed Hermitian $K$ of operator norm one acting across accessible and erased shares; this preserves isometry exactly. Compare the fixed decoder with one reoptimized for entanglement fidelity. Plot error against $\epsilon$ and erasure choice, including a perturbation supported only on the accessible side as a control. A maximum over sampled input states is a lower bound on worst-case error, not a uniform certificate. Use a channel-norm optimization or a proved bound if making a worst-case statement.

Failure criterion. A high fidelity for a single logical state does not certify a quantum code. A decoder that accesses erased qubits violates the erasure model. Check that every claimed recovery map is completely positive and trace preserving.

Research extension. Restrict decoder locality or circuit depth and derive an error-cost tradeoff. Alternatively, study an operator algebra with a nontrivial center and compare algebraic recovery with tensor-factor recovery.

Choosing the bridge responsibly

This direction is attractive for students with information-theory skills, but novelty requires more than attaching geometric vocabulary to a standard code. A strong proposal identifies an existing holographic question and explains which code property will answer it.

Progress would be a sharper approximate-recovery bound, an explicit resource limitation, or a controlled model that captures an additional independent gravitational feature. No finite qubit experiment alone establishes that the universe has a holographic dual.

References

• A. Almheiri, X. Dong, and D. Harlow, “Bulk Locality and Quantum Error Correction in AdS/CFT,” Journal of High Energy Physics 2015(4), 163 (2015). arXiv:1411.7041.
• R. Cleve, D. Gottesman, and H.-K. Lo, “How to share a quantum secret,” Physical Review Letters 83, 648–651 (1999). DOI — exact secret-sharing and erasure baseline.
• J. Cotler, P. Hayden, G. Penington, G. Salton, B. Swingle, and M. Walter, “Entanglement Wedge Reconstruction via Universal Recovery Channels,” Physical Review X 9, 031011 (2019). DOI — approximate recovery under holographic assumptions.
• D. Harlow, “TASI Lectures on the Emergence of the Bulk in AdS/CFT,” Proceedings of Science, TASI2017, 002 (2018). arXiv:1802.01040.

Entry points

Read the code-subspace construction in the primary paper, then the corresponding lecture discussion. Reproduce one exact recovery before interpreting an approximate numerical model.