---
title: Measurement-Device-Independent Network Steering
url: https://www.emergentmind.com/topics/measurement-device-independent-network-steering
type: topic
---

# Measurement-Device-Independent Network Steering

Searching arXiv for recent papers on measurement-device-independent network steering and closely related network steering frameworks.
Measurement-device-independent network steering is a steering-certification framework for multipartite quantum networks in which all-except-one parties are untrusted and treated device-independently, while trust in the remaining party’s local hardware is removed by replacing calibrated measurements with trusted quantum input states. In this setting, certification is based on observed network correlations and fiduciary quantum states rather than on a model of the measured device. The underlying network-steering formalism generalizes standard EPR steering to independent-source networks through network assemblages and network local hidden state models, and the MDI refinement shows that steering certification can be lifted to a regime in which even the last trusted party treats its apparatus as a black box [2606.25690] [2106.04482].

## 1. Formal setting: network assemblages, trust structure, and NLHS models

The basic object in network steering is a network assemblage. In the multipartite minimal-trust formulation with untrusted parties \(A_1,\dots,A_{n-1}\) and one trusted party \(B\), the assemblage is
\[
\tilde{\sigma}_B^{\vec a|\vec x},
\qquad
\vec x=(x_1,\dots,x_{n-1}),\ \vec a=(a_1,\dots,a_{n-1}),
\]
with
\[
p(\vec a|\vec x)=\operatorname{Tr}[\tilde{\sigma}_B^{\vec a|\vec x}],
\]
and normalized conditional state \(\tilde{\sigma}_B^{\vec a|\vec x}/p(\vec a|\vec x)\). The assemblage obeys the classical-to-quantum no-signalling condition
\[
\sum_{\vec a}\tilde{\sigma}_B^{\vec a|\vec x}=\sigma_B
\qquad \forall \vec x,
\]
with \(\sigma_B\) independent of \(\vec x\) [2606.25690].

Unsteerability is expressed by a network-local-hidden-state model. In the multipartite form,
\[
\tilde{\sigma}_B^{\vec a|\vec x}
=
\sum_{\vec\lambda}
\sigma_B^{(\vec\lambda)}\, p(\vec a|\vec x,\vec\lambda)\,\mu(\vec\lambda),
\]
where \(\vec\lambda\) are hidden variables shared between \(B\) and subsets of the untrusted parties, and the untrusted subnetwork is itself network-local. In the bilocal trusted-endpoints scenario introduced in line networks, the conditioned state
\[
\sigma^{AC}_{b}
=
\sum_{\beta,\gamma} p(\beta)p(\gamma)\, p(b|\beta,\gamma)\, \sigma^{A}_\beta\otimes\sigma^{C}_\gamma
\]
is the corresponding NLHS decomposition [2606.25690] [2106.04482].

A key structural consequence is that every \(\sigma_b^{AC}\) in an NLHS model is separable. Therefore, if even one conditioned state \(\sigma_b^{AC}\) is entangled, the NLHS model is impossible and network steering is certified [2106.04482]. This criterion is specific to networks: the untrusted device is not merely a party in a bipartite steering experiment, but a network node fed by independent sources and possibly performing only a fixed measurement.

## 2. Measurement-device-independent lifting

The MDI refinement replaces trust in the last party’s detector by trusted quantum input states \(\psi^y\). The observed correlations become
\[
p(\vec a,b|\vec x,\psi^y)
=
\operatorname{Tr}\!\left[
M_{BB'}^{b}
\left(\tilde{\sigma}_B^{\vec a|\vec x}\otimes \psi^y_{B'}\right)
\right],
\]
where \(M_{BB'}^b\) is Bob’s unknown joint measurement on his share \(B\) and the input system \(B'\). Equivalently,
\[
p(\vec a,b|\vec x,\psi^y)
=
\operatorname{Tr}\!\left[\mathcal M_{B'}^{b,\vec a|\vec x}\,\psi^y_{B'}\right],
\]
with effective measurement operators
\[
\mathcal M_{B'}^{b,\vec a|\vec x}
=
\operatorname{Tr}_B\!\left[
M_{BB'}^b\left(\tilde{\sigma}_B^{\vec a|\vec x}\otimes \mathbb I_{B'}\right)
\right].
\]
If the original assemblage is NLHS, then
\[
\mathcal M_{B'}^{b,\vec a|\vec x}
=
\sum_{\vec\lambda}
\mathcal M_{B'}^{b|\vec\lambda}\,
p(\vec a|\vec x,\vec\lambda)\,
\mu(\vec\lambda),
\]
so the MDI statistics cannot reveal steering [2606.25690].

The central network result is that every steerable network assemblage can be certified in an MDI experiment by choosing an entangled measurement on Bob’s side. In finite dimensions one may take
\[
M_{BB'}^{b=0}=|\psi^+\rangle\langle\psi^+|,
\qquad
|\psi^+\rangle = \frac{1}{\sqrt d}\sum_{i=1}^d |ii\rangle,
\]
which yields
\[
\mathcal M_{B'}^{b=0,\vec a|\vec x}
=
\frac{1}{d}\left(\tilde{\sigma}_B^{\vec a|\vec x}\right)^{\mathsf T}.
\]
Since steerability is invariant under transpose, this maps a steerable assemblage to a steerable effective assemblage. Consequently, all network-steerable assemblages can be detected measurement-device-independently [2606.25690].

The witness-theoretic implementation is inherited from the bipartite MDI steering formalism. Given a tomographically complete set of trusted quantum inputs \(\{\omega_y\}\), any steering witness can be expanded as
\[
F_{a|x}-\frac{\alpha}{|\mathcal X|}\openone
=
\sum_y \beta_{a,1}^{x,y}\,\omega_y^\mathcal T,
\]
and the MDI payoff is
\[
I(\mathbf P,\beta)=\sum_{a,x,y}\beta_{a,1}^{x,y}\,p(a,1|x,\omega_y).
\]
For unsteerable assemblages, \(I\le 0\); for a steerable assemblage, choosing Bob’s measurement as the maximally entangled projection gives \(I>0\). The same work proves that all steerable assemblages can be detected in this way, and that the optimal Bob measurement is
\[
E_1^{BB_0}=|\Phi^+\rangle\langle\Phi^+|,
\qquad
|\Phi^+\rangle=\frac{1}{\sqrt{d_B}}\sum_{i=1}^{d_B}|i\rangle\otimes|i\rangle
\]
[1807.08901].

This establishes the characteristic MDI asymmetry: trust is shifted from local measurements to state preparation of the quantum inputs. It also clarifies a recurrent misconception. MDI network steering is not fully device-independent Bell certification; rather, it is a minimal-trust certification regime in which a single trusted ingredient remains, namely the fiduciary quantum input ensemble.

## 3. Entangled measurements, swap steering, and one-sided-device-independent detection

A major 2025 development reinterprets network steering as a detector of entanglement in composite measurements. A composite measurement \(\{M_i\}\) is separable if every measurement element is separable,
\[
M_i \propto \sum_j \sigma_{i,j}\otimes \sigma'_{i,j}\otimes \cdots
\qquad \forall i,
\]
with local positive semidefinite operators; it is entangled if at least one POVM element is not separable [2502.06986].

The corresponding witness formalism is directly analogous to state-entanglement witnesses. For every entangled measurement \(\{E_i\}\), there exists an operator \(\mathcal W\) such that
\[
\min_i \operatorname{Tr}(\mathcal W E_i) < 0,
\qquad
\min_i \operatorname{Tr}(\mathcal W M_i) \ge 0
\]
for every separable measurement \(\{M_i\}\). Operationally, the measurement device is untrusted, while the input quantum states are trusted. In the fully trusted tomography implementation one prepares a tomographically complete set of known density operators \(\{\rho_j\}\), records
\[
p_{ij} = \operatorname{Tr}(M_i \rho_j),
\]
and infers entanglement of the measurement from a negative witness value. For the Bell basis, the paper gives the witness
\[
\mathcal W'_{BM}
=
\frac{3}{2}\mathbb{1} - \sum_{i,j=0,1} \mu_{i,j}\otimes \mu_{i,j},
\]
which requires only \(4\) correlations and satisfies
\[
\operatorname{Tr}(\mathcal W'_{BM}\sigma)\ge 0
\quad \text{for all separable }\sigma,
\qquad
\operatorname{Tr}(\mathcal W'_{BM}|\phi_+\rangle\langle\phi_+|)=-\frac12
\]
[2502.06986].

The same paper then lifts the witness to network steering without inputs, also called swap steering. Alice and Bob receive \(N\) independent bipartite sources; Bob holds all \(B_j\) systems and performs a single joint measurement \(\{E_b\}\); Alice is trusted and performs tomographically complete local measurements. The observed correlations are
\[
p(a,b|i_1\ldots i_N)
=
\operatorname{Tr}\!\left[
\left(M_{a|i_1\ldots i_N}\otimes E_b\right)
\bigotimes_{j=1}^{N}\rho_{A_jB_j}
\right].
\]
Classical network behavior is characterized by a separable outcome-independent hidden state model,
\[
p(a,b|i_1\ldots i_N)
=
\sum_{\lambda_1,\ldots,\lambda_N}
\prod_{j=1}^{N} p(\lambda_j)\,
\operatorname{Tr}\!\big[M_{a|i_1\ldots i_N}\, \rho_{\lambda_1}\otimes\cdots\otimes\rho_{\lambda_N}\big]\,
p(b|\lambda_1,\ldots,\lambda_N).
\]

Expanding the measurement witness as
\[
\mathcal W
=
-\sum_{i_1,\ldots,i_N=1}^{d^2}
\beta_{i_1,\ldots,i_N}\,
\tau_{i_1}\otimes\cdots\otimes\tau_{i_N},
\]
one defines the swap-steering functional
\[
\mathcal S_{\{E_b\}}
=
\max_b
\sum_{i_1,\ldots,i_N}
\beta_{i_1,\ldots,i_N}\, p(0,b|i_1\ldots i_N).
\]
The resulting witness bound is
\[
\mathcal S_{\{E_b\}} \le 0
\quad \text{for all SOHS models,}
\]
while for entangled \(\{E_b\}\),
\[
\mathcal S_{\{E_b\}} > 0.
\]
This proves the operational conclusion that every entangled measurement yields swap-steerable correlations, and thus measurement entanglement can be witnessed in a one-sided device-independent way [2502.06986].

The same work also derives a fully device-independent star-network witness. Starting from a Bell inequality
\[
\mathcal B = \sum_{\mathbf a,\mathbf x} c_{\mathbf a,\mathbf x} p(\mathbf a|\mathbf x) \le \beta_{LHV},
\]
it defines
\[
\mathbb E
=
\max_b\left[
\sum_{\mathbf a,\mathbf x} c_{\mathbf a,\mathbf x}\, p(\mathbf a,b|\mathbf x) - \beta_{LHV} p(b)
\right],
\]
with bound
\[
\mathbb E \le 0
\quad \text{for all bilocal (star-local) correlations.}
\]
Using Gisin’s theorem, the paper concludes that any rank-one projective entangled bipartite measurement gives a device-independent advantage in the star network, and notes that higher-rank projective measurements or entangled POVMs also qualify whenever at least one measurement element violates some Bell inequality in the swapping picture [2502.06986].

## 4. Input-free network steering: minimality, topology, and ring-network universality

A distinct but closely related branch studies network steering without measurement inputs. In the minimal two-party, two-source scenario, Alice and Bob each perform one fixed four-outcome measurement, Alice is trusted, and Bob is untrusted. Alice’s measurement is the Bell-basis measurement
\[
M_A=\{\ket{\phi_+},\ket{\phi_-},\ket{\psi_+},\ket{\psi_-}\}_{A_1A_2}.
\]
The paper terms the resulting phenomenon swap-steering and characterizes its classical benchmark by the separable outcome-independent hidden-state model
\[
p(a,b)=\sum_{\lambda_1,\lambda_2} p(\lambda_1,\lambda_2)\, p(a\mid \rho_{\lambda_1,\lambda_2})\, p(b\mid \lambda_1,\lambda_2).
\]
The linear witness
\[
W = p(0,0)+p(1,1)+p(2,2)+p(3,3)
\]
obeys
\[
W \le \beta_{\mathrm{SOHS}}=\frac12,
\]
whereas two Bell pairs and Bob’s Bell measurement reach the algebraic maximum \(W=1\) [2307.08797].

This scenario is minimal in a precise sense. The same work proves that Bell nonlocality cannot be observed at all in the two-party fixed-measurement no-input setting, since any such correlations admit a network-local hidden-variable model. Thus a scenario exists where one can observe quantum steering but not Bell non-locality. At maximal violation, the witness also self-tests two singlets shared across the two sources and Bob’s Bell-basis measurement up to local isometries, and this self-testing result yields
\[
G(\vec p)=\frac14,
\qquad
H_{\min}=2\ \text{bits},
\]
so two bits of randomness are certified from Bob’s four-outcome measurement without seed randomness [2307.08797].

The no-input program was then generalized to triangle and ring networks with one trusted node. In the triangle network, the trusted node \(A_1\) performs the Bell-basis measurement, the untrusted nodes perform arbitrary local four-outcome measurements, and the observed distribution is
\[
p(a_1,a_2,a_3)
=
\operatorname{Tr}\!\left[
\left(N_{A_1}^{a_1}\otimes M_{A_2}^{a_2}\otimes M_{A_3}^{a_3}\right)
\left(\rho_{A_1^2A_2^1}\otimes\rho_{A_2^2A_3^1}\otimes\rho_{A_3^2A_1^1}\right)
\right].
\]
The triangle witness \(W_3\) is a linear functional of \(16\) accepted events and satisfies
\[
W_3 \le \beta_{\mathrm{SOHS}}=\frac12,
\]
while Bell-pair sources and Bell measurements at every node give \(W_3=1\) [2506.23637].

The same paper extends the construction to an arbitrary \(n\)-node ring network. The ring witness \(W_n\) is defined by a recursive coarse-graining condition on the outcome string \(\{a\}_n\), and the bound remains
\[
W_n \le \frac12
\]
for all SOHS models, with \(W_n=1\) in the ideal Bell-pair/Bell-measurement realization. A stronger benchmark, called TSOHS, allows arbitrary additional links among the untrusted parties and constrains only the trusted node to receive a separable two-qubit state from the adjacent sources. The same witness still obeys
\[
W_n \le \frac12
\]
for these more general models, which the paper terms topological robustness [2506.23637].

Noise robustness is also explicit. For triangle sources
\[
\rho(v_i)=v_i|\phi_+\rangle\langle\phi_+|+\frac{1-v_i}{4}\mathbb{1},
\]
and noisy untrusted measurements
\[
\tilde M^{a_i}(w_i)=w_i N^{a_i}+\frac{1-w_i}{4},
\]
the witness becomes
\[
W_3=\frac14(1+3v_1v_2v_3w_2w_3)=\frac14(1+3\tilde v),
\]
so swap-steering is observed iff
\[
\tilde v>\frac13.
\]
Finally, if the trusted node is allowed tomographically complete measurements
\[
\mathcal A_{st}=\{\tau_s\otimes \omega_t,\ \mathbb{1}-\tau_s\otimes\omega_t\},
\]
then for every bipartite entangled state \(\tilde\rho\) with entanglement witness
\[
W_{\tilde\rho}=\sum_{s,t}\beta_{\tilde\rho,s,t}\,\tau_s\otimes\omega_t,
\]
the ring functional
\[
\mathcal S_{\tilde\rho} = -\sum_{s,t}\beta_{\tilde\rho,s,t}\,p(0,0,\ldots,0\mid st)
\]
satisfies \(\mathcal S_{\tilde\rho}\le 0\) for SOHS models and \(\mathcal S_{\tilde\rho}>0\) in the swapped entangled-state realization. This gives a universal input-assisted swap-steering construction for every entangled bipartite state in the ring setting [2506.23637].

## 5. Quantification, optimization, and computational certification

Beyond binary witnessing, the theory of MDI network steering includes quantitative certification. A central route is the assemblage moment matrix hierarchy. In a steering scenario one defines
\[
\rho_{a|x} := \mathrm{tr}_{A}\!\left[(A_{a|x}\otimes \openone)\,\rho_{AB}\right],
\]
and constructs assemblage moment matrices
\[
\chi[\rho_{a|x}] = \chi^{\rm fixed}(P) + \chi^{\rm open}(u)
= \sum_{b,y} P(a,b|x,y)\,F_{abxy} + \sum_v u_v F_v.
\]
These matrices obey positivity,
\[
\chi[\rho_{a|x}] \succeq 0,
\]
consistent reduced state,
\[
\sum_a \chi[\rho_{a|x}] = \sum_a \chi[\rho_{a|x'}],
\]
and normalization,
\[
\sum_a \chi[\rho_{a|x}]_{\rm tr} = 1.
\]
The resulting semidefinite relaxations lower-bound steering robustness directly from observed correlations \(P(a,b|x,y)\), and therefore also lower-bound generalized robustness of entanglement and incompatibility robustness [1603.08532].

The MDI steering program also provides explicit quantitative monotones. One work defines
\[
\mathcal{W}(\mathbf{P},\beta)=\sum_{a,x,y}\beta_{a,1}^{x,y}\,p(a,1|x,\tau_y),
\]
and from this the MDI steerability measure
\[
\mathcal{S}^{\mathrm{MDI}}
=
\max\left\{
\sup_{\beta,\mathbf{P}}
\frac{\mathcal{W}(\mathbf{P},\beta)}{\mathcal{W}_{\mathrm{LHS}}(\beta)}-1,\;0
\right\}.
\]
A practically accessible lower bound is
\[
\mathcal{S}^{\mathrm{LB}}(\mathbf{P})
=
\max\left\{
\sup_{\beta}
\frac{\mathcal{W}(\mathbf{P},\beta)}{\mathcal{W}_{\mathrm{LHS}}(\beta)}-1,\;0
\right\},
\]
and the optimization is an SDP. The same line of work proves that the optimal Bob measurement is the maximally entangled projection and that the MDI measure is equivalent to steering robustness and steering fraction; hence it is a convex steering monotone [1807.08901] [1901.08298].

The quantitative implications extend beyond steerability itself. AMM methods give device-independent lower bounds on steering robustness, generalized entanglement robustness, steerable weight, and measurement incompatibility robustness, while the experimentally accessible MDI measure yields lower bounds on entanglement robustness and incompatibility robustness of the underlying devices [1603.08532] [1901.08298].

For genuine network quantification, a neural-network-based method defines network steerability as the distance to the closest NLHS assemblage. For a target assemblage \(\{\sigma_a^{\bar{\mathcal B}}\}_a\),
\[
\mathcal S^{\mathrm N}
:=
\min
\mathcal D_{\mathrm A}
\left(
\{\sigma_a^{\bar{\mathcal B}}\}_a,\,
\{\sigma_a^{\mathrm{NLHS}}\}_a
\right),
\]
subject to \(\{\sigma_a^{\mathrm{NLHS}}\}_a\in \mathrm{NLHS}\) and fixed reduced trusted state. The ANN is organized to mirror the causal DAG of the network and outputs hidden-variable weights, response functions, and hidden states. In the bilocal steering scenario with a central Bell-state measurement and two isotropic sources
\[
\rho_1 = \nu |\Phi_2^+\rangle\langle\Phi_2^+| + (1-\nu)\frac{\mathcal I}{4},
\qquad
\rho_2 = \omega |\Phi_2^+\rangle\langle\Phi_2^+| + (1-\omega)\frac{\mathcal I}{4},
\]
the paper analytically demonstrates that the network steering threshold is
\[
\nu\omega = \frac13,
\]
with network steering present iff \(\nu\omega>1/3\) under the specified configuration [2502.17084].

## 6. Continuous variables, Gaussian protocols, and experimental infrastructure

Measurement-device-independent network steering extends beyond finite-dimensional systems. In bosonic continuous-variable systems, the finite-dimensional Bell-type projection is replaced by a regularized two-mode squeezed-state projection
\[
| \Phi_{BB'}^\beta\rangle
=
\sqrt{1-e^{-2\beta}}
\sum_{n=0}^{\infty} e^{-\beta n}|nn\rangle,
\]
and the corresponding random displacement channel approaches the identity in the strong operator topology as \(\beta\to 0^+\). Using a gentle measurement / regularization argument, the MDI lifting is proved for all finite-energy assemblages. In the bipartite CV case, the same work also gives a full operational characterization of the steering preorder through MDI steering games [2606.25690].

A notable practical outcome is a Gaussian-only MDI protocol. Fully Gaussian resources and measurements cannot certify Bell nonlocality in a fully DI sense, because Gaussian states and Gaussian measurements do not violate Bell inequalities in the standard fully device-independent setting. Once a single trusted quantum input is inserted, however, Gaussian operations become sufficient for MDI steering detection. In the basic protocol Bob mixes a coherent input state \(|\beta\rangle\) with his system on a \(50{:}50\) beam splitter,
\[
\hat b_1=\frac{\hat x_B+\hat x_\beta}{\sqrt2},
\qquad
\hat b_2=\frac{-\hat p_B+\hat p_\beta}{\sqrt2},
\]
and the witness is
\[
\mathcal W =
\left\langle \left(b_1-\frac{a}{\sqrt2-\beta_x}\right)^2 \right\rangle_{r=x}
+
\left\langle \left(b_2+\frac{a}{\sqrt2-\beta_p}\right)^2 \right\rangle_{r=p}.
\]
For unsteerable states sampled with Gaussian-distributed coherent inputs of width \(\Delta_\beta\),
\[
\mathcal W \ge \frac{\Delta_\beta^2}{1+\Delta_\beta^2},
\]
while any Gaussian steerable state violates the bound after a suitable local squeezing. For a pure two-mode squeezed state \(|t^{(\gamma)}\rangle\),
\[
\mathcal W_{\rm protocol}=\frac12+\frac{1}{2\cosh\gamma},
\]
so any nonzero squeezing can be detected. The construction extends to a line network with witness \(\mathcal W_{\rm line}\) of the same Gaussian form [2606.25690].

The experimental infrastructure for MDI network steering was built largely in bipartite settings. Quantum-refereed steering games demonstrated that steering can be verified without trusting either measurement device by sending nonorthogonal quantum inputs to the untrusted side; an early photonic implementation certified steering for a Werner state with \(W = 0.698 \pm 0.005\), below the CHSH threshold, using a calibrated referee-state parameter \(r^*=1.081\pm0.009\) [1408.0563]. High-dimensional MDI steering was then realized with entangled photonic qutrits, where the quantum-refereed witness verified qutrit steering and certified \(1.106\pm0.023\) bits of private randomness per photon pair, exceeding the one-bit limit for projective measurements on qubit systems [1812.11465].

Loss tolerance is another enabling feature. A measurement-device-independent steering criterion based on the QRS game proves that, for relative frequency of valid measurement outcomes \(\eta_H\), steering can be verified with maximally entangled states whenever the number of settings satisfies
\[
n>\frac{1}{\eta_h},
\]
and Bob’s measurement inefficiency only rescales the score, so certification remains possible for any nonzero \(\eta_m\) [1806.05807]. Quantitative MDI steering has also been extracted experimentally through an entangled-photon implementation of the MDI steerability measure, which is robust against losses and detector biases and simultaneously yields lower bounds on entanglement robustness and incompatibility robustness [1901.08298].

A further trust-reduction step combines MDI steering with self-testing. By self-testing the source of the quantum inputs, one obtains a fully device-independent verification framework for EPR steering. In the three-setting qubit construction, the protocol can verify all bipartite EPR-steerable states in principle, requires about \(99.7\%\) average self-testing fidelity in the analyzed noisy case, and experimentally certified a Werner state with \(v=0.6742(4)\), below the Bell-CHSH threshold \(1/\sqrt2\approx 0.707\) [1909.13432]. A plausible implication is that analogous self-testing layers will be central whenever MDI network steering is pushed toward fully device-independent network certification.

Taken together, these results define a hierarchy. Standard network steering begins with trusted tomography on the steered subsystem; MDI network steering replaces this trust by fiduciary quantum inputs; one-sided-device-independent swap-steering witnesses entanglement in joint measurements; and, in restricted architectures such as star networks or self-tested input schemes, parts of the framework become fully device-independent. The distinctive contribution of the MDI layer is that it preserves the loss tolerance and broad state coverage of steering while adapting the certification task to realistic network settings in which full calibration of every measurement node is unavailable.

Source: https://www.emergentmind.com/topics/measurement-device-independent-network-steering