Can Quantum discord increase in a quantum communication task?

Shubhayan Sarkar1 and Chandan Datta2, 3 1School of Physical Sciences, National Institute of Science Education and Research,HBNI, Jatni - 752050, India 2Institute of Physics, Sachivalaya Marg, Bhubaneswar 751005, Odisha, India. 3Homi Bhabha National Institute, Training School Complex, Anushakti Nagar, Mumbai 400085, India. (Dated: August 13, 2018) Quantum teleportation of an unknown quantum state is one of the few communication tasks which has no classical counterpart. Usually the aim of teleportation is to send an unknown quantum state to a receiver. But is it possible in some way that the receiver’s state has more quantum discord than the sender’s state? We look at a scenario where Alice and Bob share a pure quantum state and Alice has an unknown quantum state. She performs joint measurement on her qubits and channel to prepare Bob’s qubits in a mixed state which has higher quantum discord than hers. We also observe an interesting feature in this scenario, when the quantum discord of Alice’s qubits increases, then the quantum discord of Bob’s prepared qubits decreases. Furthermore, we show that the fidelity of one-qubit quantum teleportation using Bob’s prepared qubits as the channel is higher than using Alice’s qubits.

I. INTRODUCTION has been used in the language of quantum information where the measurements are not noisy. The notion of Non-classical correlations has been identified as a re- Quantum discord was first introduced by Zurek [10] and source for different communication tasks from quantum independently by Vedral [11] and Horodecki [6]. Oliver teleportation [1] to remote state preparation [2]. Most and Zurek then went on to give a measure for Quantum of the communication tasks uses entanglement as a re- discord [10]. Quantum discord was first identified as a source like teleportation, but only a few tasks use quan- resource for computation by Datta et al.[12], where they tum discord as a resource. Essentially the use of quan- took a separable state as the resource and showed that tum discord as a resource is not very well explored or the computation was better than using a classical state is controversial. For instance, let’s take the example of but not better than using an entangled state. Later Dakic remote state preparation. It was shown that separable et al. showed the role of quantum discord in remote state states can outperform entangled states for remote state preparation [2], provided that Alice and Bob don’t share preparation. However it was later shown that, its valid- a reference frame or the operations are bio-stochastic [3]. ity is restricted to some special conditions [3]. Quantum Many other applications of quantum discord have been discord has also been shown to be generated in the state proposed whose descriptions can be found in [13, 14]. conversion process even when there is no entanglement Olliver and Zurek [10] proposed a measure of quantum [4,5]. discord in terms of the mutual information. In classical A two qubit mixed state, is called classically correlated information theory, one can express the mutual infor- [6] if it can be written as mation between two random variables X and Y in two different ways - X A B ρcc = pij|iihi| ⊗ |jihj| , (1) I(X : Y ) = H(X) + H(Y ) − H(X,Y ) and i,j J(X : Y ) = H(X) − H(X|Y ), (2) A where |ii are the orthogonal states over the hilbert space where H(X) = − P p log p is the Shannon entropy B x x 2 x HA and |ji are the orthogonal states over the hilbert of X, p is the probability that X takes the value x A B x space HB. However if |ii or |ji are not orthogonal and H(X,Y ) is the joint Shannon entropy of X and arXiv:1808.04363v1 [quant-ph] 13 Aug 2018 then the state AB might be non-classically correlated. Y . H(X|Y ) is the conditional entropy and defined as This led to the idea that mixed states might posses some P H(X|Y ) = y pyH(X|y), where py is the probability of non-classical correlations which is different from entan- Y taking value y and H(X|y) is the conditional entropy glement as the state could be separable. The first non- of X, such that Y take the value y. Unlike classical do- classical correlations which differ from entanglement is main these two expression are different in quantum do- Quantum discord. However, it should be noted that the main and their difference serve as quantum discord. The term “non-classicality” may have different meanings in generalization of I(X : Y ) in quantum theory is quantum information and quantum optics, as was shown in Ref. [7,8] that in the context of quantum optics non- I(ρAB) = S(ρA) + S(ρB) − S(ρAB), (3) zero quantum discord might have some classical explana- tions and as well as in Ref. [9] it was shown that classical where I(ρAB) is the mutual information between A and state may posses some nonzero discord if the measure- B for the state ρAB, S is the von Neumann entropy and ments are noisy and can be represented by a stochastic ρA, ρB are the reduced density matrix. However, the channel. In our manuscript the term “non-classicality” generalization of J(X : Y ) is not that straight forward. 2

Olliver and Zurek [10] extended this to quantum realm 1.0 as λ≈0.82047 AB A B 0.8 J(ρ ) B = S(ρ ) − S(A|Π ) (4) Πi i initial state discord 0.6 and S(A|ΠB) is given by i δ final state average discord B X 0.4 S(A|Πi ) = piS(ρi), (5) i 0.2 B AB B AB B where ρi = T rB[Πi ρ ]/pi, pi = T r[Πi ρ ] and Πi are the measurement operators on the subsystem B. 0.0 Therefore, the measure of Quantum discord as proposed 0.0 0.2 0.4 0.6 0.8 1.0 in [10] λ

B|A AB AB δ = min B [I(ρ ) − J(ρ ) B ], (6) Πi Πi FIG. 1: The red line shows the quantum discord as function of parameter λ for the initial Alice’s Werner where the quantity I represents mutual information and state. The green line shows the average discord of the the quantity J represents the amount of information final states with variation in λ (using the cluster state gained about the subsystem A by measuring the subsys- (7) as the channel). tem B. The minimization occurs over the set of mea- surement operators such that the quantum discord is measurement independent. For two-qubit states a par- tial analytic approach was given by Girolami and Adesso and I is a four-dimensional identity matrix). Then, Al- [15], however this also needed numerical minimization ice jointly measures her two-qubit state and the chan- scheme. nel in such a way that Bob receives a two-qubit state. In Quantum teleportation [16, 17] two spatially sep- For this joint measurement Alice can choose a basis ar- arated parties Alice and Bob share a quantum channel. bitrarily. As a result she would find different outcomes Alice has an unknown qubit which she wants to prepare corresponding to the basis elements. Depending on Al- at Bob’s end without physically sending it. Then opti- ice’s outcomes, the two-qubits with Bob will collapse to mizations are done over the measurement basis and chan- different states. Alice classically communicates her out- nel such that Alice’s and Bob’s state have greatest degree come to Bob. Now Bob measures the quantum discord of overlap (essentially this is maximizing the fidelity). for each of the separate outcomes and then averages it We look at an almost similar scenario where Alice and over all the outcomes. It should be noted here that Bob Bob are spatially separated and have a shared quantum doesn’t need to apply any specific unitary measurement channel. Alice wants to prepare a two qubit state, at (U1 ⊗ U2), as discord remains conserved under a unitary Bob’s end such that the quantum discord of Bob’s state transformation. Also this protocol is different from the is higher than Alice’s state. Moreover, we have shown remote state preparation in the sense that Alice doesn’t that when the average quantum discord of Bob’s state is know about the state she wants to send to Bob. higher than Alice’s initial state and if we use Bob’s state We compare the average quantum discord of the states as a resource to teleport a single qubit, instead of Alice’s with Alice and Bob and find that average quantum dis- state then the teleportation fidelity increases. As, Quan- cord of Bob’s state is higher than Alice’s initial Werner tum discord is very difficult to calculate analytically we state for most of the range of λ. We keep the basis chosen used Mathematica (Qdensity [18]) for numerically finding by Alice fixed, and vary the parameter λ of the Werner the results. state. The paper is organized as follows. SectionII contains the description of the protocol and contains a detailed worked out example with the protocol. A possible appli- A. Example 1 cation of our result is discussed in sectionIII and finally we conclude in sectionIV. Let’s take a four-qubit Cluster state   |ΨiC = 1/2 |00iA|00iB+|01iA|10iB+|10iA|01iB−|11iA|11iB , II. DESCRIPTION OF THE PROTOCOL (7) where first two qubits are with Alice and rest two qubits Alice and Bob share a quantum channel which is a are with Bob. Alice has a two-qubit Werner state ρA = pure state of dimension more than 4 (for eg. three-qubit λ|φ+ihφ+| + (1 − λ)I/4. So the total state of the 26 W state, or four-qubit cluster state, etc.). Alice has an dimensional system is unknown quantum state (a two-qubit Werner state ρA = λ|φ+ihφ+| + (1 − λ)I/4, where |φ+i = √1 (|00i + |11i) 2 ρT = ρA ⊗ |ΨiC hΨ|. (8) 3

Alice performs joint measurements on the qubits 1, 2, 3, 4 1.0 λ≈ in the basis which is chosen arbitrarily (we want to 0.8876 project the state of those 4-qubits onto an entangled ba- 0.8 sis similar to teleportation) and sends 4 bits of classical initial state discord information to Bob. We are interested in the properties 0.6 final state average discord

(specifically discord) of the state of the qubits 5, 6 (which δ is with Bob) after Alice performs her measurement. Al- 0.4 ice chooses a complete orthonormal measurement basis as 0.2 1  |b1i = |0001i + |0010i + |0100i + |1000i , 2 0.0 1  0.0 0.2 0.4 0.6 0.8 1.0 |b2i = − |0001i + |0010i + |0100i − |1000i , 2 λ

1  FIG. 2: The red line shows the quantum discord as |b i = − |0001i + |0010i − |0100i + |1000i , 3 2 function of parameter λ for the initial Alice’s Werner 1  state. The green line shows the average discord of the |b i = − |0001i − |0010i + |0100i + |1000i , 4 2 final states with variation in λ (using the state (14) as 1  the channel). |b i = |1110i + |1101i + |1011i + |0111i , 5 2 1  |b6i = − |1110i + |1101i + |1011i − |0111i , i 2 Then we calculate the quantum discord of the state ρB as 1  given by [10]. Since there are sixteen different outcomes |b7i = − |1110i + |1101i − |1011i + |0111i , possible for Alice, therefore there are sixteen different 2 i ρB’s. Note that as Alice is communicating her results 1  i |b8i = − |1110i − |1101i + |1011i + |0111i , to Bob, Bob’s state would collapse to one of the ρB’s. 2 Thus, we compute the discord of each such state and then 1  |b9i = |0000i + |0011i + |1100i + |1111i , average it over (which means multiplying the probability 2 (N ) of the occurrence of state to the quantum discord 1  i |b i = |0000i + |0011i − |1100i − |1111i , of the state ρi ) to find the average discord δ. 10 2 B 1  P N δ(ρi ) |b11i = |0000i − |0011i + |1100i − |1111i , δ = i i B . (13) 2 P N 1  i i |b12i = − |0000i + |0011i + |1100i − |1111i , 2 Interestingly, from the Fig.1, we find that upto λ ≈ 1  0.82047 the average discord of Bob’s state is more than |b i = |0101i + |0110i + |1010i + |1001i , 13 2 the initial Werner state posses by Alice. 1  Note that for some of the post-selected outcomes ρi ’s |b i = − |0101i − |0110i + |1010i + |1001i , B 14 2 have a higher Quantum discord than the initial Werner 1  state. Hence, we calculate the average discord and it |b i = |0101i − |0110i − |1010i + |1001i and 15 2 comes out to be more than that of Alice’s initial state 1  for most of the range of λ. However, here we are not |b i = |0101i − |0110i + |1010i − |1001i . (9) 16 2 interested in the individual states which have a higher quantum discord and focus on the quantum discord that The measurement is mathematically defined as Bob would find after averaging over all the outcomes.

Mˆ i = |biihbi| ⊗ I, (10) B. Example 2 where I is the four dimensional identity matrix. Alice sends her outcome to Bob using a classical channel. Then Bob’s state would collapse to We took another 4-qubit channel, 1  † ρi = Tr [Mˆ ρ Mˆ ], (11) |ΨiΩ =√ |00iA|11iB + |01iA|01iB + |01iA|10iB + |10iA|01iB B 1234 i T i 6  upto some normalization constant Ni which is the prob- +|10iA|10iB + |11iA|00iB . (14) B ability of occurrence of outcome ρi , is We choose the measurement basis same as in Eq. (9). ˆ ˆ † Ni = Tr[MiρT Mi ]. (12) Here also we get a similar kind of result. From Fig.2, it 4

1.0 1.0 initial state discord λ≈0.678 final state average discord 0.8 0.8 λ≈0.7722 0.6 0.6 δ ℱ 0.4 0.4 initial state fidelity 0.2 0.2 final state average fidelity 0.0 0.0 0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0 λ λ

FIG. 3: The red line shows the quantum discord as FIG. 4: The red line shows the teleportation fidelity for function of parameter λ for the initial Alice’s Werner the Werner state as variation in λ, the green line shows state. The green line shows the average discord of the the upper bound or the maximum average fidelity final states with variation in λ (using the W state (15) possible for the final states res. as λ is varied (using the as the channel). cluster state given inIIA as the channel).

is clear that upto λ ≈ 0.8876 the average discord of final III. APPLICATION state is larger than the initial state. Essentially Werner state is maximally entangled state with some isotropic noise added. Consider a scenario C. Example 3 where Alice wants to teleport a 1-qubit state using the Werner state as the channel. We ask the question that We can get similar kind of result if we take a 3-qubit if the above protocol (see sectionII) is applied and then state as a channel instead of the 4-qubit channel. How- the final states are used as channels instead of the initial ever, Alice needs to send only 3 bits of classical informa- Werner state, can one increase the teleportation fidelity? tion for this case. Let’s take a three qubit W state Verstraete and Verschelde in [19] have found the upper and lower bounds of fidelity for any teleportation scheme. 1   ∗ |ΨiW = √ |0iA|01iB + |0iA|10iB + |1iA|00iB . (15) Given the channel ρ the fidelity F (ρ) is 3 1 N (ρ)  1 We choose the measurement basis as 1 + ≤ F ∗(ρ) ≤ (1 + N (ρ)), 2 q N (ρ) 2 1  1 + 1 − ( )2 |b i = |000i + |100i + |011i + |111i , C(ρ) 1 2 (17) 1  where N (ρ) is the negativity [20] of the state ρ and C(ρ) |b i = |000i + |100i − |011i − |111i , 2 2 is the concurrence [21]. We calculate these two quantities 1  as given in [20, 21] for the initial state and then find the |b i = |000i − |100i + |011i − |111i , 3 2 bounds. For Werner state these two bounds are same. 1  We do the same calculation for the final states (which |b i = − |000i + |100i + |011i − |111i , 4 2 can be obtained by applying our protocol as described in 1  sectionII) to find the upper and lower bounds and then |b i = |001i + |010i + |101i + |110i , ∗ 5 2 average it over. The average teleportation fidelity F (ρ) 1  will belong to this range specified by lower and upper |b6i = |001i − |010i − |101i + |110i , bound. We will compare our results graphically for those 2 1  three examples described in the previous section. In the |b i = |001i + |010i − |101i − |110i and figures we have only showed the upper bound as this is 7 2 the maximum achievable fidelity. 1  |b i = − |001i + |010i − |101i + |110i . (16) First we consider the example in subsectionIIA. We 8 2 find (as shown in Fig.4) that when the final states are We get similar kind of behavior like two previous exam- used as the resource state instead of the Werner state, ples. From Fig.3, it is clear that upto λ ≈ 0.7722 the the upper bound or the maximum average teleportation average discord of final state is larger than the initial fidelity is more than the fidelity of the initial Werner state. state (for most of the range of λ). Interestingly, we find 5

1.0 λ≈0.7582 1.0 λ≈0.6858 0.8 0.8

0.6 0.6 ℱ ℱ 0.4 0.4 initial state fidelity initial state fidelity 0.2 0.2 final state average fidelity final state average fidelity 0.0 0.0 0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0 λ λ

FIG. 5: The red line shows the teleportation fidelity for FIG. 6: The red line shows the teleportation fidelity for the Werner state as variation in λ, the green line shows the Werner state as variation in λ, the green line shows the upper bound or the maximum average fidelity the upper bound or the maximum average fidelity possible for the final states res. as λ is varied (using the possible for the final states res. as λ is varied (using the state given inIIB as the channel). state given inIIC as the channel). that the upper bound of the average fidelity and discord λ ≈ 0.6858 to λ ≈ 0.7722. Although we see for the exam- both are decreasing with λ. This can be easily checked ples considered above that when quantum discord is more by looking at Fig.1 and Fig.4. Moreover, we find that than the initial state discord, the average teleportation the λ for which the average quantum discord of the fi- fidelity is also more than the fidelity of the initial state nal states is equal to the initial Werner state is differ- (for some range of λ) but a direct connection can’t be es- ent from the λ where the teleportation fidelity using the tablished from these results. These above examples give Werner state is equal to the fidelity when using the final us some insight about the usefulness of discord in such states. In this region one can see that average quantum a protocol. But whether it will be true for any general discord for the final states is more, but the teleportation case needs to be further explored. fidelity is less than the initial Werner state. From Fig.1 and Fig.4, this region corresponds to the values from λ ≈ 0.678 to λ ≈ 0.82047. This difference may be due to the fact that quantum discord and quantum telepor- IV. CONCLUSION tation are not comparable. Proper optimization of the measurement basis may reduce this gap. Nonetheless, We have shown that it is possible for Alice to prepare a one important fact is that even when the initial Werner state at Bob with a higher amount of quantum discord by state is separable, we can use the final state as a channel sharing an entangled channel, local operations and clas- for quantum teleportation. sical communications. We also see that when the quan- Now we find the teleportation fidelity of the final state tum discord of Alice’s state increases then the average for the example given in subsectionIIB. We see a similar quantum discord of Bob’s state decreases. Moreover, we nature as the 4-qubit cluster state. In this example as showed above that the increment of the quantum discord we can see from Fig2 and Fig5, the region of λ (for in our protocol may be used as a way to increase fidelity which quantum discord for the final states are more in of one qubit teleportation. Our results could also be ap- this region, but the teleportation fidelity is less than the plied to other quantum communication tasks like remote initial Werner state) is from λ ≈ 0.7582 to λ ≈ 0.8876. state preparation, entanglement distribution, etc. How- Again we find the teleportation fidelity using the 3- ever our results are not optimized, so it might be possible qubit W state as the resource state. The measurement that there could exist some basis for the given quantum basis is given by Eq. (16). For the 3-qubit state the na- channel such that the quantum discord of Alice’s Werner ture of teleportation fidelity is different from the 4-qubit state is always less than Bob’s prepared state. states. From Fig.3 and Fig.6 we see that the teleporta- tion fidelity behaves completely opposite of the quantum discord. The quantum discord decreases, but the tele- portation fidelity increases as λ increases. However, in V. ACKNOWLEDGMENT this case the region (for which quantum discord for the final states are more in this region, but the teleporta- We would like to thank Satyabrata Adhikari and Pras- tion fidelity is less than the initial Werner state) is from anta Kumar Panigrahi for useful discussions. 6

[1] C. H. Bennett, G. Brassard, C. Cr´epeau, R. Jozsa, A. [11] L. Henderson and V. Vedral, J. Phys. A: Math. Gen. 34 Peres, and W. K. Wootters, Phys. Rev. Lett. 70 1895, 6899 (2001). (1993). [12] A. Datta, A. Shaji, and C. M. Caves, Phys. Rev. Lett [2] B. Daki´cet. al., Nat. Phys. 8 666 (2012). 100 050502 (2008). [3] P. Horodecki, J. Tuziemski, P. Mazurek, and R. [13] G. Adesso, T. R. Bromley and M. Cianciaruso, J. Phys. Horodecki, Phys. Rev. Lett. 112 140507 (2014). A: Math. Theor. 49 473001 (2016). [4] Michal Micuda, Robert Starek, Petr Marek, Martina [14] A. Streltsov, Quantum Correlations Beyond Entangle- Mikova, Ivo Straka, Miroslav Jezek, Toshiyuki Tashima, ment and Their Role in Quantum Information Theory Sahin K. Ozdemir, and Mark Tame, Optics Express 25 (Springer, Berlin, 2015). 7839-7848 (2017). [15] D. Girolami, and G. Adesso, Phys. Rev. A 83 052108 [5] T. Tashima, M. S. Tame, S. K. Ozdemir, F. Nori, (2011). M. Koashi, and H. Weinfurter, Phys. Rev. A 94 [16] Sahin Kaya Ozdemir, Karol Bartkiewicz, Yu-xi Liu, and 052309(2016). Adam Miranowicz, Phys. Rev. A76 042325(2007). [6] J. Oppenheim, M. Horodecki, P. Horodecki, and R. [17] Karol Bartkiewicz, Adam Miranowicz, and Sahin Kaya Horodecki, Phys. Rev. Lett. 89 180402 (2002). Ozdemir, Phys. Rev. A80 032306(2009). [7] Alessandro Ferraro, and Matteo G. A. Paris, Phys. Rev. [18] B. J.-D´ıaz,J. M. Burdis, and F. Tabakin, Comp. Phys. Lett. 108 260403 (2012). Comm. 180 474 (2009). [8] E. Agudelo, J. Sperling, and W. Vogelr, Phys. Rev. A87 [19] F. Verstraete, and H. Verschelde, Phys. Rev. Lett. 90 033811 (2013). 097901 (2003). [9] Vlad Gheorghiu, Marcos C. de Oliveira, and Barry C. [20] G. Vidal, and R. F. Werner, Phys. Rev. A 65 032314 Sanders, Phys. Rev. Lett. 115 030403 (2015). (2002). [10] H. Ollivier and W. H. Zurek, Phys. Rev. Lett. 88 017901 [21] William K. Wootters, Phys. Rev. Lett. 80 2245 (1998). (2001).