{"id":1575,"date":"2023-07-10T03:40:02","date_gmt":"2023-07-09T18:40:02","guid":{"rendered":"https:\/\/saraheee.com\/?p=1575"},"modified":"2023-08-23T12:24:16","modified_gmt":"2023-08-23T03:24:16","slug":"review-do-you-really-need-to-disguise-normal-servers-as-honeypots","status":"publish","type":"post","link":"https:\/\/saraheee.com\/ko\/2023\/07\/review-do-you-really-need-to-disguise-normal-servers-as-honeypots\/","title":{"rendered":"[review #1] Do You Really Need to Disguise Normal Servers as Honeypots?"},"content":{"rendered":"

Contents<\/h3>\n\n\n\n
0. How I got into <\/strong>(Game Theory)\n\n1. Basic Notions<\/strong>\n  1) Pooling \/ (Semi-)Separating equilibrium\n  2) Cho and Kreps criterion (Intuitive criterion)\n\n2. Summary (paper)<\/strong>\n\n3. Review (paper)<\/strong>\n  - Background\n  - Honeypot deception game model\n  - Analysis\n\n4. Analysis<\/strong>\n  1) Identify an attacker's strategy for which there exists a semi-separating PBE in Theorem IV.1.\n+ finding separating\/pooling equilibrium\n  2) Find all PBEs in a full-featured honeypot game. Then, check if there is a PBE that satisfies the Cho-Kreps criterion.\n\n5. Conclusions<\/strong>\n  1) Existence of semi-separating PBEs\n  2) PBE that satisfy the Cho-Kreps criterion\n\n6. References<\/strong><\/pre>\n\n\n\n
0. How I got into Game Theory<\/b><\/mark><\/span><\/h3>\n\n\n\n
\uac8c\uc784 \uc774\ub860\uc5d0 \ub300\ud55c \uc81c \uad00\uc2ec\uc740 \uc218\ud559\uc744 \uc2e4\uc0dd\ud65c\uc5d0 \uc801\uc6a9\ud558\ub824\ub294 \uc5f4\uc815\uacfc, \uc81c \uad00\uc2ec\uc0ac\ub97c \uc54c\uc544\ubd10 \uc900 \ud55c \uc120\ubc30\uc758 \uc81c\uc758\ub85c \uc2dc\uc791\ub418\uc5c8\uc2b5\ub2c8\ub2e4.<\/p>\n\n\n\n
\uc0ac\uc2e4 \uc800\ub294 \uc9c1\uc811\uc801\uc73c\ub85c \uc218\ud559 \uae30\uc220\uc744 \ubcf4\uc548\uc5d0 \uc801\uc6a9\ud560 \uc218 \uc788\ub294 \uc0ac\ub840\ub85c, \uc554\ud638\ud559\uc774\ub098 \uc804\uc790\uc11c\uba85 \ub4f1 \ud1b5\uacc4\uc801\uc778 \uaddc\uce59\uc774\ub098 \ud328\ud134\uc744 \ubd84\uc11d\ud558\ub294 \uc6a9\ub3c4\ub85c\ub9cc \uc0dd\uac01\ud588\uc2b5\ub2c8\ub2e4. \ud558\uc9c0\ub9cc \uac8c\uc784 \uc774\ub860\uc744 \ud0d0\uad6c\ud558\uba74\uc11c \ubcf4\uc548 \uae30\uc220\uc758 \uc6d0\ub9ac\ub098, \uc2e4\uc81c \uc2dc\ub098\ub9ac\uc624\uc5d0 \uac8c\uc784 \uc774\ub860\uc774 \uc5b4\ub5bb\uac8c \uc801\uc6a9\ub418\ub294\uc9c0 \uad6c\uccb4\ud654\ud560 \uc218 \uc788\uc5c8\uc2b5\ub2c8\ub2e4.<\/p>\n\n\n\n
\ud655\ub960 \uc774\ub860\uc774\ub098 \ucd5c\uc801\ud654, \ubbf8\uc801\ubd84\ud559\uacfc \uac19\uc740 \uc218\ud559\uc801 \uae30\uc220\uc5d0 \ud06c\uac8c \uc758\uc874\ud558\ub294 \uac8c\uc784 \uc774\ub860 \ubd84\uc57c\uc5d0\uc11c, \uc800\ub294 \ub2e4\uc591\ud55c \uc804\ub7b5\uc801 \uc2dc\ub098\ub9ac\uc624\ub97c \ubaa8\ub378\ub9c1\ud558\uace0 \ubd84\uc11d\ud558\uace0\uc790 \uacf5\ubd80 \uc911\uc785\ub2c8\ub2e4. \uadf8\ub9ac\uace0 \uc5f0\uad6c \uc900\ube44 \ub2e8\uacc4\ub85c \ub2e4\uc74c\uc758 \ub17c\ubb38\uc744 \uccab \ubc88\uc9f8 \ub9ac\ubdf0 \ub17c\ubb38\uc73c\ub85c \uc18c\uac1c\ud569\ub2c8\ub2e4.<\/p>\n\n\n\n
\uc774 \ub17c\ubb38\uc744 \ubd84\uc11d\ud558\uba74\uc11c \uc800\ub294 \ucd5c\uc801\uc758 \uc758\uc0ac \uacb0\uc815 \uc804\ub7b5\uc5d0 \ub300\ud55c \ud1b5\ucc30\ub825\uc744 \ud30c\uc545\ud558\uace0 \ucd5c\uc801\uc758 \uacb0\uacfc\ub97c \uc608\uce21\ud560 \uc218 \uc788\uc5c8\uc2b5\ub2c8\ub2e4.<\/p>\n\n\n\n
1. Basic Notions<\/mark><\/strong><\/h3>\n\n\n\n
#1. Pooling \/ (Semi-)Separating equilibrium<\/mark><\/strong><\/h4>\n\n\n\n
A. Pooling equilibrium<\/strong><\/h5>\n\n\n\n
an equilibrium in which all types of sender send the same message.<\/p>\n\n\n\n
B. Separating equilibrium<\/strong><\/h5>\n\n\n\n
an equilibrium in which all types of sender send different messages.<\/p>\n\n\n\n
C. Semi-Separating equilibrium<\/strong><\/h5>\n\n\n\n
(Partially-pooling equilibrium<\/strong> is) an equilibrium in which some types of sender send the same message, while some others sends some other messages.<\/p>\n\n\n\n
#2. Cho and Kreps criterion (Intuitive criterion)<\/mark><\/strong><\/h4>\n\n\n\n
\uc9c1\uad00\uc801 \uae30\uc900(Intuitive criterion)\uc740 \uc2e0\ud638 \uac8c\uc784\uc5d0\uc11c \ud3c9\ud615\uc744 \uac1c\uc120\ud558\uae30 \uc704\ud55c \uae30\ubc95\uc774\ub2e4.<\/p>\n\n\n\n
\uc2e0\ud638 \uac8c\uc784\uc5d0\ub294 \uc77c\ubc18\uc801\uc73c\ub85c \ub9ce\uc740 \ub0b4\uc26c \ud3c9\ud615\uc774 \uc874\uc7ac\ud558\uace0, \uc77c\ubd80 \ud3c9\ud615\uc740 \ud3c9\ud615\uc5d0 \uc18d\ud55c \ud50c\ub808\uc774\uc5b4\uac00 \ud3c9\ud615 \uacbd\ub85c\uc5d0\uc11c \ubc97\uc5b4\ub09c \ubd80\ub2f9\ud55c \ubbff\uc74c\uc744 \uac16\uac8c \ub41c\ub2e4\ub294 \uc810\uc5d0\uc11c \ubd88\ud569\ub9ac\ud558\ub2e4.
\uc774\ub7ec\ud55c \ubd88\ud569\ub9ac\ud55c \ud3c9\ud615\uc744 \ubc30\uc81c\ud558\uae30 \uc704\ud574 \uc77c\ubc18\uc801\uc73c\ub85c \ub0b4\uc26c \ud3c9\ud615\uc744 \ub2e4\uc591\ud55c \ubc29\uc2dd\uc73c\ub85c \uac1c\uc120\ud558\uace0 \uc788\uc73c\uba70,
\uc774\ub7ec\ud55c \uac1c\uc120 \uc0ac\ud56d \uc911 Cho\uc640 Kreps\uc758 \uc9c1\uad00\uc801\uc778 \uae30\uc900\uc774 \uac00\uc7a5 \uac15\ub825\ud558\uace0 \uac00\uc7a5 \ub110\ub9ac \uc0ac\uc6a9\ub418\ub294 \uac83\uc73c\ub85c \ubcf4\uc778\ub2e4.<\/p>\n\n\n\n
\uc774\ub294 \uac00\ub2a5\ud55c \ubc1c\uc2e0\uc790 \uc720\ud615\uc744 \ud3c9\ud615\uc5d0\uc11c \ubc97\uc5b4\ub09c \uba54\uc2dc\uc9c0\ub85c \uc774\ud0c8\ud558\uc5ec \ub354 \ub192\uc740 \ud6a8\uc6a9 \uc218\uc900\uc744 \uc5bb\uc744 \uc218 \uc788\ub294 \uc720\ud615\uacfc, \ud3c9\ud615\uc5d0\uc11c \ubc97\uc5b4\ub09c \uba54\uc2dc\uc9c0\uac00 \uc9c0\ubc30\uc801\uc774\uc9c0 \uc54a\uc740 \uc720\ud615\uc73c\ub85c \uc81c\ud55c\ud55c\ub2e4.<\/p>\n\n\n\n
\ub530\ub77c\uc11c \uc774 \uae30\uc900\uc740 \uac00\ub2a5\ud55c \uacb0\uacfc \uc2dc\ub098\ub9ac\uc624\ub97c \uc904\uc774\ub294 \uac83\uc744 \ubaa9\ud45c\ub85c \ud55c\ub2e4.<\/p>\n\n\n\n
<\/p>\n\n\n\n
Do You Really Need to Disguise Normal Servers as Honeypots?<\/strong><\/mark><\/h4>\n\n\n\n
2. Summary<\/mark><\/strong><\/h3>\n\n\n\n
Index Terms – cybersecurity, game theory, honeypot, signaling game<\/p>\n\n\n\n
Compare and analyze three attacker-defender games against honeypot detection techniques
Research on honeypot deception tactics and their effectiveness in mitigating Cyber Threats<\/p>\n\n\n\n
3. Review<\/mark><\/strong><\/h3>\n\n\n\n
II. BACKGROUND<\/strong><\/h4>\n\n\n\n
B. Signaling Game<\/strong><\/h5>\n\n\n\n
For off-equilibrium paths, the beliefs can be arbitrary.
However, some arbitrary beliefs can be irrational so that a PBE which relies on such beliefs can be eliminated with advanced refinement rules.<\/p>\n\n\n\n
    \n
  • I.-K. Cho and D. M. Kreps, \u201cSignaling games and stable equilibria,\u201d The Quarterly Journal of Economics, vol. 102, no. 2, pp. 179\u2013221, 1987.<\/em><\/li>\n<\/ul>\n\n\n\n
    III. HONEYPOT DECEPTION GAME MODEL<\/strong><\/h4>\n\n\n\n
    A. The attacker-defender honeypot game scenario<\/strong><\/h5>\n\n\n\n
    \"\"<\/figure>\n\n\n\n
    \"\"<\/figure>\n\n\n\n
    <\/p>\n\n\n\n
    type \\ signal<\/td>honeypot (signal h)<\/td>normal (signal n)<\/td><\/tr>
    Honeypot<\/td>Obvious Honeypot (Fake service)<\/td>Honeypot (Fake service)
    – honeypot-as-normal<\/td><\/tr>
    Normal<\/td>Fake Honeypot (Real service)
    – normal-as-honeypot<\/td>    		Server (Real service)<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n
    B. The attacker-defender Game Model<\/strong><\/h5>\n\n\n\n
    (a) Honest game
    : defenders do not use any honeypot techniques<\/p>\n\n\n\n
    (b) Semi-featured honeypot game
    1) \\(c_t > c_p\\) : revealing concrete attack methods will be more costly than simply proving nodes
    2) \\(c_h < c_a\\) : the cost by successful attacks will be the most threatening<\/p>\n\n\n\n
    (c) Full-featured honeypot game
    1) \\(c_n < c_h\\) : the cost to deploy a normal-as-honeypot node can be handled in a software method to an existing normal node
    2) \\(c_n < c_a\\), \\(c_h < c_a\\) : the cost by successful attacks will be the most threatening<\/p>\n\n\n\n
    \ub530\ub77c\uc11c, \ubc29\uc5b4\uc790\uc5d0\uac8c \ubc1c\uc0dd\ud558\ub294 \ube44\uc6a9\uc740 \\(c_n, c_h, c_a\\) \uc21c\uc73c\ub85c \uc801\uc73c\uba70, \uacf5\uaca9\uc790\ub294 \\(c_p\\)\ubcf4\ub2e4 \\(c_t\\)\uac00 \ub354 \ub9ce\uc740 \ube44\uc6a9\uc774 \ubc1c\uc0dd\ud55c\ub2e4.
    (Thus, the cost figure for the defender is \\(c_n < c_h < c_a\\)<\/strong> and for the attacker is \\(c_p < c_t\\)<\/strong>.)<\/p>\n\n\n\n
    IV. ANALYSIS<\/strong><\/h4>\n\n\n\n
    B. Equilibria Analysis on Semi-featured Honeypot Game<\/strong><\/h5>\n\n\n\n
      \n
    • Let (a, (B, C)) denote a strategy set where \u2018a\u2019 is the H- type defender\u2019s action, \u2018B\u2019 is the attacker\u2019s action when the signal (h) is received, and \u2018C\u2019 is the attacker\u2019s action when the signal (n) is received.<\/li>\n<\/ul>\n\n\n\n
      C. Equilibria Analysis on Full-featured Honeypot Game<\/strong><\/h5>\n\n\n\n
        \n
      • Let ((a, b), (C, D)) denote a strategy set where \u2018a\u2019 is the H-type defender\u2019s action, \u2018b\u2019 is the N-type defender\u2019s action, \u2018C\u2019 is the attacker\u2019s action when the signal (h) is received, and \u2018D\u2019 is the attacker\u2019s action when the signal (n) is received.<\/li>\n<\/ul>\n\n\n\n
        4. Analysis<\/mark><\/strong><\/h3>\n\n\n\n
        Objectives\n#1: Theorem IV.1\uc5d0\uc11c \ubc18\ubd84\ub9ac PBE\uac00 \uc874\uc7ac\ud558\ub294 \uacf5\uaca9\uc790\uc758 \uc804\ub7b5\uc744 \uad6c\ud55c\ub2e4.\n#2: Full-featured \ud5c8\ub2c8\ud31f \uac8c\uc784\uc5d0\uc11c \ubaa8\ub4e0 PBE\ub97c \ucc3e\ub294\ub2e4. \uadf8 \uc774\ud6c4, Cho-Kreps\uc758 \uc9c1\uad00\uc801\uc778 \uae30\uc900\uc5d0 \uc758\ud574 (\ub9cc\uc871\ud558\ub294\uc9c0 \uc5ec\ubd80\uc5d0 \ub530\ub77c) PBE\uac00 \uc9c0\uc6cc\uc9c0\uac70\ub098 \ud3c9\ud615\uc774 \uc720\uc9c0\ub418\ub294\uc9c0 \ud655\uc778\ud55c\ub2e4).\n\n#1: Identify an attacker's strategy for which there exists a semi-separating PBE in Theorem IV.1.\n#2: Find all PBEs in a Full-featured honeypot game. Then, check that the PBE is erased by the intuitive criterion of Cho-Kreps or that equilibrium is preserved(satisfies the Cho-Kreps criterion).<\/pre>\n\n\n\n
        comments #1<\/mark><\/strong><\/h4>\n\n\n\n
        Proof. Let there exist a semi-separating PBE with an attacker\u2019s strategy \\(((\\sigma_a, 1-\\sigma_a), (\\sigma_b, 1-\\sigma_b))\\) with a belief (p, q).<\/mark><\/p>\n\n\n\n
        Applying the indifference principle, the utilities of the defender\u2019s actions given the honeypot type should be same<\/mark><\/p>\n\n\n\n
        Suppose a semi-separating equilibrium exists for all arbitrary non-zero payoff parameter.<\/p>\n\n\n\n
        $$ u_h^H(A) = \\sigma_{pA}(b_d-c_h)+(1-\\sigma_{pA}(-c_h) = \\sigma_{pA}b_d=c_h $$<\/p>\n\n\n\n
        $$ u_n^H(A) = \\sigma_{qA}(b_d-c_h)+(1-\\sigma_{qA}(-c_h) = \\sigma_{qA}b_d=c_h $$<\/p>\n\n\n\n
        \u21d2 \\(\\sigma_{pA} = \\sigma_{qA}\\) – (1)<\/p>\n\n\n\n
        $$ u_h^N(A) = \\sigma_{(1-p)A}(-c_a)+(1-\\sigma_A)\\cdot0 = -c_a\\sigma_{(1-p)A} $$<\/p>\n\n\n\n
        $$ u_n^N(A) = \\sigma_{(1-q)A}(-c_a-c_n)+(1-\\sigma_{(1-q)A}(-c_n)=-\\sigma_{(1-q)A}c_a-c_n $$<\/p>\n\n\n\n
        \u21d2 \\(c_a(\\sigma_{(1-p)A}-\\sigma_{(1-q)A}) = c_n\\) – (2)<\/p>\n\n\n\n
        That is, by (1) and (2), \\(c_n = 0\\), which contradicts the assumption that \\(c_n \\neq 0\\).
        Therefore, there is no semi-separating equilibrium for any nonzero \\(c_n\\).<\/p>\n\n\n\n
        + finding separating\/pooling equilibrium<\/mark><\/strong><\/h4>\n\n\n
        \n
        \"\"<\/figure><\/div>\n\n\n
        (1) Represent defender’s strategy<\/strong> \\(\\sigma_d\\) with two parameters:<\/p>\n\n\n\n
        \\(\\sigma_d\\) = (\\(\\sigma_d^H(h), \\sigma_d^N(h)\\)) = (x, y)<\/p>\n\n\n\n
        \\(\\sigma_d^H(h)\\): for honeypot type, \\(\\sigma_d^N(h)\\): for normal type<\/p>\n\n\n\n
        (2) Compute attacker\u2019s Bayesian beliefs<\/strong> given \\(\\sigma_d\\):<\/p>\n\n\n\n
        in signal h,<\/p>\n\n\n\n
        $$ p \\equiv \\mu_{att}^h(A) = \\frac{p_{\\sigma}(h)}{p_{\\sigma}(I_h)} = \\frac{P_hy}{P_hy + (1-P_h)x} \\\\ = \\frac{P_hy}{x-P_h(x-y)} $$<\/p>\n\n\n\n
        in signal n,<\/p>\n\n\n\n
        $$ q \\equiv \\mu_{att}^n(A) = \\frac{p_{\\sigma}(n)}{p_{\\sigma}(I_n)} = \\frac{P_h(1-y)}{P_h(1-y) + (1-P_h)(1-x)} \\\\ = \\frac{P_h(1-y)}{1-x+P_h(x-y)} $$<\/p>\n\n\n\n
        (3) Find attacker\u2019s best response<\/strong> to \\(\\sigma_d\\) given beliefs:<\/p>\n\n\n\n
        attacker\u2019s BR @ \\(I_h = \\begin {cases} A & \\text{if } p \\geq \\frac{1}{2} \\\\ L & \\text{if } p < \\frac{1}{2} \\end {cases}\\)<\/p>\n\n\n\n
        attacker’s BR @ \\(I_n = \\begin {cases} A & \\text{if } q \\geq \\frac{1}{2} \\\\ L & \\text{if } q < \\frac{1}{2} \\end {cases}\\)<\/p>\n\n\n\n
        (Since one goal of this paper is to determine if the fake strategy for normal nodes is efficient, let’s set Pr(\\(t_d\\) = H) < 1\/2 for the defender’s type \\(t_d\\).
        Therefore, \\(P_h\\) < 1\/2.)<\/mark>
        – \uc544\ub798 \ud3c9\ud615\uc744 \uad6c\ud558\uae30 \uc704\ud55c \uac00\uc815<\/p>\n\n\n\n
        1) (x, y) = (1, 1)<\/strong><\/h4>\n\n\n\n
        a pooling strategy<\/em> in which both types send message h<\/strong><\/h5>\n\n\n\n
        p = \\(P_h\\) but q is unrestricted<\/p>\n\n\n\n
        If h was sent, A is BR given p = \\(P_h\\) < 0.5
        Suppose n was sent and q < 0.5 \u2192 A is BR
        Not an eqbm<\/p>\n\n\n\n
        Suppose n was sent but q \u2265 0.5 and \\(c_t > c_p\\) \u2192 L is BR
        Both types wouldn\u2019t deviate, and thus it is an eqbm<\/p>\n\n\n\n
        There exists on perfect Bayesian eqbm in pooling strategies:<\/p>\n\n\n\n
        \u27e8(hh, AL), p = \\(P_h\\), q \u2265 0.5\u27e9<\/strong><\/p>\n\n\n\n
        But, not an eqbm this normal type would deviate to n<\/p>\n\n\n\n
        if \\(P_h\\) > 0.5, pooling strategies: \u27e8(hh, LA), p = \\(P_h\\), q \u2264 0.5\u27e9<\/strong><\/p>\n\n\n\n
        But, also not an eqbm this normal type would deviate to n<\/p>\n\n\n\n
        2) (x, y) = (1, 0)<\/strong><\/h4>\n\n\n\n
        a separating strategy in which the honeypot type sends h but the normal type sends n.<\/strong><\/h5>\n\n\n\n
        attacker exactly infers defender type from the message sent A is a BR to h and L is a BR to n<\/p>\n\n\n\n
        \u27e8(hn, AL), p = \\(P_h\\), q \u2264 0.5\u27e9<\/mark><\/strong><\/p>\n\n\n\n
        In this setting, the attacker’s beliefs can be updated to p = 1 and q = 0.
        \ubc29\uc5b4\uc790\uac00 \uc5b4\ub5a4 \uc720\ud615\uc744 \uc120\ud0dd\ud558\ub4e0 \ud5c8\ub2c8\ud31f\/\uc77c\ubc18 \uc720\ud615\uc758 \uacbd\uc6b0 \uac01\uac01 \\(b_d-c_h\\), 0\ubcf4\ub2e4 \ub192\uc740 \ubcf4\uc0c1\uc744 \uc5bb\uc744 \uc218 \uc5c6\ub2e4.<\/p>\n\n\n\n
        3) (x, y) = (0, 1)<\/strong><\/h4>\n\n\n\n
        a separating strategy in which the honeypot type sends n but the normal type sends h.<\/strong><\/h5>\n\n\n\n
        attacker exactly infers defender type from the message sent A is a BR to h and L is a BR to n<\/p>\n\n\n\n
        \u27e8(nh, LA), p = \\(P_h\\), q \u2264 0.5\u27e9<\/strong><\/p>\n\n\n\n
        But, not an eqbm this normal type would deviate to n<\/p>\n\n\n\n
        4) (x, y) = (0, 0)<\/strong><\/h4>\n\n\n\n
        a pooling strategy<\/em> in which both types send message n<\/strong><\/h5>\n\n\n\n
        q = \\(P_h\\) < 0.5 but p is unrestricted<\/p>\n\n\n\n
        \u27e8(nn, AL), p \u2265 0.5, q = \\(P_h\\)\u27e9<\/strong><\/p>\n\n\n\n
        But, not an eqbm this honeypot type would deviate to h<\/p>\n\n\n\n
        If p < 0.5, attacker will Attack whether defender orders Honeypot or Normal
        So defender has an incentive to debate.<\/p>\n\n\n\n
        In the normal type, attacker has no incentive to debate because defender gets \\(-c_a\\) or 0 from n.
        So there is no PBE when p is less than 0.5.<\/p>\n\n\n\n
        comments #2<\/mark><\/strong><\/h4>\n\n\n\n
        Intuitive criterion<\/strong><\/h5>\n\n\n\n
        1) \uc6b0\uc120, out-of-equilibrium\uc744 \uc81c\ud55c\ud568\uc73c\ub85c\uc368, \uc9c1\uad00\uc801\uc774\uc9c0 \uc54a\uc740 \ud3c9\ud615\uc744 \uc81c\uac70\ud560 \uc218 \uc788\ub2e4.
        (off-the-equilibrium\uc774\ub77c\ub294 \uc6a9\uc5b4\uac00 CHO, I. AND KREPS, D. (1987) \ub17c\ubb38\uc5d0\uc11c\ub294 out-of\uc774\ub77c \uc4f0\uc784)<\/p>\n\n\n\n
        2) \ud3c9\ud615\uc744 \uad6c\uc131\ud558\ub294 \uac83\uc740\u201c\ud574\uc11d(interpretations)\u201d\uc758 \uc601\ud5a5\uc744 \ubc1b\uc73c\uba70,
        \ud574\uc11d\uc774\ub780, \ud3c9\ud615\uc5d0\uc11c \ubc97\uc5b4\ub09c \uba54\uc2dc\uc9c0\uc5d0 \ub300\ud574 \uc81c2\uc758 \ud50c\ub808\uc774\uc5b4\uac00 \uc904 \uc218 \uc788\ub294 \uac83\uc744 \uc758\ubbf8
        (\uae30\uc874 \ud3c9\ud615\uc744 \uc774\ud0c8\ud574\uc11c \ub354 \ub192\uc740 \uc218\uc900\uc744 \uc5bb\uc744 \uc218 \uc788\ub294\uc9c0\ub97c \uc758\ubbf8\ud558\ub294 \uac83\uc73c\ub85c \ud574\uc11d\ud568)<\/p>\n\n\n\n
        <\/p>\n\n\n
        \n
        \"\"<\/figure><\/div>\n\n\n
        <\/p>\n\n\n\n
        defender type\uc774 Honeypot, Normal\uc5d0\uc11c \uac01\uac01 signal h, n\uc77c \ub54c Cho and Kreps’ (1987) Intuitive Criterion\uc744 \ub9cc\uc871\ud558\ub294\uc9c0 \ud655\uc778\ud55c\ub2e4.<\/p>\n\n\n\n
        1. Check if the pooling equilibrium in which both types of defender have ‘signal h’ survives the Cho and Kreps\u2019 (1987) Intuitive Criterion.<\/strong><\/h5>\n\n\n\n
        1) if \\(P_h\\) > 0.5, <hh, LA><\/mark><\/strong><\/p>\n\n\n\n
        Cho\uc640 Kreps\uc758 (1987) \uc9c1\uad00\uc801\uc778 \uae30\uc900\uc758 \uccab \ubc88\uc9f8 \ub2e8\uacc4\uc5d0\uc11c,
        equilibrium dominated\ub418\ub294 off-the-equilibrium messages\ub97c \uc81c\uac70\ud55c\ub2e4.
        Honeypot-type defender\uc5d0\uc11c, normal signal\uc774 (honeypot signal \ub300\ube44) \ud3c9\ud615 utility level\uc744 \ud5a5\uc0c1\uc2dc\ud0ac \uc218 \uc788\ub294\uc9c0\ub97c \ud655\uc778\ud574\uc57c \ud55c\ub2e4. \uc989,<\/p>\n\n\n\n
        \\(u_d^*(signal h|Honeypot)\\) < \\(Max u_d(signal n|Honeypot)\\)<\/p>\n\n\n\n
          \n
        • \uc88c\ubcc0\uc758 equilibrium payoff\uac00 \\(-c_h\\)<\/li>\n\n\n\n
        • \uc6b0\ubcc0\uc758 Highest payoff from deviating towards signal n, \\(b_d-c_h\\)<\/li>\n<\/ul>\n\n\n\n
          But this condition is not satisfied: \ud655\uc2e4\ud788 defender\ub294 signal h\uc758 \ud3c9\ud615 \uc0c1\ud0dc\uc5d0\uc11c \\(-c_h\\)\uc758 \ubcf4\uc218\ub97c \uc5bb\uace0, signal n\uc73c\ub85c \uc774\ud0c8\ud558\ub294 \uac83\uc5d0\uc11c \uc5bb\uc744 \uc218 \uc788\ub294 \ubcf4\uc218\ub294 \\(b_d-c_h\\)\uc744 \uc5bb\uae30 \ub54c\ubb38\uc5d0 \uc774 \uc720\ud615\uc740 signal h\uc5d0\uc11c signal n\uc73c\ub85c \ubc97\uc5b4\ub0a0 \uac83\uc774\ub2e4.<\/p>\n\n\n\n
          \\(u_d^*(signal h|Normal)\\) > \\(Max u_d(signal n|Normal)\\)<\/p>\n\n\n\n
            \n
          • \uc88c\ubcc0\uc758 equilibrium payoff\uac00 \\(-c_n\\)<\/li>\n\n\n\n
          • \uc6b0\ubcc0\uc758 Highest payoff from deviating towards signal n, 0<\/li>\n<\/ul>\n\n\n\n
            attacker\uac00 Attack\uc774 \uc544\ub2cc Leave without attack\uc744 \uc120\ud0dd\ud55c\ub2e4\uba74 \uc774 \uc870\uac74\uc740 \uc2e4\uc81c\ub85c \ucda9\uc871\ub420 \uc218 \uc788\ub2e4.
            Normal defender\ub294 \uc774 pooling equilibrium\uc5d0\uc11c \\(-c_n\\) \ub300\uc2e0 0\uc758 \ubcf4\uc218\ub97c \uc5bb\uc744 \uac83\uc774\ub2e4.
            \ub530\ub77c\uc11c defender\uc5d0\uac8c \uc774 separating PBE\uc5d0\uc11c \ubc97\uc5b4\ub0a0 incentives\uac00 \uc874\uc7ac\ud55c\ub2e4.<\/p>\n\n\n\n
            \ub530\ub77c\uc11c attacker\uc758 belief\ub294, signal n\uc758 off-the-equilibrium message\ub97c \uad00\ucc30\ud55c \ud6c4, \\(\\Theta(signal n)\\) = {Honeypot, Normal} \ub450 \uacbd\uc6b0 \ubaa8\ub450 \uc874\uc7ac\ud558\uae30\uc5d0 \uc81c\ud55c\ub41c \uae30\uc900\uc744 \uc124\uc815\ud560 \uc218 \uc5c6\ub2e4.<\/p>\n\n\n\n
            2) if \\(P_h\\) < 0.5, <hh, AL><\/mark><\/strong><\/p>\n\n\n\n
            First step:<\/strong><\/p>\n\n\n\n
            \\(u_d^*(signal h|Honeypot)\\) < \\(Max u_d(signal n|Honeypot)\\)<\/p>\n\n\n\n
              \n
            • \uc88c\ubcc0\uc758 equilibrium payoff\uac00 \\(b_d-c_h\\)<\/li>\n\n\n\n
            • \uc6b0\ubcc0\uc758 Highest payoff from deviating towards signal n, \\(b_d-c_h\\)<\/li>\n<\/ul>\n\n\n\n
              But this condition is not satisfied: \ud655\uc2e4\ud788 defender\ub294 signal h\uc758 \ud3c9\ud615 \uc0c1\ud0dc\uc5d0\uc11c \\(b_d-c_h\\)\uc758 \ubcf4\uc218\ub97c \uc5bb\uace0, signal n\uc73c\ub85c \uc774\ud0c8\ud558\ub294 \uac83\uc5d0\uc11c \uc5bb\uc744 \uc218 \uc788\ub294 \ubcf4\uc218 \ub610\ud55c \\(b_d-c_h\\)\uc73c\ub85c \ub3d9\uc77c\ud558\ub2e4.
              \ub530\ub77c\uc11c \uc774 \uc720\ud615\uc740 signal h\uc5d0\uc11c \ubc97\uc5b4\ub098\uc9c0 \uc54a\uc744 \uac83\uc774\ub2e4.<\/p>\n\n\n\n
              \\(u_d^*(signal h|Normal)\\) > \\(Max u_d(signal n|Normal)\\)<\/p>\n\n\n\n
                \n
              • \uc88c\ubcc0\uc758 equilibrium payoff\uac00 \\(-c_a-c_n\\)<\/li>\n\n\n\n
              • \uc6b0\ubcc0\uc758 Highest payoff from deviating towards signal n, 0<\/li>\n<\/ul>\n\n\n\n
                attacker\uac00 Attack\uc774 \uc544\ub2cc Leave without attack\uc744 \uc120\ud0dd\ud55c\ub2e4\uba74 \uc774 \uc870\uac74\uc740 \uc2e4\uc81c\ub85c \ucda9\uc871\ub420 \uc218 \uc788\ub2e4.
                Normal defender\ub294 \uc774 pooling equilibrium\uc5d0\uc11c \\(-c_a-c_n\\) \ub300\uc2e0 0\uc758 \ubcf4\uc218\ub97c \uc5bb\uc744 \uac83\uc774\ub2e4.
                \ub530\ub77c\uc11c defender\uc5d0\uac8c \uc774 separating PBE\uc5d0\uc11c \ubc97\uc5b4\ub0a0 incentives\uac00 \uc874\uc7ac\ud55c\ub2e4.<\/p>\n\n\n\n
                Second step:<\/strong><\/p>\n\n\n\n
                \ub530\ub77c\uc11c attacker\uc758 belief\ub294, signal n\uc758 off-the-equilibrium message\ub97c \uad00\ucc30\ud55c \ud6c4, \\(\\Theta(signal n)\\) = {Normal}\ub85c \uc81c\ud55c\ub420 \uc218 \uc788\ub2e4.
                \uc989, attacker\ub294 defender\uac00 signal n\uc744 \uc120\ud0dd\ud558\ub294 \uac83\uc744 \uad00\ucc30\ud55c\ub2e4\uba74, \uacf5\uaca9\uc790\ub294 \ubc29\uc5b4\uc790\uac00 Normal Type\uc744 \uc120\ud0dd\ud588\ub2e4\uace0 \ubbff\uc744 \uac83\uc774\ub2e4.
                \uadf8 \uc774\uc720\ub294, signal h\uc758 pooling equilibrium\uc5d0\uc11c \ubc97\uc5b4\ub098\uc11c \uc774\uc775\uc744 \uc5bb\uc744 \uc218 \uc788\ub294 \uc720\uc77c\ud55c defender \uc720\ud615\uc774\uae30 \ub54c\ubb38\uc774\ub2e4.<\/p>\n\n\n\n
                \uc774\ub7ec\ud55c \ub9e5\ub77d\uc5d0\uc11c, \uc774 \uc870\uac74\uc740 Normal type\uc5d0 \ub300\ud574 \uc720\uc9c0\ub41c\ub2e4.
                attacker\ub294 \ubbff\uc74c\uc5d0 \ub530\ub77c signal n\uc744 \ubc1b\uc744 \uacbd\uc6b0, Attack\ud558\ub294 \uac83\uc774 node\ub97c Leave\ud558\ub294 \uac83\ubcf4\ub2e4 \ub354 \ud070 payoff \\((b_s > -c_p)\\)\ub97c \uc5bb\uae30 \ub54c\ubb38\uc5d0, defender \ub610\ud55c \ub354 \ud070 payoff\ub97c \uc5bb\uc744 \uc218 \uc788\ub294 signal n\uc73c\ub85c \ubc97\uc5b4\ub09c\ub2e4. \\((-c_a-c_n < -c_a)\\)<\/p>\n\n\n\n
                \ub530\ub77c\uc11c signal h\uc5d0 \ub300\ud55c pooling equilibrium\uc740 Cho and Kreps\uc758 \uc9c1\uad00\uc801\uc778 \uae30\uc900\uc5d0 \uc758\ud574 \uc0b4\uc544\ub0a8\uc9c0 \ubabb\ud55c\ub2e4.<\/p>\n\n\n\n
                2. Check if the pooling equilibrium in which both types of defender have ‘signal n’ survives the Cho and Kreps\u2019 (1987) Intuitive Criterion.<\/strong><\/h5>\n\n\n\n
                1) if \\(P_h\\) > 0.5, <nn, AL><\/mark><\/strong><\/p>\n\n\n\n
                First step:<\/strong><\/p>\n\n\n\n
                Cho\uc640 Kreps\uc758 (1987) \uc9c1\uad00\uc801\uc778 \uae30\uc900\uc758 \uccab \ubc88\uc9f8 \ub2e8\uacc4\uc5d0\uc11c,
                equilibrium dominated\ub418\ub294 off-the-equilibrium messages\ub97c \uc81c\uac70\ud55c\ub2e4.
                Honeypot-type defender\uc5d0\uc11c, honeypot signal\uc774 (normal signal \ub300\ube44) \ud3c9\ud615 utility level\uc744 \ud5a5\uc0c1\uc2dc\ud0ac \uc218 \uc788\ub294\uc9c0\ub97c \ud655\uc778\ud574\uc57c \ud55c\ub2e4. \uc989,<\/p>\n\n\n\n
                \\(u_d^*(signal n|Honeypot)\\) < \\(Max u_d(signal h|Honeypot)\\)<\/p>\n\n\n\n
                  \n
                • \uc88c\ubcc0\uc758 equilibrium payoff\uac00 \\(-c_h\\)<\/li>\n\n\n\n
                • \uc6b0\ubcc0\uc758 Highest payoff from deviating towards signal h, \\(b_d-c_h\\)<\/li>\n<\/ul>\n\n\n\n
                  attacker\uac00 Leave without attack\uc774 \uc544\ub2cc Attack\uc744 \uc120\ud0dd\ud55c\ub2e4\uba74 \uc774 \uc870\uac74\uc740 \uc2e4\uc81c\ub85c \ucda9\uc871\ub420 \uc218 \uc788\ub2e4.
                  Honeypot defender\ub294 \uc774 pooling equilibrium\uc5d0\uc11c \\(-c_h\\) \ub300\uc2e0 \\(b_d-c_h\\)\uc758 \ubcf4\uc218\ub97c \uc5bb\uc744 \uac83\uc774\ub2e4.
                  \ub530\ub77c\uc11c defender\uc5d0\uac8c \uc774 PBE\uc5d0\uc11c \ubc97\uc5b4\ub0a0 incentives\uac00 \uc874\uc7ac\ud55c\ub2e4.<\/p>\n\n\n\n
                  \\(u_d^*(signal n|Normal)\\) > \\(Max u_d(signal h|Normal)\\)<\/p>\n\n\n\n
                    \n
                  • \uc88c\ubcc0\uc758 equilibrium payoff\uac00 0<\/li>\n\n\n\n
                  • \uc6b0\ubcc0\uc758 Highest payoff from deviating towards signal h, \\(-c_n\\)<\/li>\n<\/ul>\n\n\n\n
                    But this condition is not satisfied: \ud655\uc2e4\ud788 defender\ub294 signal n\uc758 \ud3c9\ud615 \uc0c1\ud0dc\uc5d0\uc11c 0\uc758 \ubcf4\uc218\ub97c \uc5bb\uace0, signal h\uc73c\ub85c \uc774\ud0c8\ud558\ub294 \uac83\uc5d0\uc11c \uc5bb\uc744 \uc218 \uc788\ub294 \ubcf4\uc218\ub294 \\(-c_n\\)\uc73c\ub85c \uc774 \uc720\ud615\uc740 signal n\uc5d0\uc11c \ubc97\uc5b4\ub098\uc9c0 \uc54a\uc744 \uac83\uc774\ub2e4.<\/p>\n\n\n\n
                    Second step:<\/strong><\/p>\n\n\n\n
                    \ub530\ub77c\uc11c attacker\uc758 belief\ub294, signal h\uc758 off-the-equilibrium message\ub97c \uad00\ucc30\ud55c \ud6c4, \\(\\Theta(signal h)\\) = {Honeypot}\ub85c \uc81c\ud55c\ub420 \uc218 \uc788\ub2e4.
                    \uc989, attacker\ub294 defender\uac00 signal h\uc744 \uc120\ud0dd\ud558\ub294 \uac83\uc744 \uad00\ucc30\ud55c\ub2e4\uba74, \uacf5\uaca9\uc790\ub294 \ubc29\uc5b4\uc790\uac00 Honeypot Type\uc744 \uc120\ud0dd\ud588\ub2e4\uace0 \ubbff\uc744 \uac83\uc774\ub2e4.
                    \uadf8 \uc774\uc720\ub294, signal h\uc758 pooling equilibrium\uc5d0\uc11c \ubc97\uc5b4\ub098\uc11c \uc774\uc775\uc744 \uc5bb\uc744 \uc218 \uc788\ub294 \uc720\uc77c\ud55c defender \uc720\ud615\uc774\uae30 \ub54c\ubb38\uc774\ub2e4.<\/p>\n\n\n\n
                    \uc774\ub7ec\ud55c \ub9e5\ub77d\uc5d0\uc11c, \uc774 \uc870\uac74\uc740 Honeypot type\uc5d0 \ub300\ud574 \uc720\uc9c0\ub41c\ub2e4.
                    attacker\ub294 \ubbff\uc74c\uc5d0 \ub530\ub77c signal h\uc744 \ubc1b\uc744 \uacbd\uc6b0, Leave without attack\ud558\ub294 \uac83\uc774 Attack\ud558\ub294 \uac83\ubcf4\ub2e4 \ub354 \ud070 payoff \\((-c_t < -c_p)\\)\ub97c \uc5bb\uc73c\ub098, defender\ub294 payoff\uac00 \ub3d9\uc77c\ud558\ubbc0\ub85c signal h, n \uc911 \uc5b4\ub290 \uac83\uc744 \uc120\ud0dd\ud574\ub3c4 \uc0c1\uad00\uc774 \uc5c6\ub2e4. \\((-c_h = -c_h)\\)<\/p>\n\n\n\n
                    2) if \\(P_h\\) < 0.5, <nn, AA><\/mark><\/strong><\/p>\n\n\n\n
                    First step:<\/strong><\/p>\n\n\n
                    \n
                    \"\"<\/figure><\/div>\n\n\n
                    \\(u_d^*(signal n|Honeypot)\\) < \\(Max u_d(signal h|Honeypot)\\)<\/p>\n\n\n\n
                      \n
                    • \uc88c\ubcc0\uc758 equilibrium payoff\uac00 \\(b_d-c_h\\)<\/li>\n\n\n\n
                    • \uc6b0\ubcc0\uc758 Highest payoff from deviating towards signal h, \\(b_d-c_h\\)<\/li>\n<\/ul>\n\n\n\n
                      But this condition is not satisfied: \ud655\uc2e4\ud788 defender\ub294 signal h\uc758 \ud3c9\ud615 \uc0c1\ud0dc\uc5d0\uc11c \\(b_d-c_h\\)\uc758 \ubcf4\uc218\ub97c \uc5bb\uace0, signal n\uc73c\ub85c \uc774\ud0c8\ud558\ub294 \uac83\uc5d0\uc11c \uc5bb\uc744 \uc218 \uc788\ub294 \ubcf4\uc218 \ub610\ud55c \\(b_d-c_h\\)\uc73c\ub85c \ub3d9\uc77c\ud558\ub2e4.
                      \ub530\ub77c\uc11c \uc774 \uc720\ud615\uc740 signal n\uc5d0\uc11c \ubc97\uc5b4\ub098\uc9c0 \uc54a\uc744 \uac83\uc774\ub2e4.<\/p>\n\n\n
                      \n
                      \"\"<\/figure><\/div>\n\n\n
                      \\(u_d^*(signal n|Normal)\\) > \\(Max u_d(signal h|Normal)\\)<\/p>\n\n\n\n
                        \n
                      • \uc88c\ubcc0\uc758 equilibrium payoff\uac00 \\(-c_a\\)<\/li>\n\n\n\n
                      • \uc6b0\ubcc0\uc758 Highest payoff from deviating towards signal h, \\(-c_n\\)<\/li>\n<\/ul>\n\n\n\n
                        attacker\uac00 Attack\uc774 \uc544\ub2cc Leave without attack\uc744 \uc120\ud0dd\ud55c\ub2e4\uba74 \uc774 \uc870\uac74\uc740 \uc2e4\uc81c\ub85c \ucda9\uc871\ub420 \uc218 \uc788\ub2e4.
                        Honeypot defender\ub294 \uc774 pooling equilibrium\uc5d0\uc11c \\(-c_a\\) \ub300\uc2e0 \\(-c_n\\)\uc758 \ubcf4\uc218\ub97c \uc5bb\uc744 \uac83\uc774\ub2e4.
                        \ub530\ub77c\uc11c defender\uc5d0\uac8c \uc774 PBE\uc5d0\uc11c \ubc97\uc5b4\ub0a0 incentives\uac00 \uc874\uc7ac\ud55c\ub2e4.<\/p>\n\n\n\n
                        Second step:<\/strong><\/p>\n\n\n\n
                        \ub530\ub77c\uc11c attacker\uc758 belief\ub294, signal n\uc758 off-the-equilibrium message\ub97c \uad00\ucc30\ud55c \ud6c4, \\(\\Theta(signal h)\\) = {Normal}\ub85c \uc81c\ud55c\ub420 \uc218 \uc788\ub2e4.
                        \uc989, attacker\ub294 defender\uac00 signal h\uc744 \uc120\ud0dd\ud558\ub294 \uac83\uc744 \uad00\ucc30\ud55c\ub2e4\uba74, \uacf5\uaca9\uc790\ub294 \ubc29\uc5b4\uc790\uac00 Normal Type\uc744 \uc120\ud0dd\ud588\ub2e4\uace0 \ubbff\uc744 \uac83\uc774\ub2e4.
                        \uadf8 \uc774\uc720\ub294, signal h\uc758 pooling equilibrium\uc5d0\uc11c \ubc97\uc5b4\ub098\uc11c \uc774\uc775\uc744 \uc5bb\uc744 \uc218 \uc788\ub294 \uc720\uc77c\ud55c defender \uc720\ud615\uc774\uae30 \ub54c\ubb38\uc774\ub2e4.<\/p>\n\n\n
                        \n
                        \"\"<\/figure><\/div>\n\n\n
                        \uc774\ub7ec\ud55c \ub9e5\ub77d\uc5d0\uc11c, \uc774 \uc870\uac74\uc740 Normal type\uc5d0 \ub300\ud574 \uc720\uc9c0\ub41c\ub2e4.
                        attacker\ub294 \ubbff\uc74c\uc5d0 \ub530\ub77c signal h\uc744 \ubc1b\uc744 \uacbd\uc6b0, Attack\ud558\ub294 \uac83\uc774 Leave without attack\ud558\ub294 \uac83\ubcf4\ub2e4 \ub354 \ud070 payoff \\((b_s > -c_p)\\)\ub97c \uc5bb\uc73c\ub098, defender\ub294 \ub354 \ud070 payoff\ub97c \uc5bb\uc744 \uc218 \uc788\ub294 \uae30\uc874 signal n\uc5d0\uc11c \ubc97\uc5b4\ub098\uc9c0 \uc54a\ub294\ub2e4. \\((-c_a-c_n < -c_a)\\)<\/p>\n\n\n\n
                        \ub530\ub77c\uc11c signal n\uc5d0 \ub300\ud55c pooling equilibrium\uc740 Cho and Kreps\uc758 \uc9c1\uad00\uc801\uc778 \uae30\uc900\uc5d0 \uc758\ud574 \uc0b4\uc544\ub0a8\ub294\ub2e4.<\/mark><\/p>\n\n\n\n
                        5. Conclusions<\/mark><\/strong><\/h3>\n\n\n\n
                        #1. Existence of semi-separating PBEs<\/mark><\/strong><\/h4>\n\n\n\n
                        \\(c_n\\)\uc758 \uac12\uc774 0\uc77c \ub54c semi-separating PBE\ub97c \ub9cc\uc871\ud558\ub294 \uc9c0\uc810\uc774 \uc874\uc7ac\ud55c\ub2e4.(\uc870\uac74\uc5d0 \uad00\uacc4\uc5c6\uc774 \ubaa8\ub450 semi-separating PBE\ub97c \ub9cc\uc871\ud55c\ub2e4.)
                        \uc774\ub294 \ud544\uc694\ucda9\ubd84\uc870\uac74\uc744 \ub9cc\uc871\ud558\uba70, \\(c_n\\)\uc774 0\uc774 \uc544\ub2cc \uacbd\uc6b0 \ud574\ub2f9 PBE\ub294 \uc874\uc7ac\ud558\uc9c0 \uc54a\ub294\ub2e4. (\uc704\uc758 \ubd84\uc11d(4. Analysis)\uc5d0\uc11c\ub294 \uadc0\ub958\ubc95\uc73c\ub85c \uc99d\uba85\ud588\ub2e4.)<\/p>\n\n\n\n
                        \ub530\ub77c\uc11c \uc774 \ub17c\ubb38\uc758 \ubaa8\ub378\uc5d0\uc11c \uc124\uc815\ud55c \ud30c\ub77c\ubbf8\ud130\uac00 \uc77c\ubc18\uc801\uc778(0\uc774 \uc544\ub2cc) \uac12\uc744 \uac16\ub294 \uacbd\uc6b0, semi-separating PBE\ub294 \uc874\uc7ac\ud558\uc9c0 \uc54a\ub294\ub2e4.<\/p>\n\n\n\n
                        #2. PBE that satisfy the Cho-Kreps criterion<\/mark><\/strong><\/h4>\n\n\n\n
                        Full-Featured(c) \ud5c8\ub2c8\ud31f \uac8c\uc784\uc5d0\uc11c \ub2e4\uc74c\uc758 \ud480\ub9c1 \uc804\ub7b5 ((n, n), (A, A))\uc740 Cho and Kreps\uc758 \uc9c1\uad00\uc801\uc778 \uae30\uc900\uc5d0 \uc758\ud574 \uc9c0\uc6cc\uc9c0\uc9c0 \uc54a\ub294 PBE\uc774\ub2e4.
                        \uc774\ub294 \\(P_h\\) < 0.5\uc778 \uc0c1\ud669\uc5d0\uc11c, \ud5c8\ub2c8\ud31f \uc720\ud615 \ubc0f \uc815\uc0c1 \uc720\ud615\uc774 \uc2e0\ud638 n\uc744 \ubcf4\ub0bc \ub54c \uc720\uc9c0\ub418\ub294 \ud3c9\ud615\uc774\ub2e4.<\/p>\n\n\n\n
                        6. References<\/mark><\/strong><\/h3>\n\n\n\n
                          \n
                        • Suhyeon Lee, Kwangsoo Cho, and Seungjoo Kim, “Do You Really Need to Disguise Normal Servers as Honeypots?”, In Proceedings of the 40th International Conference on IEEE Military Communications Conference (MILCOM 2022), pp. 166-172, Rockville, MD, US, November 2022.<\/li>\n<\/ul>\n\n\n\n
                            \n
                          • CHO, I. AND KREPS, D. (1987), “Signaling games and stable equilibria”, Quarterly Journal of Economics, 102, pp. 179-221.<\/li>\n<\/ul>\n\n\n\n
                              \n
                            • Felix Munoz-Garcia, Ana Espinola-Arredondo, “The Intuitive and Divinity Criterion: Interpretation and Step-by-Step Examples,” Journal of Industrial Organization Education,<\/em> Volume 5, Issue 1, Pages 1\u201320, ISSN (Online) 1935-5041, DOI: 10.2202\/1935-5041.1024, March 2011.<\/li>\n<\/ul>\n\n\n\n
                                \n
                              • Muhamet Yildiz, “Signaling”, mit.edu, 14.12 Game Theory<\/li>\n<\/ul>\n\n\n\n
                                  \n
                                • Cheol Park, “Plausibility of the Intuitive Criterion of Cho and Kreps”<\/li>\n<\/ul>\n\n\n\n
                                    \n
                                  • Felix Munoz-Garcia, Daniel Toro-Gonzalez, Strategy and Game Theory – Practice Exercises with Answers, Springer Texts in Business and Economics, p.305-309<\/li>\n<\/ul>\n\n\n\n
                                    <\/p>","protected":false},"excerpt":{"rendered":"
                                    Reviewing “Do You Really Need to Disguise Normal Servers as Honeypots?” papers and looking for improvements<\/p>","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[108],"tags":[104,4,109,110,90,99],"class_list":["post-1575","post","type-post","status-publish","format-standard","hentry","category-paper-review","tag-bayesian-beliefs","tag-game-theory","tag-honeypots","tag-jul-8-2023","tag-semi-separating","tag-signaling-games"],"_links":{"self":[{"href":"https:\/\/saraheee.com\/ko\/wp-json\/wp\/v2\/posts\/1575"}],"collection":[{"href":"https:\/\/saraheee.com\/ko\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/saraheee.com\/ko\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/saraheee.com\/ko\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/saraheee.com\/ko\/wp-json\/wp\/v2\/comments?post=1575"}],"version-history":[{"count":154,"href":"https:\/\/saraheee.com\/ko\/wp-json\/wp\/v2\/posts\/1575\/revisions"}],"predecessor-version":[{"id":2552,"href":"https:\/\/saraheee.com\/ko\/wp-json\/wp\/v2\/posts\/1575\/revisions\/2552"}],"wp:attachment":[{"href":"https:\/\/saraheee.com\/ko\/wp-json\/wp\/v2\/media?parent=1575"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/saraheee.com\/ko\/wp-json\/wp\/v2\/categories?post=1575"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/saraheee.com\/ko\/wp-json\/wp\/v2\/tags?post=1575"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}