| Srinivas M. Aji and Robert J. McEliece. The generalized distributive law. IEEE Transactions on Information Theory, 46(2):325--343, 2000. |
....to exactly compute p (x s y) in ) operations. For example, when the potentials are specified by a state space model, a standard Kalman filter may be combined with a complementary reverse time recursion [6] These algorithms may be directly generalized to any graph which contains no cycles [8, 14, 19, 20]. We refer to such graphs as tree structured. Tree based inference algorithms use a series of local message passing operations to exchange statistical information between neighboring nodes. One of the most popular such methods is known as the sum product [19] or belief propagation (BP) 14] ....
S. M. Aji and R. J. McEliece. The generalized distributive law. IEEE Transactions on Information Theory, 46(2):325--343, March 2000.
....an undirected graph with pairwise cliques [24] For particular graphs with particular settings of the potentials, Equs. 4 5 yield other well known Bayesian inference algorithms, such as the forward backward algorithm in Hidden Markov Models, the Kalman Filter and even the Fast Fourier Transform [1, 13]. A related algorithm, max product , changes the integration in equation 4 to a maximization. This message passing is equivalent to Pearl s belief revision algorithm in directed graphs. For particular graphs with particular settings of the potentials, the max product algorithm is equivalent to ....
S. M. Aji and R.J. McEliece. The generalized distributive law. IEEE Transactions on Information Theory, 1999. to appear.
.... has been shown that there are on the average asymptotically 1:05 2 0:2874s many stable states in a binary Hop eld net of size s whose feedbacks and biases are zero (w jj = w j0 = 0 for j 2 V ) and whose other weights are independent identically distributed zero mean Gaussian random variables [84], 131] For a particular binary symmetric network, however, the issue of deciding whether there are e.g. at least one (when negative feedback weights are allowed) 24] two [66] or three [24] stable states, is NP complete. Indeed, the problem of determining the exact number of stable states for ....
R.J. McEliece, E.C. Posner, E.R. Rodemich, and S.S. Venkatesh, The capacity of the Hopeld associative memory, IEEE Transactions on Information Theory 33 (4) (1987) 461-482.
....memory, is important. It has been shown that in average there are asymptotically 1:05 2 0:2874s many stable states in binary Hop eld nets of size s, with zero feedbacks and biases (w jj = w j0 = 0 for j 2 V ) whose other weights are independent identically zero mean Gaussian random variables [82, 126]. For a particular binary symmetric network, however, the issue of deciding whether there are e.g. at least one (when negative feedback weights are allowed) 23] two [64] or three [23] stable states, is NP complete. Indeed, the problem of determining the exact number of stable states for a given ....
R.J. McEliece, E.C. Posner, E.R. Rodemich, and S.S. Venkatesh, The capacity of the Hopeld associative memory, IEEE Transactions on Information Theory 33 (4) (1987) 461-482.
....an undirected graph with pairwise cliques [24] For particular graphs with particular settings of the potentials, Equs. 4 5 yield other well known Bayesian inference algorithms, such as the forward backward algorithm in Hidden Markov Models, the Kalman Filter and even the Fast Fourier Transform [1, 13]. A related algorithm, max product , changes the integration in equation 4 to a maximization. This message passing is equivalent to Pearl s belief revision algorithm in directed graphs. For particular graphs with particular settings of the potentials, the max product algorithm is equivalent to ....
S. M. Aji and R.J. McEliece. The generalized distributive law. IEEE Transactions on Information Theory, 1999. to appear.
No context found.
Srinivas M. Aji and Robert J. McEliece. The generalized distributive law. IEEE Transactions on Information Theory, 46(2):325--343, 2000.
No context found.
Aji, S. M., and McEliece, R. J. The generalized distributive law. IEEE Transactions on Information Theory 46 (March 2000), 325--343.
No context found.
S. M. Aji and R. J. McEliece. The generalized distributive law. IEEE Transactions on Information Theory, 46(2):325--343, 2000.
No context found.
References #1# S. M. Aji and R.J. McEliece. The generalized distributivelaw. IEEE Transactions on Information Theory, 2000. in press. #2# S.M. Aji, G.B. Horn, and R.J. McEliece. On the convergence of iterative decoding on graphs with a single cycle. In Proc. 1998 ISIT, 1998.
No context found.
Srinivas M. Aji and Robert J. McEliece. The generalized distributive law. IEEE Transactions on Information Theory, 46(2):325--343, 2000.
No context found.
Aji, S. M., and McEliece, R. J. The generalized distributive law. IEEE Transactions on Information Theory 46 (March 2000), 325--343.
No context found.
Aji, S. M., and McEliece, R. J. The generalized distributive law. IEEE Transactions on Information Theory 46 (March 2000), 325--343.
No context found.
S. M. Aji and R. J. McEliece, \The generalized distributive law," IEEE Transactions on Information Theory, vol. 46, no. 2, pp. 325-349, March 2000.
No context found.
S. M. Aji, J. McEliece, The generalized distributive law, IEEE Transactions on Information Theory 46 (1999) 325-343.
Online articles have much greater impact More about CiteSeer.IST Add search form to your site Submit documents Feedback
CiteSeer.IST - Copyright Penn State and NEC