&lt;title&gt;Attentive Associative Architectures And Their Implications To Optical Computing&lt;/title&gt;

Athale, Ravindra A.; Friedlander, Carl; Kushner, Brian G

doi:10.1117/12.963497

Cited by 12 publications

(3 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…2, this corresponds to the placement of an aperture followed by a device (spatial light modulator) to perform the squaring operation in the correlation plane. Equation (12) can be expanded to yield f(k) = Rm0(0)fmo(k) + a(k) = N2fmo(k) + a(k) ; a(k) _ X Rmo(j)fmo(k -j) + X L (13) Rm(j)fm(k -j) . j #0 m #mo j = -L which demonstrates the trade -off that exists between the storage capacity and the degree of shift invariance.…”

Section: Distributed Nonlinearitiesmentioning

confidence: 99%

Shift-Invariant Optical Associative Memories

Hong

1987

Opt. Eng

View full text Add to dashboard Cite

Section: Distributed Nonlinearitiesmentioning

confidence: 99%

Shift-Invariant Optical Associative Memories

Hong

1987

Opt. Eng

View full text Add to dashboard Cite

“…2. These vectors are chosen [23] such that vector i has 8 bits of mismatch (Hamming distance) with any other vector j (also, note that the inner product of any two of the above vectors is equal to 4 and the inner product of any vector with itself is equal to 8).…”

Section: Performance Of Hk-recorded Anmsmentioning

confidence: 99%

New Recording Algorithm For Hopfield Model Associative Memories

Hassoun

Youssef

1988

SPIE Proceedings

View full text Add to dashboard Cite

A new recording technique for Hopfield -type associative /content addressable memories is proposed. The new technique is based on the finite and exponentially convergent algorithm of Ho and Kashyap[1] for the solution of a system of linear inequalities. Associative neural memories recorded with the proposed algorithm are shown to be superior to those recorded with the Hopfield's outer product and Kohonen's generalized inverse techniques. High capacity, high convergence rates to stored memories, and low convergence rates to false and oscillatory states are characteristics of this new recording algorithm. The issue of stable false and oscillatory states is raised, and it is shown that such states have a direct Boolean logic relationship with the stored memories.

show abstract

“…The feedback can be done electronically or optically. Another group, led by Athale, has developed a similar approach based on electro-optic components [23]. Psaltis and his group have reported several other experiences.…”

Section: Optical Neurocomputersmentioning

confidence: 99%