JPH0357554B2 - - Google Patents

Description

[Detailed description of the invention] [Industrial application field] The present invention relates to an associative memory device.

[Conventional technology]

Generally, in ordinary memories used in electronic computers, a method is adopted in which information stored therein is accessed by specifying an address.

[Problem that the invention seeks to solve]

With such conventional storage devices, there is a problem that if the contents of the address are destroyed, the contents of the stored data will become completely unknown.
Also, regardless of the stored contents, it can only be accessed by address.

The present invention is intended to solve the above problems,
To provide an associative memory device which does not become impossible to read even if a part of the storage contents is destroyed and allows access related to the storage contents.

[Means for solving problems]

To this end, the associative memory device of the present invention has a multiplexing system that multiplexes input or feedback signals and outputs the multiplexed data, an enlargement system that enlarges and outputs the input or feedback signals, and output and enlargement of the multiplexing system. A first arithmetic unit that multiplies and outputs the output of the system, a memory that stores two-dimensional information, performs addition and subtraction to the contents as necessary, and outputs the stored contents, and a memory that outputs the output of the multiplexing system and the memory. A second arithmetic unit that multiplies the output, a demultiplexing system that demultiplexes and outputs the output of the second multiplier, and stores the output from the demultiplexing system while performing a threshold value operation, and stores the output as necessary. The present invention is characterized by having a memory/recall unit consisting of a threshold value memory that can be read out in accordance with the requirements, and a feedback system that returns output results to input.

[Effect]

The associative memory device of the present invention obtains a multiplexed image and an enlarged image of input, processes them, multiplies them, and memorizes them, and also demultiplexes the multiplication result of the stored contents and reference data. This makes it possible to realize an associative memory device by recalling the information.

〔Example〕

Examples will be described below based on the drawings.

First, the principle of the associative memory device according to the present invention will be explained.

As a method for realizing associative memory, the present invention uses an autocorrelation matrix, and the memory matrix is formed by the autocorrelation of the contents to be stored.

The arithmetic processing at the time of memorization is expressed as follows.

M=Σχ·χ' (1) Here, χ is an input vector representing the content to be stored, χ is a transposed vector of χ, and M is a storage matrix. That is, the autocorrelation of the content to be stored is added many times to form the autocorrelation.

At the time of recollection, by operation with this memory matrix,
You can recall the whole from a part. The calculation process during recall can be expressed as the following formula.

y=φ(M・χ) ...(2) Here, y is the output vector, χ is the input data,
φ represents threshold operation. By operating M・χ, χ
Even if the data is incomplete due to some missing parts, as long as the memory matrix M is created by the calculation process in equation (1), the recalled data y will be the original data with the missing parts filled in. Data close to χ can be obtained. Note that the noise portion is removed by detecting data above a predetermined level by operating the threshold value of φ.

Next, as a method when recollection using the memory matrix in equation (1) above does not provide sufficient separation, a sequential calculation method for forming a memory matrix with a higher degree of separation will be described.

M _o+1 = M _o + α [χ − φ {Σ (M _o + χ)}] χ′
...(3) Here, α is the learning gain, and Σ is an operator for matching the order with χ′ by taking the partial sum of M _o ·χ. n+1th memory matrix M _o+1
is the result recalled by M _o for the nth time φ
It is obtained by correcting M _o by multiplying the correlation between the error component at the time of recall, which is represented by the difference between {Σ(M _o · χ)} and χ, and the learning gain α. Note that the learning gain α is selected to a value such that M _o converges.

In this way, by performing the calculation process of equation (3) until M _o converges, a correlation matrix M with improved separability can be obtained,
A memory matrix is formed by autocorrelation of the contents to be memorized, and at the time of association, the whole can be recalled from a part by correlating this memory matrix with the memory matrix.

FIG. 6 is a diagram for explaining the configuration and operation of a spatial light modulation tube, which is a basic component in an embodiment of the associative memory device according to the present invention, in which 1 indicates an input image;
2 is a lens, 3 is a photocathode, 4 is a microchannel plate, 5 is a mesh electrode, 6 is a crystal, 6
1 is a charge storage surface, 7 is a half mirror, 8 is monochromatic light, 9 is an analyzer, and 10 is an output image.

In the figure, an input image 1 incident on a photocathode 3 of a spatial light modulation tube via a lens 2 is converted into a photoelectron image. After this photoelectron image is multiplied by the microchannel plate 4, the charge storage surface 61 of the crystal 6
form a charge pattern. The electric field across the crystal 6 changes depending on the charge pattern, and the refractive index of the crystal 6 changes due to the Pockels effect.

Here, when the crystal 6 is irradiated with linearly polarized monochromatic light 8, the polarization state of the reflected light from the charge accumulation surface 61 has changed due to the birefringence of the crystal 6, so if it is passed through the analyzer 9, An output image 10 having a light intensity corresponding to the light intensity of the input image 1 is obtained.

Next, the main functions of such a spatial light modulation tube related to the present invention will be explained.

(a) Memory function The spatial light modulation tube has a memory function that retains the charge distribution on the surface of the electro-optic crystal for a long time. Since the crystal 6 has a very high electrical resistance value, the charge distribution on the crystal surface 61 can be maintained for several days or more.

(b) Subtraction function The spatial light modulation tube can selectively form a positive or negative charge distribution on the surface of the electro-optic crystal. FIG. 7A is a graph showing secondary electron emission characteristics of an electro-optic crystal.

1 incident on the charge accumulation surface 61 as shown in the figure.
The secondary electron energy E is the first crossover point E
If it is smaller than 1 or larger than the second crossover point E2, the number of primary electrons is larger than the number of secondary electrons emitted at the crystal surface (δ<1), so the crystal surface is negatively charged. do. When the energy of the primary electron is between E1 and E2, the number of secondary electrons is greater than the number of primary electrons (δ>1),
The crystal surface is positively charged.

Figure 6 shows whether to write with positive charge or negative charge when accumulating charge in the crystal.
This is done by controlling the voltages of Vc and Vb. Here, the subtraction function can be provided by two methods: first writing a negative voltage and then charging positively, or first charging positively and then writing negative charges. The amount of subtraction is the following 3
It can be controlled in two ways.

That is, there are a method of changing the intensity of the incident light during subtraction, a method of changing the duration of the voltage applied to the microchannel plate 4, and a method of changing the voltage applied to the microchannel plate 4.

The writing and erasing methods are well known, but the seventh
This will be explained using a positive charge image as an example with reference to FIGS. A and B. In addition, the voltage applied to the secondary electron collecting electrode
Vc is set at the second crossover _E2 in FIG. 7A.

[Erase operation]

The crystal back voltage Vb is changed to the second crossover point E ₂
When set to a potential corresponding to , the crystal surface potential
Vs is a value obtained by adding the potential increase of positive charges caused by writing to the potential _E2 . For this surface potential, the energy of the incident primary electron is
Since E is more than ₂ , the secondary electron emission ratio δ<1,
Negative charges are accumulated until the surface potential reaches E ₂ . When the potential reaches E ₂ , δ=1, an equilibrium state is reached, and the surface charge becomes zero.

[Write operation]

The crystal back voltage Vb is changed to the first crossover point E ₁
and the second crossover point _E2 , the voltage E' is set to provide a sufficient dynamic range.
At this time, the crystal surface potential Vs also becomes approximately E′, so
The energy of the incident primary electron is between E ₁ and E ₂ . Therefore, since the secondary electron emission ratio δ>1, a positive charge image is formed. The emitted secondary electron is
The secondary electrons are collected by the secondary electron collecting electrode, which is at a potential Vc higher than Vs.

(C) Real-time threshold operation function The spatial light modulation tube can perform real-time threshold operation according to the setting conditions of Vc and Vb shown in FIG. The mesh electrode 5 is provided near the crystal surface, and when it is set to a predetermined potential, if sufficient electrons are supplied to the crystal surface, the crystal surface potential becomes the potential of the mesh electrode, and this potential This is the crossover point.
In other words, if the crystal surface potential is lower than the mesh electrode, the secondary electrons emitted from the crystal surface will be collected by the mesh electrode, so the secondary electron emission will increase relative to the incident electrons, and the crystal surface potential will rise. Conversely, if the crystal surface potential is higher than the mesh electrode potential, the number of incident electrons will be greater than the secondary electrons, causing the crystal surface potential to drop, and eventually become constant when the crystal surface potential becomes equal to the mesh electrode potential. Now, the voltage Vc of the mesh electrode 5
When Vb is lowered to about 0.1KV and Vb is lowered stepwise below 0.1KV, the potential of the crystal surface also decreases stepwise, and the charge accumulation surface 61 of the crystal 6 becomes a negative potential, and electrons no longer reach it. This results in a so-called lockout state. However, when Vb is slowly lowered in a ramp-like manner, electrons are supplied to the crystal surface in areas where the intensity of the incident light is high and a large amount of electrons are supplied, and secondary electron emission increases, causing negative potential. However, in areas where the intensity of the incident light is low and the amount of electrons supplied is small, the supply of electrons cannot keep up with the potential drop.
Therefore, the potential becomes negative and electrons no longer reach the crystal surface. Therefore, in response to the intensity of light incident on the photocathode 3, there are areas where the crystal surface becomes negative potential and writing is not performed, and areas where electrons reach the crystal surface and writing occurs without the surface becoming negative potential. As a result, a threshold operation is performed depending on the intensity of the incident light. (d) AND operation function based on real-time threshold operation As described above, the spatial light modulation tube has a real-time threshold operation function. Using this, do the following
It is possible to perform an AND operation. That is, as shown in FIG. 8, when two inputs are superimposed, the input light intensity becomes three levels of 0, 1, and 2. If real-time threshold operation is performed at level 2 of these, the result will be as shown in Figure 9.
This means that an AND operation has been executed.

Next, other components necessary for the present invention will be explained. Note that the input image is assumed to be a 2×2 pattern.

Enlargement system As shown in Figure 10, the 2x2 input image is
4 to enlarge and project. In reality, this is done using a lens or the like.

Multiplexing System A 2×2 input image is repeatedly projected as shown in FIG. 11A to obtain a pattern as shown in FIG. 11B. Actually, this is realized by a lens array or the like.

Demultiplexing system As shown in Figure 12, the 4x4 input image is
Project it so that it overlaps with 2. Actually, this is realized by a lens array or the like.

Next, a coding method that performs coding suitable for two-dimensional data using the aforementioned enlarged image, multiplexed image, and demultiplexed image will be explained using a 2×2 image.

The method of memorization and recollection that was separately filed by the present applicant will be explained.

The vector χ and its transposed vector χ′ are expressed as follows.

χ＝x ₁ x ₂ x ₃ x ₄ , χ′=(x ₁ , x ₂ , x ₃ , x ₄ ) In this case, the memorization is M=χ・χ′=x ₁ x ₂ x ₃ x ₄・( x ₁ , x ₂ , x ₃ , x ₄ ) = x ₁ x ₁ x ₁ x ₂ x ₁ x ₃ x 1 x ₄ x ₂ x ₁ x ₂ x ₂ x ₂ x ₃ x ₂ x ₄ x ₃ x _{1 x 3} x ₂ x ₃ x ₃ x 3 x _{4 x 4} x ₁ x ₄ x ₂ x ₄ x ₃ x ₄ x ₄ , while recollection is expressed as z′=M×y=x ₁ x ₁ x ₁ x ₂ x ₁ x ₃ x ₁ x ₄ x ₂ x ₁ x ₂ x ₂ x ₂ x _{3 x 2} x ₄ x ₃ x ₁ x ₃ x ₂ x ₃ x ₃ x ₃ x ₄ x ₄ x ₁ x ₄ x _{2 x 4} x ₃ x ₄ x ₄ y ₁ y ₂ y ₃ y ₄ = x ₁ x ₁ y ₁ x 1 x ₂ y ₂ x 1 x _{3 y 3 x 1} x ₄ y _{4 x 2 x 1} y ₁ x ₂ x ₂ y ₂ x ₂ x ₃ y ₃ x ₂ x ₄ y ₄ x ₃ x ₁ y ₁ x ₃ x ₂ y ₂ x 3 x ₃ y ₃ x ₃ x ₄ y _{4 x 4 x 1} y ₁ x ₄ x ₂ y ₂ x It is expressed as ₄ x ₃ y ₃ x ₄ x ₄ y ₄ .

Here, if we make z' one-dimensional and call it z, then z=x ₁ x ₁ y ₁ +x ₁ x ₂ y ₂ +x ₁ x ₃ y ₃ +x ₁ x ₄ y ₄ x ₂ x ₁ y ₁ +x ₂ x ₂ y ₂ +x ₂ x ₃ y ₃ +x ₂ x ₄ y ₄ x ₃ x ₁ y ₁ +x ₃ x ₂ y ₂ +x ₃ x ₃ y ₃ +x ₃ x ₄ y ₄ x ₄ x ₁ y ₁ +x ₄ x ₂ y ₂ +x ₄ x ₃ y ₃ + x ₄ x ₄ y ₄ .

However, in order to perform image processing with this method, it is necessary to convert the two-dimensional data of the image into one-dimensional data.

In contrast, in the present invention, memorization and recall are performed by the following calculations. Now let χ be the content you want to remember, and χ = x ₁ x ₂ x ₃ x ₄ .

Here, due to the expansion system, x mag=x ₁ x ₁ x ₂ x ₂ x 1 x ₁ x _{2 x 2 x 3} x ₃ x ₄ x ₄ x ₃ x ₃ x ₄ x ₄ Due to the multiplex system, x mlt We get =x ₁ x ₂ x ₁ x ₂ x ₃ x ₄ x ₃ x ₄ x ₁ x ₂ x ₁ x ₂ x ₃ x ₄ x ₃ x ₄ . At this time, if the multiplication of the elements of x mag and x mlt, x mag・x mlt, is M, then M=x mag・x mlt=x ₁ x ₁ x ₁ x ₂ x ₂ x ₁ x
₂ x ₂ x ₁ x ₃ x ₁ x ₄ x ₂ x ₃ x ₂ x ₄ x ₃ x ₁ x ₃ x ₂ x ₄ x _{1 x 4 x 2} x ₃ x ₃ x 3 x ₄ x ₄ x ₃ x _{4 x 4} (... multiplication between elements is called Hadamard product), and the same result as that obtained by the matrix operation described above is obtained (however, the array is different),
It can be seen that it is possible to memorize by x mag x mlt.

Also, let y mlt=y ₁ y ₂ y ₁ y ₂ y ₃ y ₄ y _{3 y 4 y 1} y ₂ y ₁ y ₂ y ₃ y ₄ y ₃ y ₄ , and multiply the elements of M and y mlt by M・y
If mlt is z', z'=M・y mlt = x ₁ x ₁ x ₁ x ₂ x ₂ x 1 x ₂ x ₂ x ₁ x ₃ x ₁ x ₄ x ₂ x ₃ x ₂ x ₄ x _{3 x 1} x ₃ x ₂ x ₄ x ₁ x ₄ x ₂ x ₃ x _{3 x 3 x 4} x ₄ x ₃ x ₄ x ₄ y ₁ y ₂ y ₁ y ₂ y ₃ y ₄ y ₃ y ₄ y ₁ y ₂ y ₁ y ₂ y ₃ y ₄ y ₃ y ₄ =x ₁ x ₁ y ₁ x 1 x ₂ y ₂ x ₂ x ₁ y ₁ x ₂ x ₂ y ₂ x ₁ x ₃ y _{3 x 1 x 4} y ₄ x ₂ x ₃ y ₃ x ₂ x ₄ y ₄ x ₃ x ₁ y ₁ x ₃ x ₂ y ₂ x ₄ x ₁ y ₁ x ₄ x 2 y ₂ x ₃ x ₃ y ₃ x ₃ x ₄ y ₄ x _{4 x 3} y ₃ x ₄ x ₄ y ₄ . Here, let z be the demultiplexing of z', then z=x ₁ x ₁ y ₁ +x ₁ x ₂ y ₂ +x ₁ x ₃ y ₃ +x ₁ x ₄ y ₄ x ₂ x ₁ y ₁
+x ₂ x ₂ y ₂ +x ₂ x ₃ y ₃ +x ₂ x ₄ y ₄ z=x ₁ x ₁ y ₁ +x ₁ x ₂ y ₂ +x ₁ x ₃ y ₃ +x ₁ x ₄ y ₄ x ₂ x ₁ y ₁
+x ₂ x ₂ y ₂ +x ₂ x ₃ y ₃ +x ₂ x ₄ y ₄ x ₃ x ₁ y ₁ +x ₃ x ₂ y ₂ +x ₃ x ₃ y ₃ +x ₃ x ₄ y ₄ x ₄ x ₁ y ₁ +x ₄ x ₂ y ₂ + x ₄ x
₃ y ₃ + x ₄ x ₄ y ₄ , which is the same result as the above matrix operation (however, the array is different), and M・y
It can be seen that a recalled pattern can be obtained by demultiplexing the mlt image.

In this way, memorization can be performed using the multiplexing system and expansion system, and recall can be performed using the demultiplexing system, so it is possible to realize associative memory entirely through two-dimensional processing. Become.

Next, the associative memory according to the present invention will be specifically explained. Note that the calculation is performed using the following equation, which is similar to equation (3) above.

M _o+1 = M _o +α[x mag−[φ｛Σ(M _o・x mlt)
}]mag]xmlt...(4) Here, α is a learning gain, xmag is an enlarged image of x, xmlt is a multiplexed image of x, and Σ is a demultiplexed image. The n+1st memory matrix M _o+1 is the result of recalling by the nth M _o [φ{Σ(M _o・
xmlt)}mag and xmag, and the product of _xmalt multiplied by the learning gain α. As before, the learning gain α is M _o
Select a value such that converges, and perform this calculation process as M _o
If this is done until convergence, M _o with improved resolution can be obtained.

FIG. 1 is a block diagram showing the configuration of an associative memory device according to the present invention, in which 101 is an enlargement system;
2 is a multiplexing system, 103 is a first arithmetic unit, 104 is a memory, 105 is a second arithmetic unit, 106 is a demultiplexing system, and 107 is a threshold memory.

In the figure, as mentioned above, the enlarging system 101 converts, for example, a 4×4 input x into a 16×16 x mag
The multiplexing system 102 similarly creates 16×16 x mlt from the 4×4 input x. The first arithmetic unit 103 multiplies the elements of x mag and x mlt. The first memory 104 has x mag x
This is the memory for adding mlt and subtracting feedback values. The second arithmetic unit 105 multiplies the contents of the memory 104 by xmlt to create recall data. The demultiplexing system 106 converts 16×16 data into 4×
This is for converting into data of 4. The threshold value memory 107 is a memory that performs a threshold value operation, detects data above a predetermined level, and cuts out noise. The output y of this threshold memory is fed back to the input side through a feedback system.

Next, the operation during memorization will be explained according to equation (4). The flow of memorization is shown by thin lines in the figure.

First, an enlargement system 101 creates x mag from two-dimensional input data x to be memorized, and at the same time, a multiplex system 102 creates x mlt.

In the second arithmetic unit 105, the
M _o and x mlt from multiplexing system 102,
M _o ·x mlt is calculated, passed through the demultiplexing system 106, subjected to threshold value operation in the threshold value memory 107, and then stored in memory. Here, it is assumed that M _o is initially set to, for example, "0". Of course, “0”
It may be set to a value other than .

Using the created x mag and x mlt as input data, the first arithmetic unit 103 multiplies them x
Execute mag・x mlt and multiply by α to memory 104
Add to M _o stored in .

The result of
(M _o x)}.

The first arithmetic unit 103 calculates φ{Σ(M _o x)} x
mlt is calculated, multiplied by α, and subtracted from M _o in the memory 104.

x mag x mlt is multiplied by α and added to M _o , and then φ{Σ(M _o x)} x mlt
The result of multiplying α by α and subtracting _{it from M o becomes} M _o+1 .

The input data is memorized by performing the above operations until M _o converges. expect
Once M _o is obtained, processing is performed based on that value.

Next, the time of recollection will be explained.

In this case, it is assumed that the memory matrix M has already been calculated and memorized. The flow during recall is shown by the thick line in the figure.

First, xmlt is created by the multiplexing system 102 from two-dimensional input data x to be recalled.

In the second arithmetic unit 105, the memory 104
Multiplication with M read from , M.

The output is read from the threshold memory 107, and if one recollection is insufficient, the output is returned to the input by the feedback system, recollection is performed, and desired data is read.

In this way, memorization and recall can be performed.

Next, an embodiment in which the associative memory device shown in FIG. 1 is constructed using a spatial light modulation tube will be described.

Figure 2A is a diagram showing an example of an optical associative memory device with a complete optical system using four spatial light modulation tubes, Figure 2B is a time chart during memorization, and Figure 2C is a time chart during recall. This is a time chart. In the figure,
201 to 204 are spatial light modulation tubes, 211 is an enlarged imaging system, 212 is a multiplex imaging system, 213 is an inverse multiplex imaging system, 221 to 228 are half mirrors, 231 to 2
38 is a reflecting mirror, S1, S2, S3 are shutters,
Vb ₁ to _{Vb 4} are spatial light modulation tubes 201 to 20, respectively.
4, the crystal backside voltages Vm ₁ to Vm ₄ are the voltages of the microchannel plates (MCPs) of the spatial light modulation tubes 201 to 204. These devices are arranged to function as devices corresponding to FIG. 1, and the computing units 103 and 105 in FIG. The memory 104 is made up of a spatial light modulation tube 203, and the threshold memory 107 is made up of a spatial light modulation tube 204. Note that the calculation is performed based on equation (4) as in FIG. 1, and the learning gain α in the equation is given by appropriately setting the voltage of the MCP. Moreover, the display of ~ in FIGS. 2B and 2C corresponds to the processing process ~ below.

First, the operation during memorization will be explained with reference to FIGS. 2A and 2B.

Input x and half mirror 228, reflector 2
38 to a magnifying imaging system 211 and a multiplexing imaging system 212, respectively x mag and x
mlt, and at the same time as the input is turned on, the crystal back voltage Vb ₁ of the spatial light modulation tubes 201 and 202 is set.
Vb ₂ is the write voltage, and microchannel plate (MCP) voltages Vm ₁ and Vm ₂ are set for the specified time.
Turn on and write x mag and x mlt to the spatial light modulation tubes 201 and 202, respectively.

Next, open the shutter S1 and open the half mirror 22.
1,222, reflecting mirror 231, half mirror 22
The spatial light modulation tube 202 is continued to be read out by monochromatic light through the mirrors 3 and 224, and as a result, the spatial light modulation tube 203 constituting the memory is read out through the half mirrors 224 and 225. M _o ·x mlt is calculated by successively reading out the continuation result x mag of the spatial light modulation tube 202 and the memory content Mn of the spatial light modulation tube 203 . The calculation results are demultiplexed by the demultiplexing imaging system 213 via the half mirror 225 and converted into Σ(M _o x mlt)
is created and written to the spatial light modulation tube 204 while operating the threshold value by real-time hard clip operation. This threshold processing gradually lowers the crystal back voltage Vb ₄ of the spatial light modulation tube 204, and
During this time, the MCP voltage Vm ₄ is turned ON, and the threshold-processed digital quantity φ{Σ(M _o ·x mlt)} is stored in the spatial light modulation tube.

Only the shutter S2 is opened, and in the same way, the half mirrors 221, 222 and the reflecting mirror 23 are exposed to monochromatic light.
2. Read out the spatial light modulation tube 201 through the half mirror 223, and use the result to read out the half mirror 2.
By successively reading out the spatial light modulation tube 202 via 24, x mag x mlt is calculated. At the same time as the shutter S2 is turned on, the crystal back voltage Vb ₃ of the spatial light modulation tube 203 is switched to addition mode, the MCP voltage Vm ₃ is turned on, and the read x mag x mlt is transferred to the half mirror 22.
5. Add to the spatial light modulation tube 203 via reflecting mirrors 233, 234, and 235.

The crystal back voltage Vb ₁ of the spatial light modulation tube 201 is switched to erase, and the MCP voltage Vm ₁ is switched to erase.
After turning ON and erasing the spatial light modulation tube 201, only the shutter S3 is opened, and monochromatic light is transmitted from the spatial light modulation tube 204 via the half mirror 221, the reflecting mirror 236, and the half mirror ₂₂₆ .
mlt)}, half mirror 227,
Reflector 237, half mirror 228, reflector 2
38, an enlarged image is formed in the enlarged imaging system 211, and the image is sent to the spatial light modulation tube 201 [φ{Σ(M _o x
mlt)}] Write mag.

Open only the shutter S2 and open the spatial light modulation tube 20.
1 and half mirror 22 with the result.
3,224 to read out the spatial light modulation tube 202 and obtain [φ{Σ(M _o x mlt)}] mag x
Calculate mlt. At the same time as the shutter S2 is turned on, the crystal back voltage Vb ₃ of the spatial light modulation tube 203 is switched to subtraction, and the MCP voltage Vm ₃ is turned on for a predetermined period of time.
(M _o x mlt)}] Subtract mag x mlt.
At this time, the result does not become smaller than 0 due to the effect of the mesh electrode of the spatial light modulation tube 203 (see FIG. 6).

The crystal back voltages Vb ₁ , Vb ₂ , Vb ₄ of the spatial light modulation tubes 201 , 202 and 204 are switched to erase, and the MCP voltages Vm ₁ , Vm ₂ , Vm ₄ are switched to erase.
Turn on to erase each.

Memorization is completed by performing the above operations until convergence for each input pattern.

Next, the recollection operation will be explained with reference to FIGS. 2A and 2C.

Turn on the input, and at the same time switch the crystal back voltage Vb ₂ of the spatial light modulation tube 202 to write mode, turn on the MCP voltage Vm ₂ , and send x to the spatial light modulation tube 202 via the half mirror 228 and the multiplex imaging system 212. Write mlt.

Only the shutter S1 is opened and the spatial light modulation tube 202 is read out using monochromatic light through the half mirrors 221, 222, the reflecting mirror 231, and the half mirrors 223, 224.
24, 225 to read out the spatial light modulation tube 203 and calculate M _o ·x mlt. The calculation results are demultiplexed by the demultiplexing imaging system 213 and Σ
Create (M _o x mlt). Spatial light modulation tube 2
While slowly lowering the crystal back voltage Vb ₂ of 04, the MCP voltage Vm ₄ is turned on, and φ{Σ(M _o ·x mlt)} is written to the spatial light modulation tube 204 while performing a real-time hard clip operation.

Half mirror 221, reflecting mirror 2 due to monochromatic light
36, the spatial light modulation tube 204 is read out through the half mirror 226, and the output φ{Σ(M _o ·x mlt)} is obtained from the half mirror 227.

The crystal back voltage Vb ₂ of the spatial light modulation tube 202 is switched to erase, and the MCP voltage Vm ₂ is switched to erase.
After turning on and erasing the spatial light modulation tube 202, only the shutter S3 is turned on. Half mirror 221, reflecting mirror 236, half mirror 22 by monochromatic light
6, 227, the content recalled from the spatial light modulation tube 204 by the feedback loop consisting of the reflecting mirror 27 is transferred to the half mirror 228 and the multiplex imaging system 2.
Spatial light modulation tube 20 as x mlt through 12
Write to 2.

It is possible to switch the crystal back voltage Vb ₄ of the spatial light modulation tube 204 to erase, turn on the MCP voltage Vm ₄ to erase, and return to the process.

FIG. 3A is a diagram showing an embodiment of a hybrid optical content addressable memory device in which a part of the optical content addressable memory device using the complete optical system of FIG. 2 is replaced with an electronic circuit, and FIG. Time chart, 3rd
Figure C is a time chart during recall. In the diagram, 3
01 to 303 are spatial light modulation tubes, 311 is an enlarged imaging system, 312 is a multiplex imaging system, 313 is an inverse multiplex imaging system, 321 to 324 are half mirrors, 331 to 3
35 is a reflecting mirror, 341 is a light receiving matrix, 35
1 is a CPU, 361 and 362 are pattern display devices,
S1 and S2 are shutters.

In this embodiment, interface with electronic circuits is facilitated by replacing the input section and feedback section with a CPU. Additionally, the number of spatial light modulation tubes used is reduced to three, resulting in a system that can be implemented more easily.

The memorization operation will be explained with reference to FIGS. 3A and 3B.

Turn on the display devices 361 and 362, and turn on the CPU 35.
1, the input x is the pattern display device 361, 36
2 and displayed. The displayed pattern is magnified by a magnifying imaging system 311 and multiplexed by a multiplexing imaging system 312 to generate x mag and x mlt, respectively. At the same time, the crystal back voltage Vb ₁ of the spatial light modulation tubes 301 and 302,
At the same time as switching Vb ₂ to writing, MCP voltages Vm ₁ and Vm ₂ are turned on for a predetermined period of time to send x mag and x mlt to the spatial light modulation tubes 301 and 302, respectively.
Write.

Shutter S1 only and light receiving matrix 34
Turn on 1. Then, by monochromatic light, a half mirror 321, a reflecting mirror 331, a half mirror 32
The spatial light modulation tube 302 is read out through 2, 323, and the half mirrors 323, 324 are read out with the result.
Continue reading out the spatial light modulation tube 303 through
Calculate Mn x mlt. The calculation results are demultiplexed by the demultiplexing imaging system 313 to obtain Σ(Mn・x
mlt) and convert it into a light receiving matrix 341
converts it into an electrical signal and supplies it to the CPU 351.
In the CPU 351, [φ{Σ
(Mn x mlt)] is obtained.

At the same time only the shutter S2 is opened, the crystal back voltage Vb ₃ of the spatial light modulation tube 303 is switched to addition, and the MCP voltage Vm ₃ is turned ON.
As a result, the monochromatic light causes the half mirror 321,
The contents of the spatial light modulation tube 301 are read out through the reflecting mirror 332 and the half mirror 322, and as a result, the contents of the spatial light modulation tube 302 are read out through the half mirrors 322 and 323, and x mag x
mlt is calculated and the calculation result is half mirror 32
3, 324, and is added to the spatial light modulation tube 303 via reflecting mirrors 333, 334, and 335.

The crystal back voltage Vb ₁ of the spatial light modulation tube 301 is switched to erase, and the MCP voltage Vm ₁ is switched to erase.
After turning ON and erasing the spatial light modulation tube 301, the CPU
[φ{Σ(Mn·x mlt)}] from 351 is supplied to a display device 361 for display, and is enlarged and imaged by an enlarged imaging system 311. At the same time, by switching the crystal back voltage of the spatial light modulation tube 301 to write Vb ₁ and turning on the MCP voltage Vm ₁ , enlarged imaging [φ{Σ(Mn x
mlt)}]mag is written into the spatial light modulation tube 301.

At the same time only the shutter S2 is opened, the crystal back voltage Vb ₃ of the spatial light modulation tube 303 is switched to the subtraction side, and the MCP voltage Vm ₃ is turned on for a predetermined period of time.
do. The monochromatic light reads out the contents of the spatial light modulation tube 301 through the half mirror 321, the reflecting mirror 332, and the half mirror 322, and as a result, the contents of the spatial light modulation tube 302 are read out through the half mirrors 322 and 323, and [φ {Σ(Mn・x
mlt)}]mag・xmlt is calculated, and the calculation results are shown in the half mirrors 323, 324 and the reflecting mirror 33.
Spatial light modulation tube 30 via 3,334,335
Subtracted from the contents of 3. In this case, the calculation result will not become lower than "0" due to the mesh effect.

While switching the crystal back voltages Vb ₁ and Vb ₂ of the spatial light modulation tubes 301 and 302 to writing,
They are erased by turning on the MCP voltages Vm ₁ and Vm ₂ for a predetermined period of time.

Thereafter, the operations from to are repeated until the memorization is completed.

Next, the recall process of the second embodiment will be explained with reference to FIGS. 3A and 3C.

Turn on the display device 362 and input x to the CPU 35
1 is supplied to the display device 362 and displayed,
Multiple images are formed by a multiple imaging system 312. xmlt is written into the spatial light modulation tube 302 by switching the crystal back voltage of the spatial light modulation tube 302 to writing and simultaneously turning on the MCP voltage for a certain period of time.

Shutter S1 only and light receiving matrix 34
1 is turned on, and the half mirror 32 is illuminated by monochromatic light.
1. Read out xmlt from the spatial light modulation tube 302 through the reflecting mirror 331 and half mirror 323, read out xmlt from the spatial light modulation tube 302 through the half mirrors 323 and 324, and read out Mm.
Calculate x mlt. The calculation results are demultiplexed by a demultiplexing imaging system 313 via a half mirror 324 to obtain Σ(Mn x mlt), which is converted into an electrical signal by a light receiving matrix 341.
Supplied to CPU351. [φ{Σ(Mn·x mlt)}] is obtained by threshold processing in the CPU 351, and data recall is executed.

Switch the crystal back voltage Vb ₂ of the spatial light modulation tube 302 to erase, and turn on the MCP voltage Vm ₂ for a predetermined time.
After deleting the data, turn on the display device 362,
[φ{Σ(Mn・x
mlt)}] is displayed and multiplexed by the multiplex imaging system 312. Data is written into the spatial light modulation tube 302 by switching the crystal back voltage Vb ₂ of the spatial light modulation tube 302 to writing and simultaneously turning on the MCP voltage Vm ₂ for a predetermined period of time. It is also possible to return from this process to subsequent processes.

As described above, in the systems in the two embodiments, the spatial light modulation tubes 201, 202, 301, 30
2 is used for multiplication, but in reality it is {0,
1}, so the logical operation AND
It is the same as the operation. Therefore, by using AND operation using real-time threshold operation for this part,
It is possible to reduce the number of spatial light modulation tubes by one.

FIG. 4A is a diagram showing an example of an optical associative memory device using a complete optical system, in which the number of spatial light modulation tubes is reduced by one and is realized by three by introducing a real-time threshold operation AND operation. Figure B is a time chart for memorization, and Figure 4 C is a time chart for recall. In the figure, 401 to 403 are spatial light modulation tubes, 411 is an enlarged imaging system, 412 is a multiplex imaging system,
413 is an inverse multiplex imaging system, 421 to 430 are half mirrors, 431 to 437 are reflecting mirrors, S1, S2,
S3 and S4 are shutters.

In the figure, a spatial light modulation tube 401 performs an AND operation, a spatial light modulation tube 402 functions as a memory, and a spatial light modulation tube 403 functions as a threshold memory.

Memorization will be explained with reference to FIGS. 4A and 4B.

Only the shutter S1 is opened and x is input, and the multiplex imaging system 412 forms multiple images to form the spatial light modulation tube 4.
The crystal back voltage Vb ₁ of 01 is switched to writing, and the MCP voltage Vm ₁ is turned on for a predetermined period of time to write x mlt.

Half mirror 421, reflecting mirror 4 due to monochromatic light
31. Read out the spatial light modulation tube 401 through the half mirror 422, read out the spatial light modulation tube 402 through the half mirrors 422 and 423 with the result, calculate M _o x mlt, and use the inverse multiplex imaging system 413. Perform demultiplexing by Σ(M _o・x
mlt). The crystal back voltage Vb ₃ of the spatial light modulation tube 403 is slowly lowered, while the MCP voltage Vm ₃ is turned on to perform threshold processing using real-time threshold operation.
Write φ{Σ(M _o x mlt)} into 3.

After switching the crystal back voltage Vb ₁ of the spatial light modulation tube 401 to erase and turning on the MCP voltage Vm ₁ for a predetermined period of time to erase, input x, open the shutters S1 and S2, and erase the half mirrors 427 and 42.
9. x mlt by multiple imaging system 412, and enlarged image x by half mirrors 427, 428, 429, reflecting mirror 437, and magnifying imaging system 411
Create mag. At the same time, the crystal back voltage Vb ₁ of the spatial light modulation tube 401 is slowly lowered, and during this time the MCP voltage Vm ₁ is turned on, writing is performed in areas where the intensity of the incident light is strong, and the lockout state is placed in areas where the intensity of the incident light is weak, in real time. By threshold operation, an AND operation of x mag and x mlt is executed and x mag · x mlt is written.

The crystal back voltage Vb ₂ of the spatial light modulation tube 402 is switched to addition, and the MCP voltage is
Turn on Vm ₂ , read out the spatial light modulation tube 401, and add x mag · x mlt to the spatial light modulation tube 402.

After switching the crystal back voltage Vb ₁ of the spatial light modulation tube 401 to erase and turning on the MCP voltage Vm ₁ for a predetermined period of time to erase, x is input, shutters S1 and S4 are opened, and the half mirror 429 and multiple imaging system are switched on. x mlt is created by 412, and φ
{Σ(M _o・x mlt)} is read out, and [φ{Σ(M _o・x mlt)}]
Create mag, and use real-time threshold operation as in the case of [φ{Σ(M _o x mlt)}] mag and x
Execute the AND operation of mlt, [φ{Σ(M _o・x
mlt)} mag・x mlt spatial light modulation tube 401
Write to.

Reflector 431, half mirror 4 due to monochromatic light
22 to read out the spatial light modulation tube 401,
Switch the crystal back voltage Vb ₂ of the spatial light modulation tube 402 to subtraction, and turn on the MCP voltage Vm ₂ .
and subtracted from the spatial light modulation tube 402. At this time, the result does not become smaller than 0 due to the effect of the mesh electrode.

Crystal back voltage of spatial light modulation tubes 401 and 403
At the same time as switching Vb ₁ and Vb ₃ to erase, MCP
The voltages Vm ₁ and Vm ₃ are turned on for a predetermined period of time to erase each one.

Memorization is completed by performing the above operations until convergence for each input pattern.

Next, recall will be explained with reference to FIGS. 4A and 4C.

Open only the shutter S1, input x, switch the crystal back voltage Vb ₁ of the spatial light modulation tube 401 to writing, and turn on the MCP voltage Vm ₁ .
Then, xmlt created by the multiplex imaging system 412 is written into the spatial light modulation tube 401 via half mirrors 427 and 429.

Reflector 431, half mirror 4 due to monochromatic light
22 to read out the spatial light modulation tube 401,
As a result, the spatial light modulation tube 402 is read out via the half mirrors 422 and 423, and M _o x
Calculate mlt. The calculation results are transferred to the inverse multiplex imaging system 41
3 to create Σ(M _o x mlt), slowly lower the crystal back voltage Vb ₃ of the spatial light modulation tube 403, and turn on the MCP voltage Vm ₃ during that time to perform real-time threshold operation. and writes φ{Σ(M _o ·x mlt)} into the spatial light modulation tube 403.

Half mirror 421, reflecting mirror 4 due to monochromatic light
35, the spatial light modulation tube 403 is read out through the half mirror 424, and φ{Σ(M _o ·x mlt)} is obtained as an output from the half mirror 425. It is also possible to re-recall this output by opening the shutters S1 and S3 and returning it through the feedback system.

The crystal back voltage Vb ₁ of the spatial light modulation tube 401 is switched to erase, and the MCP voltage Vm ₁ is switched to erase.
After turning ON and erasing, open the shutters S1 and S3, switch the crystal back voltage Vb ₁ of the spatial light modulation tube 401 to writing, and change the MCP voltage.
Turn on Vm ₁ and write the output of the feedback loop to the spatial light modulation tube 401 through the multiplex imaging system 412.

While lowering the crystal back voltage Vb ₃ of the spatial light modulation tube 403, the MCP voltage Vm ₃ is turned on for a predetermined time.
After deletion, recall is performed by repeating the process from.

FIG. 5A is a diagram showing an embodiment of a hybrid optical content addressable memory device in which a part of the optical content addressable memory device using the complete optical system of FIG. 4 is replaced with an electronic circuit, and FIG. Time chart, 5th
Figure C is a time chart during recall. In the figure, 5
01 and 502 are spatial light modulation tubes, 511 is an enlarged imaging system, 512 is a multiplex imaging system, 513 is an inverse multiplex imaging system, 521 to 523 are half mirrors, 531 to 5
34 is a reflecting mirror, 541 is a light receiving matrix, 55
1 is a CPU, and 561 and 562 are pattern display devices.

In this embodiment, the input part and feedback part of the optical content addressable memory system shown in FIG.
By replacing the spatial light modulation tube with an electronic circuit, the interface with the electronic circuit is facilitated, and by replacing the output section with an electronic circuit, a hybrid optical content addressable memory system with one less spatial light modulation tube is obtained.

The operation during memorization will be explained with reference to FIGS. 5A and 5B.

The input x is supplied from the CPU 551 to the display device 561 for display, and the displayed data is transferred to the multiplex imaging system 5
Convert to xmlt at 12, spatial light modulation tube 501
At the same time, switch the crystal back voltage Vb ₁ to writing and turn on the MCP voltage Vm ₁ for a predetermined period of time.
Write mlt.

The light receiving matrix 541 is turned ON, and the monochromatic light is read out from the spatial light modulation tube 501 through the half mirror 521, and the result is read out from the half mirror 5.
21,522 to read out the spatial light modulation tube 502 and calculate M _o ·x mlt. The calculation results are transferred to the inverse multiplex imaging system 51 via a half mirror 522.
3 to create Σ(M _o ·x mlt), which is converted into an electrical signal by the light reception matrix 541. This electrical signal is taken into the CPU 551, subjected to threshold processing, and converted into data φ{Σ(M _o x
mlt)} is obtained.

At the same time as switching the crystal back voltage Vb ₁ of the spatial light modulation tube 501 to writing, the MCP voltage Vm ₁
After turning on for a predetermined period of time and erasing, the display device 56
1,562 is turned on, magnification imaging system 511 and multiplex imaging system 512 produce x mag and x
Create mlt. Then, while slowly lowering the crystal back voltage Vb ₁ of the spatial light modulation tube 501, the MCP voltage Vm ₁ is turned ON during this time, and the AND operation of x mag and x mlt is performed by real-time threshold operation, and x mag x mlt Write.

The spatial light modulation tube 501 is read out via the half mirror 521 with monochromatic light. At this time, the crystal back voltage Vb ₂ of the spatial light modulation tube 502 is switched to addition, and the MCP voltage Vm ₂ is changed for a predetermined period of time.
Turn on x mag and x mlt to spatial light modulation tube 5
Add to 02.

After switching the crystal back voltage Vb ₁ of the spatial light modulation tube 501 to erase and turning on the MCP voltage Vm ₁ for a predetermined period of time to erase, the display device 561,
562 is turned ON, input x is input from the OPU to the display device 561, and φ{Σ(M _o・x
mlt)} for display, and the magnifying imaging system 51
1, each φ{Σ
(M _o x mlt)} Create mag, x mlt.
And the crystal back voltage of the spatial light modulation tube 501
While slowly lowering Vb ₁ , the MCP voltage
By turning on Vm ₁ during that time and operating the real-time threshold,
φ{Σ(M _o・x mlt)} between mag and x mlt
Execute the AND operation, [φ{Σ(M _o・x mlt)}
mag]·x mlt is written into the spatial light modulation tube 501.

The crystal back voltage Vb ₂ of the spatial light modulation tube 502 is switched to subtraction, and the MCP voltage Vm ₂ is turned on for a predetermined period of time, and the half mirror 52 is illuminated by monochromatic light.
1 from the spatial light modulation tube 501 through [φ{Σ
Read (M _o・x mlt)}mag〕・x mlt,
The result is written into the spatial light modulation tube 502 via the half mirror 522 and subtracted from the stored data. In this case, the result will not be less than "0" due to the effect of the mesh electrode.

The crystal back voltage Vb ₁ of the spatial light modulation tube 501 is switched to erase, and the MCP voltage Vm ₁ is turned on for a predetermined period of time to erase.

Memorization is performed by repeating the above operations.

The retrieval operation will be explained with reference to FIGS. 5A and 5C.

Turn on the display device 561 and input x to the CPU 55
1 and displayed, the multiplex imaging system 512
x mlt is created by Spatial light modulation tube 5
The crystal back voltage Vb ₁ of 01 is switched to writing, and the MCP voltage Vm ₁ is turned on for a predetermined period of time to enter the writing state, and x mlt is written.

The light receiving matrix 541 is turned on, and the monochromatic light is transmitted to the spatial light modulation tube 5 via the half mirror 521.
01, xmlt is read out, and based on the result, the contents of the spatial light modulation tube 502 are read out via the half mirror 522, and M _o ·xmlt is calculated. The calculation results are demultiplexed by the demultiplexing imaging system 513 to create Σ(M _o x mlt), which is converted into an electrical signal by the light receiving matrix 541 and sent to the CPU 55.
1. The CPU 551 performs threshold processing to obtain [φ{Σ(M _o ·x mlt)}. This output is fed back to the input side and a re-recall process is executed.

The crystal back voltage Vb ₁ of the spatial light modulation tube 501 is switched to erase, and the MCP voltage Vm ₁ is turned on for a predetermined period of time to erase, and then the display device 561 is turned on.
Then, [φ{Σ(M _o x mlt)} is the display device 5
61. At this time, the spatial light modulation tube 50
The data is written into the spatial light modulation tube 501 by switching the crystal back voltage Vb ₁ of 1 to write mode and turning on the MCP voltage Vm ₁ for a predetermined period of time. Then return to the process.

Recollection is performed by repeating the above process.

〔Effect of the invention〕

As described above, according to the present invention, a part of the memory contents is stored by an associative memory device that finds information that is the same as reference input from the outside or that meets certain conditions and retrieves the remaining related parts. Even if the memory is damaged, reading is not completely impossible, and access related to the stored contents becomes possible, which is useful for information retrieval.

[Brief explanation of drawings]

FIG. 1 is a block diagram of an associative memory device according to the present invention, and FIGS. 2, 3, 4, and 5 are examples of optical systems of the associative memory device of the present invention using spatial light modulation tubes. 6 is a diagram for explaining the configuration of a spatial light modulation tube, which is a basic component in an embodiment of the associative memory device according to the present invention, and FIG. 7 is a diagram for explaining the structure of a spatial light modulation tube between spatial light modulation devices. Figures 8 and 9 are diagrams for explaining secondary electron emission characteristics and writing and erasing; Figures 8 and 9 are diagrams for explaining AND operation using the real-time hard clip function of the space modulation tube;
The figures are diagrams for explaining an enlargement system, FIG. 11 is a diagram for explaining multiplexing, and FIG. 12 is a diagram for explaining a demultiplexing system. 1... Image, 2... Lens, 3... Photocathode, 4
...Microchannel plate, 5...Mesh electrode, 6...Crystal, 61...Charge storage surface, 7...
...half mirror, 8...monochromatic light, 9...analyzer,
10...Output light, 101...Enlargement system, 102...
...Multiplexing system, 103... First arithmetic unit, 104...
...Memory, 105...Second arithmetic unit, 106...
Demultiplexing system, 107... Threshold memory, 201-2
04...Spatial light modulation tube, 211...Enlarged imaging system,
212...Multiple imaging system, 213...Inverse multiple imaging system, 221-228...Half mirror, 231-
238...Reflector, S1, S2, S3...Shutter.

Claims (1)

[Claims] 1. A multiplexing system that multiplexes input or feedback signals and outputs the result, an expansion system that expands and outputs the input or feedback signal, and multiplies the output of the multiplexing system and the output of the expansion system and outputs the result. A first arithmetic unit, a memory that stores two-dimensional information, performs addition and subtraction on the content as necessary, and outputs the stored content, and a second arithmetic unit that multiplies the output of the multiplexing system and the output of the memory. and a demultiplexing system that demultiplexes and outputs the outputs of the first multiplier and the second multiplier, and a demultiplexing system that performs threshold operation on the output from the demultiplexing system and stores it and reads it out as necessary. An associative memory device having a memorization and recall section consisting of a threshold memory and a feedback system that returns output results to input.