Optical computation device

Abstract

An optical computation device (1) is provided with an optical computation unit (11) that includes an optical modulation element (11ai), and an image sensor (12) that generates an electric signal representing an intensity distribution of signal light output from the optical computation unit (11). A phase modulation amount of each of cells constituting the optical modulation element (11ai) is set such that a direction in which the signal light (SLi) is emitted is different from a direction in which noise light (NSi) is emitted. In this way, an optical computation device is achieved in which the influence of noise light on an electric signal generated by an image sensor is reduced.

Description

[Technical field]

[0001] The present invention relates to an optical arithmetic device that performs optical arithmetic using an optical modulation element having a plurality of cells in which the phase modulation amount is set independently or the phase modulation amount can be set independently. [Background technology]

[0002] There is known an optical modulation element that has a plurality of cells and is designed to optically execute a predetermined operation by causing the signal light transmitted through each cell to interfere with each other. The optical operation using such an optical modulation element has the advantage of being faster and consuming less power than an electrical operation using a processor. In addition, by sequentially applying the signal light to a plurality of optical modulation elements, it is possible to easily realize a high-level optical operation that is difficult to achieve with a single optical modulation element.

[0003] Patent Document 1 discloses an optical neural network having an input layer, an intermediate layer, and an output layer. The above-mentioned optical modulation element can be used, for example, as the intermediate layer of such an optical neural network. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] U.S. Patent No. 7,847,225 Summary of the Invention [Problem to be solved by the invention]

[0005] However, when the signal light before the operation is incident on the light modulation element, various noise lights are emitted from the light modulation element in addition to the signal light representing the operation result. When the light modulation element is a transmissive light modulation element, the signal light representing the operation result is generated by interference of transmitted light phase-modulated by each cell constituting the light modulation element. Of the noise lights, the noise light generated by the signal light before the operation passing through the light modulation element (specular transmission) without being phase-modulated by each cell is emitted in the same direction as the signal light. Also, when the light modulation element is a reflective light modulation element, the signal light representing the operation result is generated by interference of reflected light phase-modulated by each of the multiple cells. Of the noise lights, the noise light generated by the signal light before the operation being reflected by the light modulation element (specular reflection) without being phase-modulated by each cell is emitted in the same direction as the signal light.

[0006] Among the noise light emitted from each light modulation element, the noise light emitted in the same direction as the signal light is incident on the image sensor together with the signal light. Therefore, the electrical signal generated by the image sensor is affected by such noise light. If the effect of such noise light is large, it becomes difficult to correctly read the result of the optical calculation from the electrical signal generated by the image sensor.

[0007] One aspect of the present invention has been made in view of the above problems, and has an object to provide an optical computing device in which the effect of noise light on an electrical signal generated by an optical sensor is reduced. [Means for solving the problem]

[0008] An optical arithmetic device according to a first aspect of the present invention includes an optical arithmetic unit including an optical modulation element having a plurality of cells in which the phase modulation amount is set independently or the phase modulation amount can be set independently, and an optical sensor that detects signal light output from the optical arithmetic unit and generates an electrical signal representing the detection result, the optical modulation element being (1) a transmissive optical modulation element that emits signal light generated by interference of transmitted light phase-modulated by each of the plurality of cells and noise light that has passed through the optical modulation element without being phase-modulated by the plurality of cells, or (2) a reflected light modulated by each of the plurality of cells. a reflective optical modulation element that emits signal light generated by interference between reflected light from the plurality of cells, and noise light reflected to the optical modulation element without being phase-modulated by the plurality of cells, wherein in the optical modulation element, the amount of phase modulation of each of the plurality of cells is set so that a direction in which the signal light exits the optical modulation element and a direction in which the noise light exits the optical modulation element are different, and the effect of the noise light on the electrical signal is smaller than when the direction in which the signal light exits the optical modulation element and the direction in which the noise light exits the optical modulation element are the same. Effect of the Invention

[0009] According to one embodiment of the present invention, an optical computing device in which the influence of noise light on an electrical signal generated by an image sensor is reduced can be realized. [Brief description of the drawings]

[0010] [Figure 1] 1 is a side view showing a configuration of an optical calculation device according to a first embodiment of the present invention. [Diagram 2] 1. FIG. 4 is a side view showing a modified example of the optical arithmetic device shown in FIG. [Diagram 3] 2A and 2B are a plan view and a cross-sectional view showing a first configuration example of a light modulation element included in the optical arithmetic device shown in FIG. [Figure 4] 1. FIG. 4 is a cross-sectional view showing a second configuration example of the light modulation element included in the optical arithmetic device shown in FIG. [Diagram 5]1. FIG. 4 is a perspective view showing a third configuration example of the light modulation element included in the optical arithmetic device shown in FIG. [Figure 6] FIG. 11 is a side view showing the configuration of an optical calculation device according to a second embodiment of the present invention. [Figure 7] 7 is a side view showing a first modified example of the optical computing device shown in FIG. 6. [Figure 8] 7A to 7C are a top view, a side view, and a plan view showing a second modified example of the optical computing device shown in FIG. [Figure 9] 7A and 7B are a plan view and a cross-sectional view showing a first configuration example of an optical modulation element included in the optical arithmetic device shown in FIG. 6. [Figure 10] 7 is a cross-sectional view showing a second configuration example of the light modulation element included in the optical arithmetic device shown in FIG. 6. [Figure 11] 7 is a perspective view showing a third configuration example of the light modulation element included in the optical arithmetic device shown in FIG. 6. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0011] [First embodiment] (Configuration of Optical Computing Unit) The configuration of an optical arithmetic device 1 according to a first embodiment of the present invention will be described with reference to Fig. 1. Fig. 1 is a side view showing the configuration of the optical arithmetic device 1.

[0012] As shown in FIG. 1, the optical calculation device 1 includes an optical calculation unit 11 and an image sensor 12.

[0013] The optical calculation unit 11 is a set of at least one optical modulation element 11a1-11an (n is any natural number equal to or greater than 1). In this embodiment, in order to realize multi-stage optical calculation, a set of three optical modulation elements 11a1-11a3 is used as the optical calculation unit 11. The optical modulation elements 11a1-11an may be integrated. For example, n optical diffraction layers formed in a structure that transmits signal light, such as a dry gel, may be used as the optical modulation elements 11a1-11an.

[0014] Each light modulation element 11ai (i is a natural number between 1 and n) is composed of a plurality of cells in which the phase modulation amount is set independently of each other. Each light modulation element 11ai is a transmissive light modulation element, and performs optical calculation by causing transmitted light phase-modulated by each cell to interfere with each other. Here, performing optical calculation means converting the two-dimensional intensity distribution of the signal light from a two-dimensional intensity distribution representing information before calculation to a two-dimensional intensity distribution representing information after calculation. An example of the configuration of a transmissive light modulation element will be described later with reference to a different drawing.

[0015] In addition to the signal light described above, the light modulation element 11ai emits noise light that has passed through the light modulation element 11ai without being phase-modulated by each cell. For example, if there is a gap between the cells that constitute the light modulation element 11ai, the light that has passed through the gap becomes noise light. The optical axis of the noise light emitted from the light modulation element 11ai is located in a plane that is spanned by the optical axis of the signal light incident on the light modulation element 11ai and the normal to the incident surface of the light modulation element 11ai. The emission angle of the noise light emitted from the light modulation element 11ai is the same as the incidence angle at which the signal light is incident on the light modulation element 11ai. In other words, the optical axis of the noise light coincides with the optical axis of the transmitted light when the signal light is directly transmitted through the light modulation element 11ai.

[0016] In order to separate such noise light from the signal light, the phase modulation amount of each cell constituting the light modulation element 11ai is set so that the direction in which the signal light leaves the light modulation element 11ai differs from the direction in which the noise light leaves the light modulation element 11ai. In a transmissive light modulation element, the exit angle of the noise light is the same as the incident angle of the incident light. Therefore, the phase modulation amount of each cell is set so that the exit angle of the signal light is not the same as the incident angle of the incident light.

[0017] In this embodiment, the signal light SL0 is perpendicularly incident on the incident surface of the first light modulation element 11a1. The first light modulation element 11a1 perpendicularly emits the noise light NL1 from the emission surface and obliquely emits the signal light SL1. The second light modulation element 11a2 is arranged such that (1) its incident surface is parallel to the emission surface of the first light modulation element 11a1, (2) its incident surface intersects with the signal light SL1 emitted from the first light modulation element 11a1, and (3) its incident surface does not intersect with the optical axis of the noise light NL1 emitted from the first light modulation element 11a1. Instead of imposing the condition (3), a condition may be imposed that the optical axis of the noise light NL1 emitted from the first light modulation element 11a1 does not intersect with the intersection point between the incident surface of the second light modulation element 11a2 and the signal light SL1 emitted from the first light modulation element 11a1. Alternatively, a condition may be imposed that at least a part of the noise light NL1 emitted from the first light modulation element 11a1 is not incident on the incident surface of the second light modulation element 11a2.

[0018] In this embodiment, the second light modulation element 11a2 emits the noise light NL2 obliquely and emits the signal light SL2 perpendicularly. The third light modulation element 11a3 is arranged such that (1) its incident surface is parallel to the exit surface of the second light modulation element 11a2, (2) its incident surface intersects with the signal light SL2 emitted from the second light modulation element 11a2, and (3) its incident surface does not intersect with the optical axis of the noise light NL2 emitted from the second light modulation element 11a2. Instead of imposing the condition (3), a condition may be imposed that the optical axis of the noise light NL2 emitted from the second light modulation element 11a2 does not intersect with the intersection point between the incident surface of the third light modulation element 11a3 and the signal light SL2 emitted from the second light modulation element 11a2. Alternatively, a condition may be imposed that at least a part of the noise light NL2 emitted from the second light modulation element 11a2 is not incident on the incident surface of the third light modulation element 11a3. The third light modulation element 11a3 emits the noise light NL3 vertically and emits the signal light SL3 obliquely. In this embodiment, the optical axes of the signal lights SL0 to SL3 and the noise lights NL1 to NL3 are located in the same plane.

[0019] The image sensor 12 is an example of an optical sensor that detects the signal light SL3 output from the optical calculation unit 11 and generates an electrical signal representing the detection result. The image sensor 12 is composed of a plurality of photoelectric conversion units to generate an electrical signal representing the intensity distribution of the signal light SL3 output from the optical calculation unit 11. The image sensor 12 is arranged such that (1) its incident surface is parallel to the exit surface of the third light modulation element 11a3, (2) its incident surface intersects with the optical axis of the signal light SL3 output from the optical calculation unit 11, and (3) its incident surface does not intersect with the optical axes of the noise lights NL1 to NL3 output from the optical calculation unit 11. Note that instead of imposing the condition (3), a condition may be imposed that the optical axes of the noise lights NL1 to NL3 output from the optical calculation unit 11 do not intersect with the intersection point between the incident surface of the image sensor 12 and the optical axis of the signal light SL3 output from the optical calculation unit 11.

[0020] By arranging the light modulation elements 11a1-11a3 and the image sensor 12 as described above, the influence of the noise lights NL1-NL3 on the electrical signal generated by the image sensor 12 becomes smaller than that of a conventional optical calculation device in which the signal light SLi and the noise light NLi are emitted in the same direction from each light modulation element 11ai. Instead of imposing the condition that (any point on) the incident surface of the image sensor 12 does not intersect with the optical axis of the noise lights NL1-NL3, a condition may be imposed that the point at which the incident surface of the image sensor 12 intersects with the optical axis of the signal lights SL1-SL3 is different from the point at which the incident surface of the image sensor 12 intersects with the optical axis of the noise lights NL1-NL3. Even in this case, the intensity of the noise lights NL1-NL3 incident on the image sensor 12 can be reduced compared to the conventional configuration in which the center of the incident surface of the image sensor 12 intersects with the optical axis of the noise lights NL1-NL3. This makes it possible to suppress the influence of the noise lights NL1-NL3 on the electrical signal generated by the image sensor 12. The same can be said if a condition is imposed that at least a part of the noise light NL1 to NL3 is not incident on the incident surface of the image sensor 12.

[0021] In this embodiment, an arrangement is adopted for all photoelectric conversion units constituting the image sensor 12 so that noise light NL1 to NL3 is not incident at an intensity equal to or greater than the detection limit, but the present invention is not limited to this. For example, among the photoelectric conversion units constituting the image sensor 12, a photoelectric conversion unit to which signal light SL3 is incident at an intensity equal to or greater than the detection limit may be arranged so that noise light NL1 to NL3 is not incident at an intensity equal to or greater than the detection limit. Even in this case, the influence of the noise light NL1 to NL3 on the electrical signal generated by the image sensor 12 can be made smaller than that of a conventional optical computing device.

[0022] In addition, in this embodiment, a configuration is adopted in which the signal light SL0 two-dimensionally intensity modulated by the input data is input to the optical calculation unit 11, but the present invention is not limited to this. That is, a configuration may be adopted in which a carrier light is input to the optical calculation unit 11, and a signal light SL1 two-dimensionally modulated by the input data is generated by the optical modulation element 11a1. Furthermore, the optical calculation device 1 may further include a display that generates the signal light SL0 or a light source that generates the carrier light.

[0023] The setting of the phase modulation amount of each cell of each light modulation element 11ai can be realized by using, for example, machine learning. As a model used in this machine learning, for example, a model in which the two-dimensional intensity distribution of the signal light SL0 incident on the light modulation element 11a1 is input and the two-dimensional intensity distribution of the signal light SL3 incident on the image sensor 12 is output, and a model including the phase modulation amount in each cell of each light modulation element 11ai as a parameter can be used. Here, the two-dimensional intensity distribution of the signal light SL0 incident on the light modulation element 11a1 refers to, for example, a collection of intensities of the signal light SL0 incident on each cell constituting the light modulation element 11a1. Also, the two-dimensional intensity distribution of the signal light SL3 incident on the image sensor 12 refers to, for example, a collection of intensities of the signal light SL3 incident on each photoelectric conversion unit constituting the image sensor 12.

[0024] Furthermore, when the optical arithmetic unit 11 includes optical modulation elements 11ai whose cells have variable phase modulation amounts, the optical arithmetic device 1 may include a control unit that sets the phase modulation amounts of the cells of the optical modulation elements 11ai. This makes it possible to change the content of the optical arithmetic operation performed by the optical arithmetic device 1.

[0025] In addition, when the n optical diffraction layers formed in the dry gel are used as the optical modulation elements 11a1 to 11an, it is preferable to use a gel that shrinks while maintaining a similar shape by dehydration shrinkage, for example, a gel used in the Implosion Fabrication method. In this way, by drying the swollen gel in which the n optical diffraction layers are formed, the optical calculation unit 11 in which the n optical diffraction layers are precisely arranged can be easily manufactured.

[0026] (Modification of the optical computing device) A modified example of the optical arithmetic device 1 will be described with reference to Fig. 2. Fig. 2 is a side view showing the configuration of the optical arithmetic device 1 according to this modified example (hereinafter, referred to as optical arithmetic device 1A).

[0027] The optical arithmetic device 1A, like the optical arithmetic device 1, includes an optical arithmetic section 11 and an image sensor 12. The optical arithmetic device 1A differs from the optical arithmetic device 1 in the positions of the light modulation elements 11a1 to 11a3 and the image sensor 12.

[0028] In this modification, the signal light SL0 is obliquely incident on the incident surface of the first light modulation element 11a1. The first light modulation element 11a1 obliquely emits the noise light NL1 and the signal light SL1 in different directions. The second light modulation element 11a2 is arranged such that (1) its incident surface is parallel to the exit surface of the first light modulation element 11a1, (2) its incident surface intersects with the signal light SL1 emitted from the first light modulation element 11a1, and (3) its incident surface does not intersect with the optical axis of the noise light NL1 emitted from the first light modulation element 11a1. The second light modulation element 11a2 obliquely emits the noise light NL2 and the signal light SL2 in different directions. The third light modulation element 11a3 is arranged so that (1) its incident surface is parallel to the exit surface of the second light modulation element 11a2, (2) its incident surface intersects with the signal light SL2 emitted from the second light modulation element 11a2, and (3) its incident surface does not intersect with the optical axis of the noise light NL2 emitted from the second light modulation element 11a2. The third light modulation element 11a3 obliquely emits the noise light NL3 and the signal light SL3 in different directions. The image sensor 12 is arranged so that (1) its incident surface is parallel to the exit surface of the third light modulation element 11a3, (2) its incident surface intersects with the optical axis of the signal light SL3 emitted from the optical calculation unit 11, and (3) its incident surface does not intersect with the optical axes of the noise lights NL1 to NL3 emitted from the optical calculation unit 11. The optical axes of the signal lights SL0 to SL3 and the noise lights NL1 to NL3 are located in the same plane.

[0029] By arranging the light modulation elements 11a1 to 11a3 and the image sensor 12 as described above, the influence of the noise lights NL1 to NL3 on the electrical signal generated by the image sensor 12 is smaller than that of a conventional optical computing device in which the signal light SLi and the noise light NLi are emitted in the same direction from each light modulation element 11ai.

[0030] (Example of the configuration of a transmissive light modulation element) A configuration example of the light modulation element 11ai included in the optical calculation device 1 will be described with reference to Fig. 3. Fig. 3(a) is a plan view of the light modulation element 11ai according to this example. Fig. 3(b) is a cross-sectional view of a cell C constituting the light modulation element 11ai according to this example.

[0031] 3(a), the optical modulation element 11ai is composed of a plurality of cells C in which the phase modulation amount is set independently of each other. When the signal light SLi-1 is incident on the optical modulation element 11ai, the signal light SLi-1 that is phase-modulated and transmitted through each cell C interferes with each other to form the signal light SLi that represents the calculation result.

[0032] The cell C constituting the light modulation element 11ai is a microcell. Here, the term "microcell" refers to a cell with a cell size of less than 10 μm. Also, the term "cell size" refers to the square root of the area of ​​the cell. For example, when the shape of the microcell in plan view is a square, the cell size of the microcell is the length of one side of the microcell. The lower limit of the cell size of the microcell is, for example, 1 nm.

[0033] The light modulation element 11ai illustrated in (a) of Fig. 3 is composed of 200 x 200 cells C arranged in a matrix. The shape of each cell C in a plan view is a square of 500 nm x 500 nm, and the shape of the light modulation element 11ai in a plan view is a square of 100 µm x 100 µm. In this case, the signal light SLi-1 transmitted through the gaps between the cells C or the periphery of the cells C without being phase-modulated by the cells C becomes noise light.

[0034] Each cell C constituting the light modulation element 11ai can be composed of, for example, a polarizing plate C11, a polarizing plate C12, a first electrode C13, a magnetization free layer C14, an insulating layer C15, a magnetization fixed layer C16, and a second electrode C17, as shown in (b) of Figure 3.

[0035] The polarizing plate C11 and the polarizing plate C12 are disposed so as to face each other. The first electrode C13, the magnetization free layer C14, the insulating layer C15, the magnetization fixed layer C16, and the second electrode C17 are laminated in this order and sandwiched between the polarizing plate C11 and the polarizing plate C12. Here, the lamination direction of the first electrode C13, the magnetization free layer C14, the insulating layer C15, the magnetization fixed layer C16, and the second electrode C17 is the same as that of the polarizing plate C11 and the polarizing plate C12. Polarizing plateTherefore, a first side surface of the magnetization free layer C14 is in surface contact with one of the main surfaces of the polarizing plate C11, and a second side surface of the magnetization free layer C14 facing the first side surface is Polarizing plate The polarizing plate C11 is in surface contact with one of the principal surfaces of the magnetization free layer C12. The signal light SLi-1 (1) enters the inside of the magnetization free layer C14 via the polarizing plate C11, (2) passes through the magnetization free layer C14, and (3) exits from the magnetization free layer C14 via the polarizing plate C12.

[0036] The magnetization free layer C14 is made of, for example, a soft magnetic material (for example, CoFeB) having electrical conductivity and light transmission. The magnetization fixed layer C16 is made of, for example, a hard magnetic material (for example, Permalloy) having electrical conductivity. The polarizers C11 and C12 are selected as polarizers that selectively transmit a polarized light component whose polarization direction P is parallel to the magnetization direction M of the magnetization fixed layer C16. FIG. 3B illustrates a case where the magnetization direction M and the polarization direction P are parallel to both the main surface of the polarizer C11 and the main surface of the magnetization fixed layer C16.

[0037] When a potential difference is applied between the first electrode C13 and the second electrode C17, a spin current (a flow of spin-polarized electrons) is injected from the magnetization fixed layer C16 through the insulating layer C15 into the magnetization free layer C14 due to the tunnel effect, and magnetization occurs in the magnetization free layer C14. Here, the magnetization generated in the magnetization free layer C14 is magnetization parallel to the magnetization direction M of the magnetization fixed layer C16, that is, magnetization parallel to the polarization direction P of the signal light SLi-1 incident on the magnetization free layer C14 through the polarizing plate C11. For this reason, the phase of the signal light SLi-1 is delayed by the transverse Kerr effect in the process of propagating through the magnetization free layer C14.

[0038] Here, the phase shift of the signal light SLi-1 in cell C is toneThe amount of phase modulation of the cell C is determined according to the magnitude of magnetization generated in the magnetization free layer C14. The magnitude of magnetization generated in the magnetization free layer C14 is determined according to the magnitude of the spin current injected into the magnetization free layer C14. The magnitude of the spin current injected into the magnetization free layer C14 is determined according to the potential difference applied between the first electrode C13 and the second electrode C17. Therefore, by controlling the potential difference applied between the first electrode C13 and the second electrode C17, the amount of phase modulation of the cell C can be set to a desired value.

[0039] In this configuration example, the cell C having the same configuration as the STT (Spin Transfer Torque) type MRAM (Magnetoresistive Random Access Memory) has been described, but the present invention is not limited thereto. For example, the cell C having the same configuration as the SOT (Spin Orbit Torque) type MRAM may be used. Such a cell C can be realized, for example, by removing the insulating layer C15, the magnetization fixed layer C16, and the second electrode C17 from the structure shown in FIG. 3(b). In this case, for example, the first electrode C13 contains a heavy metal, and a pulse voltage or pulse current is applied to the first electrode C13, thereby efficiently injecting a spin current into the magnetization free layer C14.

[0040] Furthermore, the cell C may be configured similarly to a cell of an LCD (Liquid Crystal Display). In this case, the cell C is configured of a first glass substrate C21a, a first transparent electrode C22a, a liquid crystal layer C23, a second transparent electrode C22b, and a second glass substrate C21b, as shown in Fig. 4. The first glass substrate C21a, the first transparent electrode C22a, the liquid crystal layer C23, the second transparent electrode C22b, and the second glass substrate C21b are laminated in this order from the incident side of the signal light SLi-1.

[0041] The signal light SLi-1 (1) passes through the first glass substrate C21a and the first transparent electrode C22a, (2) passes through the liquid crystal layer C23, and (3) passes through the second transparent electrode C22b and the second glass substrate C21b. The liquid crystal layer C23 is composed of liquid crystal molecules aligned in a direction parallel to the main surface of the glass substrate C21, and has a refractive index according to the potential difference between the first transparent electrode C22a and the second transparent electrode C22b. Therefore, the signal light SLi-1 is subjected to phase modulation when passing through the liquid crystal layer C23. The amount of phase modulation of the cell C can be set to a desired value by controlling the potential difference applied between the first transparent electrode C22a and the second transparent electrode C22b.

[0042] The light modulation element 11ai may be composed of a plurality of cells C whose thicknesses are set independently of each other. Fig. 5 is an enlarged perspective view showing a portion of the light modulation element 11ai composed of a plurality of cells C whose thicknesses are set independently of each other.

[0043] The light modulation element 11ai shown in Fig. 5 is composed of a transparent substrate C31 and a plurality of pillars formed on the upper surface of the transparent substrate C31. Each pillar is a rectangular columnar structure having a square base with the length of each side equal to the cell size, and functions as a cell C. In this case, the signal light SLi-1 transmitted between the pillars constituting each cell C (the peripheral areas of each cell where no pillars are formed) without being phase-modulated by the cell C becomes noise light.

[0044] The signal light SLi (1) enters from the top surface of the pillar, (2) passes through the pillar, (3) passes through the transparent substrate C31, and (4) exits from the bottom surface of the transparent substrate C31. The amount of phase modulation of the signal light SLi passing through each cell C is determined according to the height of the pillar constituting that cell C. That is, the amount of phase modulation of the signal light SLi passing through a cell C constituted by a tall pillar is large, and the amount of phase modulation of the signal light SLi passing through a cell C constituted by a short pillar is small. The amount of phase modulation of each cell C is fixed.

[0045] [Second embodiment] (Configuration of Optical Computing Unit) The configuration of an optical calculation device 2 according to a second embodiment of the present invention will be described with reference to Fig. 6. Fig. 6 is a side view showing the configuration of the optical calculation device 2.

[0046] As shown in FIG. 6, the optical calculation device 2 includes an optical calculation section 21, an image sensor 22, and a mirror 23.

[0047] The optical calculation unit 21 is a set of at least one optical modulation element 21a1-21an (n is any natural number equal to or greater than 1). In this embodiment, in order to realize multi-stage optical calculation, a set of three optical modulation elements 21a1-21a3 is used as the optical calculation unit 21. The optical modulation elements 21a1-21an may be integrated. For example, n optical diffraction layers formed in a structure that transmits signal light, such as a dry gel, may be used as the optical modulation elements 21a1-21an. Alternatively, the optical modulation elements 21a1-21an may be embedded in a single substrate, or n regions of a single optical modulation element may be used as the optical modulation elements 21a1-21an.

[0048] Each light modulation element 21ai (i is a natural number between 1 and n) is composed of a plurality of cells whose phase modulation amount can be set independently of each other. Each light modulation element 21ai is a reflective light modulation element, and performs optical calculation by causing reflected light phase-modulated by each cell to interfere with each other. Here, performing optical calculation means converting the two-dimensional intensity distribution of the signal light from a two-dimensional intensity distribution representing information before calculation to a two-dimensional intensity distribution representing information after calculation. An example of the configuration of a reflective light modulation element will be described later with reference to a different drawing.

[0049] In addition to the above-mentioned signal light, the optical modulation element 21ai outputs the optical signal without being phase-modulated by each cell. reflected toThe light modulator 21ai emits noise light. For example, the light reflected on the surface of the light modulator 21ai without entering the inside of the cells constituting the light modulator 21ai becomes the noise light. The optical axis of the noise light emitted from the light modulator 21ai is located in a plane that is spanned by the optical axis of the signal light incident on the light modulator 21ai and the normal to the incident surface of the light modulator 21ai, and the emission angle of the noise light emitted from the light modulator 21ai is the same as the incidence angle of the signal light incident on the light modulator 21ai. In other words, the optical axis of the noise light coincides with the optical axis of the reflected light when the signal light is specularly reflected by the light modulator 21ai.

[0050] In order to separate such noise light from the signal light, the phase modulation amount of each cell constituting the light modulation element 21ai is set so that the direction in which the signal light leaves the light modulation element 21ai differs from the direction in which the noise light leaves the light modulation element 21ai. In a reflective light modulation element, for example, the exit angle of the noise light is the same as the incident angle of the incident light. Therefore, the phase modulation amount of each cell is set so that the exit angle of the signal light is not the same as the incident angle of the incident light.

[0051] The mirror 23 is a means for reflecting the signal light SL1 to SL3 emitted from the light modulation elements 21a1 to 21a3, and is disposed so that its reflecting surface faces the input and output surfaces of the light modulation elements 21a1 to 21a3.

[0052] In this embodiment, the signal light SL0 is obliquely incident on the first light modulation element 21a1. The first light modulation element 21a1 obliquely emits the noise light NL1 and the signal light SL1 in different directions. The noise light NL1 and the signal light SL1 emitted from the first light modulation element 21a1 are each reflected by the mirror 23. The second light modulation element 21a2 is arranged such that (1) its incident and exit surfaces are located on the same plane as the incident and exit surfaces of the first light modulation element 21a1, (2) its incident and exit surfaces intersect with the signal light SL1 reflected by the mirror 23, and (3) its incident and exit surfaces do not intersect with the optical axis of the noise light NL1 reflected by the mirror 23. Instead of imposing the condition (3), a condition may be imposed that the optical axis of the noise light NL1 emitted from the first light modulation element 21a1 does not intersect with an intersection point between the incident / exit surface of the second light modulation element 21a2 and the signal light SL1 emitted from the first light modulation element 21a1. Alternatively, a condition may be imposed that at least a part of the noise light NL1 emitted from the first light modulation element 21a1 does not enter the incident / exit surface of the second light modulation element 21a2.

[0053] In this embodiment, the second light modulation element 21a2 obliquely emits the noise light NL2 and the signal light SL2 in different directions. The noise light NL2 and the signal light SL2 emitted from the second light modulation element 21a2 are each reflected by the mirror 23. The third light modulation element 21a3 is arranged such that (1) its incident / exit surface is located on the same plane as that of the second light modulation element 21a2, (2) its incident / exit surface intersects with the signal light SL2 reflected by the mirror 23, and (3) its incident / exit surface does not intersect with the optical axis of the noise light NL2 reflected by the mirror 23. Instead of imposing the condition (3), a condition may be imposed that the optical axis of the noise light NL2 emitted from the second light modulation element 21a2 does not intersect with the intersection point between the incident / exit surface of the third light modulation element 21a3 and the signal light SL2 emitted from the second light modulation element 21a2. Alternatively, a condition may be imposed that at least a part of the noise light NL2 emitted from the second light modulation element 21a2 is not incident on the incident surface of the third light modulation element 21a3. The third light modulation element 21a3 obliquely emits the noise light NL3 and the signal light SL3 in different directions. The optical axes of the signal lights SL0 to SL3 and the noise lights NL1 to NL3 are located in the same plane.

[0054] The image sensor 22 is an example of an optical sensor that detects the signal light SL3 output from the optical calculation unit 21 and generates an electrical signal representing the detection result. The image sensor 22 is composed of a plurality of photoelectric conversion units to generate an electrical signal representing the intensity distribution of the signal light SL3 output from the optical calculation unit 21. The image sensor 22 is arranged such that (1) its incident surface is located on the same plane as the reflecting surface of the mirror 23, (2) its incident surface intersects with the optical axis of the signal light SL3 output from the optical calculation unit 21, and (3) its incident surface does not intersect with the optical axes of the noise light NL1 to NL3 output from the optical calculation unit 21.

[0055] By arranging the light modulation elements 21a1-21a3, the image sensor 22, and the mirror 23 as described above, it is possible to prevent the noise lights NL1-NL3 from being incident on the photoelectric conversion units of all the photoelectric conversion units constituting the image sensor 22 with an intensity equal to or greater than the detection limit of the photoelectric conversion units. Therefore, the influence of the noise lights NL1-NL3 on the electrical signal generated by the image sensor 22 is smaller than that of a conventional optical arithmetic device in which the signal light SLi and the noise light NLi are emitted in the same direction from each light modulation element 21ai. Instead of imposing the condition that (any point on) the incidence surface of the image sensor 22 does not intersect with the optical axis of the noise lights NL1-NL3, a condition (weaker condition) may be imposed that the point at which the incidence surface of the image sensor 22 intersects with the optical axis of the signal lights SL1-SL3 is different from the point at which the incidence surface intersects with the optical axis of the noise lights NL1-NL3. Even in this case, it is possible to reduce the intensity of the noise lights NL1 to NL3 incident on the image sensor 22, compared to the conventional configuration in which the center of the incident surface of the image sensor 22 intersects with the optical axes of the noise lights NL1 to NL3. This makes it possible to suppress the influence of the noise lights NL1 to NL3 on the electrical signal generated by the image sensor 22. The same can be said even if a condition is imposed that at least a part of the noise lights NL1 to NL3 is not incident on the incident surface of the image sensor 22.

[0056] In this embodiment, an arrangement is adopted for all photoelectric conversion units constituting the image sensor 22 such that noise light NL1 to NL3 is not incident at an intensity equal to or greater than the detection limit, but the present invention is not limited to this. For example, among the photoelectric conversion units constituting the image sensor 22, a photoelectric conversion unit to which signal light SL3 is incident at an intensity equal to or greater than the detection limit may be arranged such that noise light NL1 to NL3 is not incident at an intensity equal to or greater than the detection limit. Even in this case, the influence of the noise light NL1 to NL3 on the electrical signal generated by the image sensor 22 can be made smaller than that of a conventional optical computing device.

[0057] In addition, in this embodiment, a configuration is adopted in which the signal light SL0 two-dimensionally intensity modulated by the input data is input to the optical calculation unit 21, but the present invention is not limited to this. That is, a configuration may be adopted in which a carrier light is input to the optical calculation unit 21 and a signal light SL1 two-dimensionally modulated by the input data is generated by the optical modulation element 21a1. Furthermore, the optical calculation device 2 may further include a display that generates the signal light SL0 or a light source that generates the carrier light.

[0058] The setting of the phase modulation amount of each cell of each light modulation element 21ai can be realized by using, for example, machine learning. As a model used in this machine learning, for example, a model in which the two-dimensional intensity distribution of the signal light SL0 incident on the light modulation element 21a1 is input and the two-dimensional intensity distribution of the signal light SL3 incident on the image sensor 22 is output, and a model including the phase modulation amount in each cell of each light modulation element 21ai as a parameter can be used. Here, the two-dimensional intensity distribution of the signal light SL0 incident on the light modulation element 21a1 refers to, for example, a collection of intensities of the signal light SL0 incident on each cell constituting the light modulation element 21a1. Also, the two-dimensional intensity distribution of the signal light SL3 incident on the image sensor 22 refers to, for example, a collection of intensities of the signal light SL3 incident on each photoelectric conversion unit constituting the image sensor 22.

[0059] In addition, when the optical arithmetic unit 21 includes the optical modulation elements 21ai whose phase modulation amount of each cell is variable, the optical arithmetic unit 2 may include a control unit that sets the phase modulation amount of each cell of the optical modulation elements 21ai. This makes it possible to change the content of the optical arithmetic operation executed by the optical arithmetic unit 1.

[0060] Also, as described above, n areas of a single light modulation element may be used as the light modulation elements 21a1 to 21an. That is, n calculation areas A1, A2, ..., AN may be set in a single light modulation element, the calculation area A1 performs optical calculation by modulating and reflecting incident light, and each calculation area Ai (i is a natural number between 2 and N) other than the calculation area A1 performs optical calculation by modulating and reflecting the signal light reflected by the mirror 23 after being modulated and reflected by the calculation area Ai-1.

[0061] In addition, when the n optical diffraction layers formed in the dry gel are used as the optical modulation elements 21a1 to 21an, it is preferable to use a gel that shrinks while maintaining a similar shape by dehydration shrinkage, for example, a gel used in the Implosion Fabrication method. This makes it possible to easily manufacture the optical calculation unit 21 in which the n optical diffraction layers are precisely arranged by drying the swollen gel in which the n optical diffraction layers are formed.

[0062] (Modification 1 of the optical computing device) A first modified example of the optical arithmetic device 2 will be described with reference to Fig. 7. Fig. 7 is a side view showing the configuration of the optical arithmetic device 2 according to this modified example (hereinafter, referred to as optical arithmetic device 2A).

[0063] The optical arithmetic device 2A, like the optical arithmetic device 2, includes an optical arithmetic section 21, an image sensor 22, and a mirror 23. The optical arithmetic device 2A differs from the optical arithmetic device 2 in the positions of the light modulation elements 21a1-21a3, the position and orientation of the image sensor 22, and the orientation of the mirror 23.

[0064] In the optical arithmetic device 2, the mirror 23 is disposed so that its reflection surface is parallel to the incident and exit surfaces of the optical modulation elements 21a1 to 21a3. In contrast, in the optical arithmetic device 2A, the mirror 23 is disposed so that its reflection surface is non-parallel to the incident and exit surfaces of the optical modulation elements 21a1 to 21a3. In addition, in the optical arithmetic device 2, the image sensor 22 is disposed so that its incident surface is parallel to the incident and exit surfaces of the optical modulation elements 21a1 to 21a3. In contrast, in the optical arithmetic device 2A, the image sensor 22 is disposed so that its incident surface is non-parallel to the incident and exit surfaces of the optical modulation elements 21a1 to 21a3. Note that the fact that the incident surface of the image sensor 22 is located on the same plane as the reflection surface of the mirror 23 is common to the optical arithmetic device 2 and the optical arithmetic device 2A.

[0065] In this modification, the signal light SL0 is obliquely incident on the first light modulation element 21a1. The first light modulation element 21a1 obliquely emits the noise light NL1 and the signal light SL1 in different directions. The noise light NL1 and the signal light SL1 emitted from the first light modulation element 21a1 are each reflected by the mirror 23. The second light modulation element 21a2 is arranged such that (1) its incident and exit surface is located on the same plane as the incident and exit surface of the first light modulation element 21a1, (2) its incident and exit surface intersects with the signal light SL1 reflected by the mirror 23, and (3) its incident and exit surface does not intersect with the optical axis of the noise light NL1 reflected by the mirror 23. The second light modulation element 21a2 obliquely emits the noise light NL2 and the signal light SL2 in different directions. The noise light NL2 and the signal light SL2 emitted from the second light modulation element 21a2 are each reflected by the mirror 23. The third light modulation element 21a3 is arranged such that (1) its incident and exit surfaces are located on the same plane as the incident and exit surfaces of the second light modulation element 21a2, (2) its incident and exit surfaces intersect with the signal light SL2 reflected by the mirror 23, and (3) its incident and exit surfaces do not intersect with the optical axis of the noise light NL2 reflected by the mirror 23. The third light modulation element 21a3 obliquely emits the noise light NL3 and the signal light SL3 in different directions. The image sensor 22 is arranged such that (1) its incident surface is located on the same plane as the reflecting surface of the mirror 23, (2) its incident surface intersects with the optical axis of the signal light SL3 emitted from the optical calculation unit 21, and (3) its incident surface does not intersect with the optical axes of the noise lights NL1 to NL3 emitted from the optical calculation unit 21. The optical axes of the signal lights SL0 to SL3 and the noise lights NL1 to NL3 are located in the same plane.

[0066] By arranging the light modulation elements 21a1-21a3, the image sensor 22, and the mirror 23 as described above, it is possible to prevent the noise light NL1-NL3 from being incident on the photoelectric conversion unit with an intensity equal to or greater than the detection limit of the photoelectric conversion unit for all the photoelectric conversion units constituting the image sensor 22. Therefore, the influence of the noise light NL1-NL3 on the electric signal generated by the image sensor 22 is smaller than that of the conventional optical arithmetic device in which the signal light SLi and the noise light NLi are emitted in the same direction from each light modulation element 21ai. Moreover, in this modification, the angles of incidence of the signal light SL0-SL2 on the light modulation elements 21a1-21a3 and the angles of emission of the signal light SL1-SL3 from the light modulation elements 21a1-21a3 can be made common. Therefore, the design and manufacture of the light modulation elements 21a1-21a3 becomes easier.

[0067] (Modification 2 of the Optical Computing Device) A second modified example of the optical arithmetic device 2 will be described with reference to Fig. 8. In Fig. 8, (a) is a top view of the optical arithmetic device 2 according to this modified example (hereinafter, referred to as optical arithmetic device 2B), (b) is a side view of the optical arithmetic device 2B, and (c) is a front view of the optical arithmetic device 2B. Note that in (a) of Fig. 8, the mirror 23 is omitted from the illustration.

[0068] The optical calculation device 2B includes an optical calculation unit 21, an image sensor 22, and a mirror 23, similar to the optical calculation device 2. B The first difference between the optical arithmetic device 2 and the optical arithmetic device 2 is the positions of the light modulation elements 21a1 to 21a3, the position and orientation of the image sensor 22, and the position and orientation of the mirror 23. B The second difference between the optical processor 2 and the optical modulator 21ai is the emission direction of the signal light SLi from each optical modulation element 21ai.

[0069] In the optical arithmetic device 2, the mirror 23 is disposed so that its reflecting surface is parallel to the incident and exit surfaces of the optical modulation elements 21a1-21a3. In contrast, in the optical arithmetic device 2B, the mirror 23 is disposed so that its reflecting surface is non-parallel to the incident and exit surfaces of the optical modulation elements 21a1-21a3. In addition, in the optical arithmetic device 2, the image sensor 22 is disposed so that its incident surface is located on the same plane as the reflecting surface of the mirror 23, whereas in the optical arithmetic device 2B, the image sensor 22 is disposed so that its incident surface is located on the same plane as the incident and exit surfaces of the optical modulation elements 21a1-21a3.

[0070] In the optical arithmetic device 2, the optical axis of the signal light SLi emitted from each optical modulation element 21ai is arranged in a plane including the optical axis of the signal light SLi-1 incident on the optical modulation element 21ai and the optical axis of the noise light NLi emitted from the optical modulation element 21ai. B In this embodiment, the optical axis of the signal light SLi emitted from each optical modulation element 21ai is positioned outside the plane that contains the optical axis of the signal light SLi-1 incident on that optical modulation element 21ai and the optical axis of the noise light NLi emitted from that optical modulation element 21ai.

[0071] In this modification, the signal light SL0 is obliquely incident on the first light modulation element 21a1. The first light modulation element 21a1 obliquely emits the noise light NL1 and the signal light SL1 in different directions. The noise light NL1 and the signal light SL1 emitted from the first light modulation element 21a1 are each reflected by the mirror 23. The second light modulation element 21a2 is arranged such that (1) its incident and exit surface is located on the same plane as the incident and exit surface of the first light modulation element 21a1, (2) its incident and exit surface intersects with the signal light SL1 reflected by the mirror 23, and (3) its incident and exit surface does not intersect with the optical axis of the noise light NL1 reflected by the mirror 23. The second light modulation element 21a2 obliquely emits the noise light NL2 and the signal light SL2 in different directions. The noise light NL2 and the signal light SL2 emitted from the second light modulation element 21a2 are each reflected by the mirror 23. The third light modulation element 21a3 is arranged such that (1) its incident and exit surfaces are located on the same plane as the incident and exit surfaces of the second light modulation element 21a2, (2) its incident and exit surfaces intersect with the signal light SL2 reflected by the mirror 23, and (3) its incident and exit surfaces do not intersect with the optical axis of the noise light NL2 reflected by the mirror 23. The third light modulation element 21a3 obliquely emits the noise light NL3 and the signal light SL3 in different directions. The noise light NL3 and the signal light SL3 emitted from the third light modulation element 21a3 are each reflected by the mirror 23. The image sensor 22 is arranged such that (1) its incident surface is located on the same plane as the reflecting surface of the mirror 23, (2) its incident surface intersects with the optical axis of the signal light SL3 reflected by the mirror 23, and (3) its incident surface does not intersect with the optical axes of the noise lights NL1 to NL3 reflected by the mirror 23.

[0072] By arranging the light modulation elements 21a1-21a3, the image sensor 22, and the mirror 23 as described above, it is possible to prevent the noise light NL1-NL3 from being incident on the photoelectric conversion unit with an intensity equal to or greater than the detection limit of the photoelectric conversion unit for all the photoelectric conversion units constituting the image sensor 22. Therefore, the influence of the noise light NL1-NL3 on the electric signal generated by the image sensor 22 is smaller than that of the conventional optical arithmetic device in which the signal light SLi and the noise light NLi are emitted in the same direction from each light modulation element 21ai. Moreover, in this modification, the angles of incidence of the signal light SL0-SL2 on the light modulation elements 21a1-21a3 and the angles of emission of the signal light SL1-SL3 from the light modulation elements 21a1-21a3 can be made common. Therefore, the design and manufacture of the light modulation elements 21a1-21a3 becomes easier.

[0073] (Example of the configuration of a reflective light modulation element) A configuration example of a reflective light modulation element 21ai included in the optical computation device 1 will be described with reference to Fig. 9. Fig. 9(a) is a plan view of the light modulation element 21ai according to this example. Fig. 9(b) is a cross-sectional view of a cell C constituting the light modulation element 21ai according to this example.

[0074] 9(a), the optical modulation element 21ai is composed of a plurality of cells C in which the phase modulation amount is set independently of each other. When the signal light SLi-1 is incident on the optical modulation element 21ai, the signal light SLi-1 reflected by each cell C while being phase modulated interferes with each other to form the signal light SLi representing the calculation result.

[0075] The cell C constituting the light modulation element 21ai is a microcell. Here, the term "microcell" refers to a cell with a cell size of less than 10 μm. Also, the term "cell size" refers to the square root of the area of ​​the cell. For example, when the shape of the microcell in plan view is a square, the cell size of the microcell is the length of one side of the microcell. The lower limit of the cell size of the microcell is, for example, 1 nm.

[0076] 9(a) is composed of 200×200 cells C arranged in a matrix. The shape of each cell C in plan view is a square of 500 nm×500 nm, and the shape of the light modulation element 21ai in plan view is a square of 100 μm×100 μm.

[0077] Each cell C constituting the light modulation element 21ai can be composed of a polarizing plate C11, a reflecting plate C18, a first electrode C13, a magnetization free layer C14, an insulating layer C15, a magnetization fixed layer C16, and a second electrode C17, as shown in (b) of Fig. 9. In this case, the signal light SLi-1 reflected by the surface of the polarizing plate C11 becomes noise light.

[0078] The polarizing plate C11 and the reflecting plate C18 are arranged to face each other. The first electrode C13, the magnetization free layer C14, the insulating layer C15, the magnetization fixed layer C16, and the second electrode C17 are stacked in this order and sandwiched between the polarizing plate C11 and the reflecting plate C18. Here, the stacking direction of the first electrode C13, the magnetization free layer C14, the insulating layer C15, the magnetization fixed layer C16, and the second electrode C17 is perpendicular to the stacking direction of the polarizing plate C11 and the reflecting plate C18. Therefore, the first side surface of the magnetization free layer C14 is in surface contact with one of the main surfaces of the polarizing plate C11, and the second side surface of the magnetization free layer C14 facing the first side surface is in surface contact with one of the main surfaces of the reflecting plate C18. The signal light SLi-1 (1) enters the inside of the magnetization free layer C14 via the polarizing plate C11, (2) is reflected by the reflecting plate C18, and (3) exits to the outside of the magnetization free layer C14 via the polarizing plate C11.

[0079] The magnetization free layer C14 is made of, for example, a conductive and light-transmitting soft magnetic material (for example, CoFeB). The magnetization fixed layer C16 is made of, for example, a conductive hard magnetic material (for example, Permalloy). The polarizer C11 is selected to selectively transmit a polarized light component whose polarization direction P is parallel to the magnetization direction M of the magnetization fixed layer C16. FIG. 9B illustrates a case where the magnetization direction M and the polarization direction P are parallel to both the main surface of the polarizer C11 and the main surface of the magnetization fixed layer C16.

[0080] When a potential difference is applied between the first electrode C13 and the second electrode C17, a spin current (a flow of spin-polarized electrons) is injected from the magnetization fixed layer C16 through the insulating layer C15 into the magnetization free layer C14 due to the tunnel effect, and magnetization occurs in the magnetization free layer C14. Here, the magnetization generated in the magnetization free layer C14 is magnetization parallel to the magnetization direction M of the magnetization fixed layer C16, that is, magnetization parallel to the polarization direction P of the signal light SLi-1 incident on the magnetization free layer C14 through the polarizing plate C11. For this reason, the phase of the signal light SLi-1 is delayed by the transverse Kerr effect in the process of propagating through the magnetization free layer C14.

[0081] Here, the phase shift of the signal light SLi-1 in cell C is tone The amount of magnetization is determined according to the magnitude of magnetization generated in the magnetization free layer C14. The magnitude of magnetization generated in the magnetization free layer C14 is determined according to the magnitude of the spin current injected into the magnetization free layer C14. The magnitude of the spin current injected into the magnetization free layer C14 is determined according to the potential difference applied between the first electrode C13 and the second electrode C17. Therefore, by controlling the potential difference applied between the first electrode C13 and the second electrode C17, the phase modulation of the cell C can be set to a desired value.

[0082] In this configuration example, the cell C having the same configuration as the STT (Spin Transfer Torque) type MRAM (Magnetoresistive Random Access Memory) has been described, but the present invention is not limited thereto. For example, the cell C having the same configuration as the SOT (Spin Orbit Torque) type MRAM may be used. Such a cell C can be realized, for example, by removing the insulating layer C15, the magnetization fixed layer C16, and the second electrode C17 from the structure shown in FIG. 9(b). In this case, for example, the first electrode C13 contains a heavy metal, and a pulse voltage or pulse current is applied to the first electrode C13, thereby efficiently injecting a spin current into the magnetization free layer C14.

[0083] The cell C may be configured similarly to an LCOS (Liquid Crystal On Silicon) cell. In this case, the cell C is configured, for example, as shown in Fig. 10, of a glass substrate C21, a transparent electrode C22, a liquid crystal layer C23, and a reflective electrode C24. The glass substrate C21, the transparent electrode C22, the liquid crystal layer C23, and the reflective electrode C24 are laminated in this order from the incident side of the signal light SLi-1.

[0084] The signal light SLi-1 (1) passes through the glass substrate C21 and the transparent electrode C22, (2) passes through the liquid crystal layer C23, (3) is reflected by the reflective electrode C24, (4) passes through the liquid crystal layer C23, and (5) passes through the transparent electrode C22 and the glass substrate C21. The liquid crystal layer C23 is composed of liquid crystal molecules aligned in a direction parallel to the main surface of the glass substrate C21, and has a refractive index according to the potential difference between the transparent electrode C22 and the reflective electrode C24. Therefore, the signal light SLi-1 is subjected to phase modulation when passing through the liquid crystal layer C23. The amount of phase modulation of the cell C can be set to a desired value by controlling the potential difference applied between the transparent electrode C22 and the reflective electrode C24.

[0085] The reflective light modulation element 21ai may be composed of a plurality of cells C each having a thickness set independently of one another. Fig. 11 is an enlarged perspective view showing a portion of the light modulation element 21ai composed of a plurality of cells C each having a thickness set independently of one another.

[0086] The light modulation element 21ai shown in Fig. 11 is composed of a reflector C32 and a number of pillars formed on the upper surface of the reflector C32. Each pillar is a rectangular columnar structure having a square base with the length of each side equal to the cell size, and functions as a cell C. In this case, in the region where the pillars are formed, the signal light SLi-1 reflected by the upper surface of the pillar becomes noise light. Also, in the region where the pillars are not formed, the reflector C The signal light SLi-1 reflected by the upper surface of 32 becomes noise light.

[0087] Signal light SLi-1 (1) enters from the top surface of the pillar, (2) passes through the pillar, (3) is reflected by reflector C32, (4) passes through the pillar, and (5) exits from the top surface of the pillar. The amount of phase modulation of signal light SLi-1 reflected from each cell C is determined according to the height of the pillar constituting that cell C. That is, the amount of phase modulation of signal light SLi-1 reflected from a cell C constituted by a tall pillar is large, and the amount of phase modulation of signal light reflected from a cell C constituted by a short pillar is small. The amount of phase modulation of each cell C is fixed.

[0088] [summary] An optical arithmetic device according to a first aspect of the present invention includes an optical arithmetic unit including an optical modulation element having a plurality of cells in which the phase modulation amount is set independently or the phase modulation amount can be set independently, and an optical sensor that detects signal light output from the optical arithmetic unit and generates an electrical signal representing the detection result, the optical modulation element being (1) a transmissive optical modulation element that emits signal light generated by interference of transmitted light phase-modulated by each of the plurality of cells and noise light that has passed through the optical modulation element without being phase-modulated by the plurality of cells, or (2) a reflected light modulated by each of the plurality of cells. a reflective optical modulation element that emits signal light generated by interference between reflected light from the plurality of cells, and noise light reflected to the optical modulation element without being phase-modulated by the plurality of cells, wherein in the optical modulation element, the amount of phase modulation of each of the plurality of cells is set so that a direction in which the signal light exits the optical modulation element and a direction in which the noise light exits the optical modulation element are different, and the effect of the noise light on the electrical signal is smaller than when the direction in which the signal light exits the optical modulation element and the direction in which the noise light exits the optical modulation element are the same.

[0089] According to the above configuration, it is possible to reduce the influence of noise light on the electrical signal generated by the optical sensor.

[0090] In an optical computing device according to aspect 2 of the present invention, in addition to the configuration of aspect 1, the optical sensor is an image sensor having a plurality of photoelectric conversion units, and the image sensor is arranged so that the noise light is not incident at an intensity above the detection limit on a photoelectric conversion unit on which the signal light is incident at an intensity above the detection limit.

[0091] According to the above configuration, the influence of noise light on the electrical signal generated by the image sensor can be further reduced.

[0092] An optical computing device according to a third aspect of the present invention employs, in addition to the configuration of the second aspect, a configuration in which the image sensor is arranged so that the noise light is not incident on any photoelectric conversion unit with an intensity equal to or higher than the detection limit. Alternatively, an optical computing device according to a third aspect of the present invention employs, in addition to the configuration of the first aspect, a configuration in which the optical sensor is arranged so that (1) its incident surface intersects with the optical axis of the signal light emitted from the optical computing unit, and (2) the intersection of the incident surface and the optical axis of the signal light emitted from the optical computing unit does not intersect with the optical axis of the noise light emitted from the optical computing unit.

[0093] According to the above configuration, the influence of noise light on the electrical signal generated by the image sensor can be further reduced.

[0094] In an optical computing device according to a fourth aspect of the present invention, in addition to the configuration of any one of the first to third aspects, the optical computing unit includes at least a first transmissive optical modulation element and a second transmissive optical modulation element as the optical modulation elements, and the second transmissive optical modulation element is arranged such that (1) its incident surface intersects with the optical axis of the signal light emitted from the first transmissive optical modulation element, and (2) the intersection of its incident surface and the optical axis of the signal light emitted from the first transmissive optical modulation element does not intersect with the optical axis of the noise light emitted from the first transmissive optical modulation element. Furthermore, in an optical computing device according to aspect 5 of the present invention, in addition to the configuration of any one of aspects 1 to 3, the optical computing unit includes a first transmissive optical modulation element and a second transmissive optical modulation element as the optical modulation elements, and the second transmissive optical modulation element is configured such that (1) its incident surface intersects with the optical axis of the signal light emitted from the first transmissive optical modulation element, and (2) the second transmissive optical modulation element is positioned such that at least a portion of the noise light emitted from the first transmissive optical modulation element does not enter its incident surface.

[0095] According to the above configuration, when the light modulation element is a transmissive light modulation element, it is possible to effectively reduce the influence of noise light on the electrical signal generated by the optical sensor.

[0096] The sixth aspect of the present invention computing device In addition to the configuration of any one of aspects 1 to 3, a mirror is further provided, and the optical calculation unit includes at least a first reflective light modulation element and a second reflective light modulation element as the light modulation element, and the second reflective light modulation element has: (1) an incident and exit surface thereof that is different from the first reflective light modulation element; element and (2) the intersection point between the input / output surface and the optical axis of the signal light output from the first reflective optical modulation element and reflected by the mirror does not intersect with the optical axis of the noise light output from the first reflective optical modulation element and reflected by the mirror, or the optical axis of the noise light output from the first reflective optical modulation element. Calculations In addition to the configuration of any one of aspects 1 to 3, the device further includes a mirror, and the optical calculation unit includes a first reflective light modulation element and a second reflective light modulation element as the light modulation elements, and the second reflective light modulation element is configured such that (1) its incident / exit surface intersects with the optical axis of the signal light emitted from the first reflective light modulation element and reflected by the mirror, and (2) the noise light emitted from the first reflective light modulation element and reflected by the mirror, or at least a portion of the noise light emitted from the first reflective light modulation element, does not enter the incident / exit surface.

[0097] According to the above configuration, when the light modulation element is a reflective light modulation element, it is possible to effectively reduce the influence of noise light on the electrical signal generated by the optical sensor.

[0098] In an optical computing device according to aspect 8 of the present invention, in addition to the configuration of aspect 6 or 7, a configuration is adopted in which the reflective surface of the mirror is non-parallel to the input and output surfaces of the first reflective light modulation element and the second reflective light modulation element.

[0099] According to the above configuration, the angle of incidence of the signal light to the first reflective optical modulation element and the angle of incidence of the signal light to the second reflective optical modulation element can be made common, and the angle of emission of the signal light from the first reflective optical modulation element and the angle of emission of the signal light from the second reflective optical modulation element can be made common, which makes it easier to design and manufacture the optical arithmetic device.

[0100] In an optical computing device according to aspect 9 of the present invention, in addition to the configuration of aspect 6 or 7, the optical computing unit is configured to be composed of a single optical modulation element having areas functioning as the first reflective optical modulation element and the second reflective optical modulation element.

[0101] According to the above configuration, since it is not necessary to individually move the light modulation elements to optimize the positional relationship of the light modulation elements, a highly accurate optical calculation unit can be easily manufactured.

[0102] In an optical arithmetic device according to a tenth aspect of the present invention, in addition to the configuration of any one of the first to ninth aspects, the optical arithmetic unit is a structure that transmits signal light, modulation The structure has an optical diffraction layer formed therein, which functions as an element.

[0103] According to the above configuration, since it is not necessary to individually move the light modulation elements to optimize the positional relationship of the light modulation elements, a highly accurate optical calculation unit can be easily manufactured.

[0104] [Additional Notes] The present invention is not limited to the above-described embodiments, and various modifications are possible within the scope of the claims. Embodiments obtained by appropriately combining the technical means included in the above-described embodiments are also included in the technical scope of the present invention. [Explanation of symbols]

[0105] 1,2 Optical calculation device 11,21 Optical calculation section 11a1 to 11a3 Light modulation element (transmissive light modulation element) 21a1 to 21a3 Light modulation element (reflective light modulation element) 12,22 Image sensor

Claims

1. an optical arithmetic unit including an optical modulation element having a plurality of cells in which a phase modulation amount is set independently or in which the phase modulation amount can be set independently; and an optical sensor that detects the signal light output from the optical arithmetic unit and generates an electrical signal representing the detection result, the light modulation element is (1) a transmissive light modulation element that outputs signal light generated by interference of transmitted light phase-modulated by each of the plurality of cells, and noise light that has passed through the light modulation element without being phase-modulated by the plurality of cells, or (2) a reflective light modulation element that outputs signal light generated by interference of reflected light phase-modulated by each of the plurality of cells, and noise light that has been reflected to the light modulation element without being phase-modulated by the plurality of cells, In the light modulation element, the phase modulation amount of each of the plurality of cells is set so that a direction in which the signal light is emitted from the light modulation element and a direction in which the noise light is emitted from the light modulation element are different from each other, an effect of the noise light on the electrical signal is smaller than that in a case where the direction in which the signal light is emitted from the optical modulation element and the direction in which the noise light is emitted from the optical modulation element are the same; An optical computing device comprising:

2. the optical sensor is an image sensor having a plurality of photoelectric conversion units; the image sensor is disposed so that the noise light is not incident at an intensity equal to or greater than a detection limit on a photoelectric conversion unit to which the signal light is incident at an intensity equal to or greater than a detection limit, 2. The optical computing device according to claim 1 .

3. The optical sensor is disposed such that (1) its incident surface intersects with the optical axis of the signal light emitted from the optical calculation unit, and (2) an intersection point between the incident surface and the optical axis of the signal light emitted from the optical calculation unit does not intersect with the optical axis of the noise light emitted from the optical calculation unit.

2. The optical computing device according to claim 1 .

4. the optical calculation unit includes a first transmissive optical modulation element and a second transmissive optical modulation element as the optical modulation elements, the second transmissive light modulation element is disposed such that (1) its incident surface intersects with the optical axis of the signal light emitted from the first transmissive light modulation element, and (2) an intersection point between the incident surface and the optical axis of the signal light emitted from the first transmissive light modulation element does not intersect with the optical axis of the noise light emitted from the first transmissive light modulation element.

2. The optical computing device according to claim 1 .

5. the optical calculation unit includes a first transmissive optical modulation element and a second transmissive optical modulation element as the optical modulation elements, the second transmissive light modulation element is arranged such that (1) its incident surface intersects with the optical axis of the signal light output from the first transmissive light modulation element, and (2) at least a part of the noise light output from the first transmissive light modulation element is not incident on the incident surface; 2. The optical computing device according to claim 1 .

6. Further comprising a mirror; the optical calculation unit includes a first reflective optical modulation element and a second reflective optical modulation element as the optical modulation elements, The second reflective light modulation element is arranged such that (1) its input / output surface intersects with the optical axis of the signal light output from the first reflective light modulation element and reflected by the mirror, and (2) an intersection between the input / output surface and the optical axis of the signal light output from the first reflective light modulation element and reflected by the mirror does not intersect with the optical axis of the noise light output from the first reflective light modulation element and reflected by the mirror, or the noise light output from the first reflective light modulation element.

2. The optical computing device according to claim 1 .

7. Further comprising a mirror; the optical calculation unit includes a first reflective optical modulation element and a second reflective optical modulation element as the optical modulation elements, The second reflective light modulation element is arranged such that (1) its incident / exit surface intersects with the optical axis of the signal light emitted from the first reflective light modulation element and reflected by the mirror, and (2) the noise light emitted from the first reflective light modulation element and reflected by the mirror, or at least a part of the noise light emitted from the first reflective light modulation element, is not incident on the incident / exit surface.

2. The optical computing device according to claim 1 .

8. a reflecting surface of the mirror is non-parallel to the input and output surfaces of the first reflective light modulation element and the second reflective light modulation element; 8. The optical computing device according to claim 6, wherein the optical computing device is a first optical computing device.

9. the optical calculation unit is configured by a single optical modulation element having an area that functions as the first reflective optical modulation element and the second reflective optical modulation element; 8. The optical computing device according to claim 6, wherein the optical computing device is a first optical computing device.

10. The optical calculation unit is a structure that transmits signal light and has an optical diffraction layer formed therein that functions as the optical modulation element.

6. The optical computing device according to claim 1, wherein the optical computing device is a first optical computing device.