Multi-vector optical multiplier

Abstract

The invention relates to a multi-vector optical multiplier. Systems and methods for performing element-by-element multi-vector multiplication are provided. Examples include a waveguide to receive an input optical signal encoded with a first vector and output an output optical signal. One or more optoelectronic (O / E) converters are provided to receive one or more optical signals encoded with one or more vectors and to generate one or more electrical signals based on the received one or more optical signals. One or more optical modulators are optically coupled to the waveguide and electrically coupled to the one or more O / E converters, the one or more optical modulators modulating an intensity of the input optical signal on the waveguide based on the one or more electrical signals. The output optical signal is encoded with a product of the first vector and the one or more vectors.

Description

Background Art

[0001] In recent years, artificial neural networks ("ANNs") have become an important tool for machine learning. ANNs have a wide range of applications, such as but not limited to image recognition, natural language processing, and gaming. For more complex tasks, larger and deeper ANNs become more important. These larger and deeper ANNs require faster and more energy-efficient computing hardware. BRIEF DESCRIPTION OF THE DRAWINGS

[0002] The present disclosure is described in detail according to one or more different embodiments with reference to the following drawings. The drawings are provided for illustrative purposes only and merely depict typical or example embodiments.

[0003] Figure 1 A schematic block diagram of an optical multiplier 100 according to the presently disclosed technology is depicted.

[0004] Figure 2 An example element-by-element optical multiplier according to the presently disclosed techniques is depicted.

[0005] Figure 3 Depicted is a schematic block diagram of an element-wise multiplication that may be performed using examples of the presently disclosed techniques.

[0006] Figure 4A Depicted is an example schematic circuit of an opto-electrical converter driven optical modulator according to an example implementation of the presently disclosed technology.

[0007] Figure 4B The diagram shows Figure 4A A schematic circuit generates a graphical representation of example input signal currents and example voltage biases.

[0008] Figure 5A Depicted is a schematic diagram of an overcoupled microring-assisted Mach-Zehnder interferometer that may be used in examples of the presently disclosed technology.

[0009] Figure 5B Depicted in Figure 5A A graphical representation of the intensity of the output optical signal as a function of the phase shift in an example implementation of an overcoupled microring assisted Mach-Zehnder interferometer in FIG.

[0010] Figure 6 is an example computational component that may be used to implement various features of an element-by-element optical multiplier according to embodiments disclosed herein.

[0011] Figure 7 is an example computer system that can be used to implement various features of the disclosed element-by-element optical multiplier.

[0012] The drawings are not exhaustive and do not limit the disclosure to the precise forms disclosed. DETAILED DESCRIPTION

[0013] Examples of the presently disclosed technology provide devices and methods for performing element-by-element multiplication (i.e., Hadamard product). Implementations disclosed herein provide optical multipliers capable of performing element-by-element multiplication. The presently disclosed technology provides high-speed, energy-efficient optoelectronic devices that can be used to produce integrated and scalable optical multipliers. The disclosed technology provides optical inputs and optical outputs that are fully compatible with optical neural networks (ONNs).

[0014] As mentioned above, ANNs can be an important tool for machine learning. As ANNs become larger and deeper, they require faster and more energy-efficient computing hardware. However, traditional electronic ANNs have several limitations, including high energy consumption, limited processing speed, and susceptibility to electromagnetic interference. Therefore, ONNs have attracted increasing attention because ONNs rely on light rather than electronic signals to perform calculations and information processing. Relying on light can overcome the above-mentioned bottlenecks of electronic ANNs. For example, the operating rate of ultra-high-speed optoelectronic devices (such as photodiodes and modulators) can be one to two orders of magnitude higher than the operating rate of their electronic counterparts. Further, relying on light can provide ONNs that are resistant to electromagnetic interference, whose resistance is not affected by length, and have low energy consumption. Therefore, ONNs are considered to be a competitive and more efficient hardware solution.

[0015] Many machine learning applications have various data sources, including text, images, and audio data. The weighting and combination of data from multiple sources can complement the information from different modalities to obtain improved machine learning performance. For example, multimodal fusion networks can provide robust performance with multiple data. Element-wise multiplication, also known as the Hadamard product, is an operation that combines different modalities / data into a single output. Electronic circuits can implement vector element-wise multiplication by using operational amplifiers and resistors. However, the bandwidth of operational amplifiers is only a few kHz to tens of MHz. In addition, the energy consumption of these operational amplifiers is high, ranging from a few μW to a few mW.

[0016] Some approaches attempt to use optoelectronic devices to provide element-by-element multiplication. One approach provides a vector multiplier using balanced homodyne detection. In this case, the electric field and The two input data signals represented by are injected into the beam splitter and then detected by the balanced photodiode. The beam splitter mixes the two input data and the balanced PD detects the differential signal output from the beam splitter. The output current of the homodyne detection is proportional to the product of the two input electric fields provided, that is, Where R is the photodiode responsivity. This method can process the field multiplication of two optical vectors at high speed. However, since the vector multiplication originates from the photoelectric effect, that is, I PD =R·P=R·|E| 2 , this method is only valid for two vectors. Homodyne detection cannot be extended to multi-vector products, so this method is not sufficient for many machine learning applications. In addition, this method requires that the input data signal provided to the beam splitter is encoded onto a coherent optical signal of the same wavelength.

[0017] In this context, the technology disclosed herein provides an element-by-element optical multiplier that improves operation bandwidth and reduces energy consumption. Since the disclosed technology provides an optical modulator that uses an optical input and provides an optical output, the disclosed technology is fully compatible with existing ONNs. In addition, the disclosed technology can be used for element-by-element multiplication of two or more vectors and does not require coherence or common wavelength between the input optical signals.

[0018] An example of the currently disclosed technology provides an optical multiplier for implementing the Hadamard product by intensity multiplication of multiple vectors. The embodiments disclosed herein may be particularly suitable for element-by-element multiplication of multiple vectors. The examples disclosed herein receive multiple vectors as encoded optical signals and output an output optical signal encoded by a vector as the element-by-element product of the input vectors. Each input vector is encoded into a different optical signal in the time domain. The input optical signal encoded with a first vector is injected into a waveguide including one or more optical modulators. In an illustrative example, each optical modulator includes a Mach-Zehnder interferometer ("MZI"). The remaining optical signals are each encoded with one of the remaining vectors and injected into a separate waveguide terminating at a corresponding photoelectric (O / E) converter. The O / E converter can be implemented as a photodetector (PD) or a similar device. Each O / E converter is electrically connected to an optical modulator. In an illustrative example, each O / E converter is electrically connected to a phase shifter (PS), which is optically coupled to one of the optical modulators. Each O / E converter converts the encoded optical signal into an electrical signal, which is applied to the corresponding PS. The input optical signal encoded with the first vector is modulated by each optical modulator based on the corresponding PS driven by the corresponding electrical signal. For example, the photocurrent I generated by each O / E converter PD The square of the absolute value of the electric field (|E| 2 ) and is proportional to the photocurrent (I PD), the element-by-element intensity of the other vectors can be multiplied by the first vector. Therefore, the resulting optical signal output from the waveguide is encoded with the element-by-element product of the first vector and the other vectors. Since the O / E converter converts the optical signal into an electrical signal, each optical signal does not need to be coherent or have the same wavelength. The optical signals can have the same or different wavelengths.

[0019] In an example embodiment, a linear transformation is achieved by connecting a plurality of load resistors to an O / E converter. For example, a load resistor in the plurality of load resistors is connected to a corresponding O / E converter and converts the photocurrent from the corresponding O / E converter into a voltage bias. The voltage bias is applied to the PS of each optical modulator with a phase shift proportional to the applied electric field. Thus, each optical modulator modulates the intensity of the input optical signal according to one of the remaining vectors. Thus, the output optical signal is intensity modulated according to the element-by-element product of the first vector and the other vectors.

[0020] According to some examples, to achieve multiplication operations, a strongly overcoupled micro-ring assisted MZI (RAMZI) can be used as part of each optical modulator. For example, each signal detected by the O / E converter can be encoded into a PS that is optically coupled to a ring that is overcoupled with a corresponding MZI. As used herein, "overcoupling" refers to evanescent coupling between waveguides with a high coupling coefficient. The coupling coefficient is a value between 0 and 1, where a value of 1 represents perfect coupling or 100% coupling of light from one waveguide to another. A high coupling coefficient can be, for example, a coupling coefficient of 0.89 or higher. In an example, the coupling coefficient can be between 0.84 and 0.94, which provides a coupling efficiency deviation of less than 0.01. In an illustrative example, the coupling coefficient used to provide an illustrative strongly overcoupled ring can be a coupling coefficient of 0.92.

[0021] Due to the above reasons, the examples of the currently disclosed technology provide several advantages over conventional devices. For example, by using optoelectronic devices in an integrated system, efficiency can be improved relative to providing a separate optical system and an electronic chip. In addition, the examples disclosed herein can be reduced to a compact size and can be operated with low energy consumption. In addition, the disclosed technology can support high data rates (e.g., about tens to hundreds of Gbit / s), which is determined by the bandwidth of the O / E converter and the PS, and both can obtain bandwidths higher than 50GHz. Therefore, vectors can be quickly encoded onto optical signals. As another example, the currently disclosed technology provides a series structure that can be scaled up to any number of vectors for element-by-element multiplication (e.g., by adding additional optical modulators and O / E converters). Therefore, by flexibly changing the number of O / E converters and optical modulator pairs, any number of modalities or data sources can be combined. As another example advantage, element-by-element products can be passively calculated at the speed of light via optical interference in an optical modulator (e.g., optical interference in an MZI), which is much faster than electronic multipliers in terms of speed and energy consumption. In another example advantage, examples disclosed herein provide optical input and optical output structures that can be inserted into any location in an ONN to meet the needs of different architectures.

[0022] Figure 1 A schematic block diagram of an optical multiplier 100 according to the presently disclosed technology is depicted.

[0023] The optical multiplier 100 includes one or more O / E converters 114 for receiving one or more optical signals encoded with one or more vectors. That is, each of the one or more O / E converters 114 receives a corresponding optical signal in the one or more optical signals. Each of the one or more O / E converters 114 is configured to generate an electrical signal based on the corresponding optical signal.

[0024] The optical multiplier 100 also includes one or more optical modulators 122. Each of the one or more optical modulators 122 corresponds to one of the one or more O / E converters 114 and receives an electrical signal generated by the corresponding O / E converter 114. The optical modulator 122 also receives an input optical signal encoded with a first vector. Each optical modulator 122 is configured to modulate the intensity of the input optical signal based on a corresponding electrical signal received from the corresponding O / E converter 114. The optical modulator 122 outputs an output optical signal encoded with an output vector, which is an element-by-element product of the first vector and one or more vectors input to the O / E converter 114.

[0025] In various embodiments, the optical multiplier 100 includes one or more linear transformation devices 134, each of which corresponds to one of the one or more O / E converters 114 and one of the one or more optical modulators 122, and is configured to perform a linear transformation on the electrical signal generated by the one or more O / E converters 114. In an example, the O / E converter 114 generates a photocurrent, and the one or more linear transformation devices 134 are configured to perform a transformation to convert the photocurrent into a voltage bias applied to the one or more optical modulators 122. That is, each linear transformation device 134 converts the photocurrent from the corresponding O / E converter 114 into a voltage bias and applies the voltage bias to the corresponding optical modulator, which then modulates the intensity of the input optical signal based on the received voltage bias.

[0026] Figure 2 A schematic diagram of an example implementation of an element-by-element optical multiplier 200 according to the presently disclosed technology is depicted. The optical multiplier 200 can be implemented on a Si platform, where the entire structure of the optical multiplier 200 can be integrated into a single silicon-on-insulator (SOI) platform (e.g., a single photonic chip).

[0027] The optical multiplier 200 includes an input waveguide 202 that receives an input optical signal 204 encoded with a first vector 206 at an input end 208. The optical multiplier 200 also includes an output end 210 that provides an output optical signal 212. The optical multiplier 200 includes one or more O / E converters 214a-214n (collectively referred to herein as "O / E converters 214") for receiving one or more optical signals 216a-n (collectively referred to herein as "optical signals 216") encoded with one or more vectors 218a-n (collectively referred to herein as "vectors 218"). The one or more optical signals 216 are each injected into a waveguide 220a-n (collectively referred to herein as "waveguide 220"), each waveguide being optically coupled to a respective corresponding O / E converter 214. The one or more O / E converters 214 each generate a voltage bias based on the optical signal 216 received by each respective O / E converter 214. The optical multiplier 200 also includes one or more optical modulators 222a-n (collectively referred to herein as “optical modulators 222”), each optically coupled to the waveguide 202 and electrically coupled to the O / E converter 214. The one or more optical modulators 222 are configured to modulate the intensity of the input optical signal 204 propagating on the waveguide 202 based on one or more voltage biases generated by the O / E converter 214. Thus, the output optical signal 212 is encoded with an output vector 224, which is an element-wise product of the first vector 206 and the one or more vectors 218.

[0028] exist Figure 2 In the example of FIG. 1 , the one or more optical modulators 222 include one or more PSs 226a-n (collectively referred to herein as “PS 226”) electrically coupled to the O / E converter 214. More specifically, as Figure 2 As shown, each of the one or more optical modulators 222 includes a corresponding PS 226 electrically coupled to a corresponding O / E converter 214. Electrical coupling refers to a circuit path between two devices for transmitting electrical energy from one device to another device in this article. The circuit path can be provided by conductive coupling, impedance matching, etc. The O / E converter 214 generates an electrical signal (e.g., a voltage bias or current) based on the optical signal 216 received by each O / E converter. More specifically, the electrical signal generated by each O / E converter 214 can be proportional to the intensity of the optical signal 216 on the O / E converter 214. In various examples, a photodetector, an avalanche photodiode, a balanced photodiode, etc. can be provided for the O / E converter 214, which is configured to detect the intensity of the amount of photons incident thereon and convert the photons into proportional electrical signals. The PS 226 disclosed herein can be provided as, but not limited to, a phase shifter utilizing the Pockels Effect (e.g., linear electro-optical effect). These phase shifters can use non-centrosymmetric materials, such as but not limited to lithium niobate (LiNbO3), lithium tantalate (LiTaO3), beta-barium borate (BBO), potassium titanyl phosphate (KTP), etc. Although these materials are not silicon, they can be integrated on silicon integrated circuits by bonding, transfer printing, etc.

[0029] The one or more optical modulators 222 may also include one or more MZIs 232a-n (collectively referred to herein as "MZIs 232") optically coupled to the one or more PSs 226. Figure 2 In an example embodiment, the one or more MZIs 232 are one or more RAMZIs 228a-n (collectively referred to herein as “RAMZIs 228”), which include one or more MZIs 232 optically coupled to one or more microrings 230a-n (collectively referred to herein as “microrings 230”). More specifically, as Figure 2 As shown, each of the one or more optical modulators 222 includes a corresponding RAMZI 228 formed by an MZI 232 optically coupled to a microring 230 including a corresponding PS 226 formed thereon.

[0030] The optical multiplier 200 may be used to perform element-by-element multiplication. Figure 3 Depicted is a schematic block diagram of an element-by-element multiplication that can be performed using an example of the presently disclosed technology. More particularly, Figure 3 304 ( ) is from a plurality of input vectors 302a-302n (e.g., ), wherein each input vector 302a-302n includes a plurality of elements. That is, for example, the first input vector 302a ( ) and one or more other input vectors 302b-302n ( ) are multiplied by the first element of the output vector 304 to produce the first element 310 of the output vector 304, wherein the first element is the product of the first elements 308a-308n. Figure 3 In the illustrative example of , each input vector 302a-302n is shown as an N×1 vector that produces an N×1 output vector 304, where N is 8. However, N may be any integer desired for a given application.

[0031] refer to Figure 2 , the input vectors 302a-302n may be examples of the first vector 206 and the one or more vectors 218. That is, for example, the first input vector 302a ( ) may be an example of a first vector 206, while one or more other input vectors 302b-302n ( ) may be an example of one or more vectors 218. Thus, the output vector 304 ( ) can be an example of output vector 224.

[0032] Back to Figure 2During operation, an input optical signal 204 is encoded in the time domain with a first input vector 206 (e.g., first input vector 302a in one example), and one or more optical signals 216 are encoded with one or more vectors 218 (e.g., one or more input vectors 302b-302n in one example). For example, the intensity of each optical signal can be modulated according to the element of the corresponding vector used to encode the optical signal. As an illustrative example, the optical signal can be modulated to various intensity levels between 0 and a maximum value, where 0 intensity can represent a 0 element value, and the maximum intensity represents the maximum (or highest) element value in the element value across all input vectors (e.g., across all first vectors 206 and one or more vectors 218). Thus, discrete intensity levels can be used to represent different element values. Then, when a given optical signal propagates in the time domain, each intensity level can represent a sequential element value (e.g., the first time intensity represents the value of the first element, the second time intensity represents the value of the second element, and so on). Once encoded, the input optical signal 204 can be injected directly or indirectly, one or more optical modulators 222 and one or more optical signals 216 are each injected into a separate O / E converter 214. The O / E converter 214 converts the encoded optical signal 216 into an electrical signal, wherein each O / E converter 214 is electrically connected to a separate PS 226. The input optical signal 204 is modulated by the one or more optical modulators 222 via the PS 226 to generate an output optical signal 212 and an output vector 224, thereby obtaining the product of the one or more vectors 218.

[0033] As described above, examples disclosed herein may utilize intensity modulation of optical signals 204 and 216 to perform element-by-element multiplication. For example, each O / E converter 214 may be configured such that the photocurrent I generated by the O / E converter 214 PD and the received optical signal 216 (eg, I PD ∝|E| 2 ) is proportional to the electric field (E) applied to the O / E converter 214. This proportionality can be obtained from the photoelectric effect, i.e., I PD =R·P=R·|E| 2 , where P is the optical power incident on the detector and R is the responsivity of the O / E converter 214. Responsivity can be defined as the detector output electrical signal (typically a current) generated in response to a given incident radiation power falling on the detector. An optical multiplier 200 can be provided to implement the phase shift (Δφ) applied to the input optical signal 204 by the PS 226 and the photocurrent (I PD), which can be used to multiply the elements of the optical signal 216 by the first vector 206 using the encoded intensity received by the O / E converter 214.

[0034] There are many ways to achieve Δφ and I PD For example, one or more O / E converters 214 are electrically connected to one or more linear conversion devices 234a-n, which are configured to convert the photocurrent I generated by the one or more O / E converters 214 PD Converted to voltage bias V PS . Voltage bias (V PS ) may be applied to the corresponding PS 226, and the phase shift (Δφ) of the corresponding PS may be proportional to the electric field (E) of the applied optical signal (eg, the input optical signal 204). PS As described above, PS226 can utilize the Pockels effect and include non-centrosymmetric materials such as, but not limited to, LiNbO3, LiTaO3, BBO, KTP, and the like.

[0035] In the illustrative example of optical multiplier 200, one or more linear transformation devices 234a-n can be implemented as one or more load resistors 234a-n (collectively referred to herein as load resistors 234), which can help provide linear transformation. Each load resistor 234 can be operated to convert the photocurrent I generated by each corresponding O / E converter 214 PD Converted to a voltage bias (V PS ), the phase shift (Δφ) corresponding to the PS can be related to the electric field (E ) of the applied optical signal (eg, the input optical signal 204) PS ). That is, for example, the phase shift (Δφ) of PS 226 may be proportional to the electric field applied to input optical signal 204 by corresponding PS 226, which is proportional to the voltage bias (V generated by O / E converter 214 and applied to PS 226. PS ) is proportional to the voltage bias (V PS ) and resistance (r load ) multiplied by the photocurrent (I PD ), which is proportional to the absolute value of the electric field incident on the O / E converter 214 (eg, Δφ∝E PS ∝V PS ∝R load I PD ∝|E| 2Thus, for each of the one or more optical modulator 222 and O / E converter 214 pairs, parameters of various properties of the one or more optical modulators 222 and O / E converter 214 may be selected to provide a linear transformation accordingly.

[0036] In an example embodiment, the optical multiplier 200 may provide the ability to weight each vector 206 and 218. For example, by using an avalanche photodiode (APD) for the O / E converter 214, additional avalanche gain may be achieved for the photocurrent generated by each O / E converter 214. The additional avalanche gain may be used to weight each vector on the optical modulator 222. Thus, weights may be applied to different vectors by adjusting the avalanche gain. For example, a first weight may be applied by adjusting the avalanche gain of the O / E converter 214a, and a second weight may be applied by adjusting the avalanche gain of the O / E converter 214b. If the first weight is greater than the second weight, the intensity modulation of the input optical signal 204 by the optical modulator 222a will be greater than the modulation applied by the optical modulator 222b.

[0037] Figure 4A An example schematic circuit 400 of an O / E converter driven optical modulator according to an example implementation of the presently disclosed technology is depicted.

[0038] Schematic circuit 400 is an example simulation of an implementation of an O / E converter 214 electrically connected to one or more optical modulators 222. For example, the encoded signal 402 (e.g., vector 218) is provided to a first oscilloscope 404 (e.g., an oscilloscope) and a Simulink to physical signal (PS) converter (S PS) 406 that converts the Simulink input signal to a physical signal. S PS 406 outputs a signal that simulates optical signal 216, which is provided to circuit 414, which is an example circuit of a photodiode (e.g., O / E converter 214). Circuit 414 includes a current source 408, which is connected in series with a resistor R series Connected in series with capacitor C PD and the shunt resistor R shunt The output of current source 408 is connected to a load resistor R load The negative terminal of a load resistor 234 (eg, a load resistor 234) which is also connected to C PD and R shunt The positive terminal of Figure 4A As shown. load The negative terminal of the voltage sensor 410 is also connected to the positive terminal of the capacitor C PS Connect in parallel. R load and C PSThe positive terminal of the voltage sensor 410 is connected to a solver configuration block (f(x)) 418 and grounded along with the negative terminal of the voltage sensor 410. The solver configuration block specifies the solver parameters that the model requires before simulation can begin. The voltage sensor 410 outputs a voltage to a PS to Simulink converter (PS S) 416 that is connected to a second oscilloscope 412 (e.g., an oscilloscope). The capacitor C is used to generate a voltage. PS To simulate PS226.

[0039] Figure 4B The diagram shows Figure 4A 4 is a graphical representation of an example input signal current and an example voltage bias generated by the schematic circuit 400 in FIG. More particularly, Figure 4B An input photocurrent signal 420 plotted as IPD monitored by a first oscilloscope 404 is depicted versus time, representing the encoded signal 402 provided to the circuit 414. In this example, the encoded signal 402 may be provided as a vector with element values ​​of 3, 0, 1, 4, 2, etc. Figure 4B Also depicted is the voltage across C detected by the voltage sensor 410 monitored by the second oscilloscope 412. PS The voltage bias signal 430 describes the voltage applied to Figure 2 The voltage bias of the PS226.

[0040] In operation, the input photocurrent detected by circuit 414 is encoded in signal 402. The photocurrent is monitored by first oscilloscope 404 providing input photocurrent signal 420. Current source 408 generates a current through R load and C PS The output photocurrent across C PS The voltage of is monitored by a second oscilloscope 412 providing a voltage bias signal 430. In the example simulation, Figure 4B shows a 50Gbit / s downspan C PS The voltage of C PD 10fF, R shunt 200MΩ, R series is 50Ω, R load is 50Ω, and C PS is 10fF. Figure 4B As shown, the input photocurrent IPD is proportional to the cross-C PS Thus, the information encoded into the input signal incident on circuit 414 can be converted into a voltage bias applied to C PS voltage bias.

[0041] Therefore, reference Figure 2, information encoded on the optical signal 216 incident on the O / E converter 214 can be converted into a voltage bias generated by the O / E converter 214, which can be applied to the PS 226 to drive the optical modulator 222. In addition, Figure 4A and Figure 4B It is shown that the optical multiplier 200 can support very high data rates (eg, 50 Gbit / s), which is much higher than the data rates achievable using corresponding electronic systems.

[0042] As described above, examples of the presently disclosed technology can utilize an over-coupled RAMZI in an optical modulator 222 to provide intensity modulation of an input optical signal 204. In this case, the O / E converter 214 encodes an electrical signal (e.g., a voltage bias) based on an optical signal 216 incident thereon, which is applied to a corresponding PS 226 provided on a microring 230 over-coupled to an MZI 232. The PS 226 applies a phase shift to the input optical signal 204, which is evanescently over-coupled from the MZI 232 into the microring 230, by modulating the electric field applied thereto based on modulating the voltage bias applied thereto. Modulating the phase shift imparted by the PS 226 causes an intensity modulation at the output of the MZI 232, which is proportional to and corresponds to the intensity modulation of the optical signal 216 incident on the O / E converter 214. Since the optical signal 216 is encoded with the vector 218 , the resulting optical signal output from the MZI 232 is encoded with an output vector that is the element-wise product of the vector 218 and the first vector 206 encoded on the input optical signal 204 .

[0043] Figure 5A Depicted is a schematic diagram of an overcoupled RAMZI 500 that may be used in examples of the presently disclosed technology.

[0044] The overcoupled RAMZI 500 may include an MZI 510 evanescently coupled to a microring 520. The overcoupled RAMZI 500 includes an input terminal 512 that receives an input electric field E i The input optical signal 502 propagates at a wavelength λ. The optical signal 502 may be encoded according to a vector. The overcoupled RAMZI 500 includes an output terminal 518 that transmits an output electric field E oThe output optical signal 506 is shown in FIG. 5 . The MZI 510 receives an input optical signal 502 from an input end, and the optical signal is split into a first branch 516 and a second branch 514 according to a splitting ratio. The MZI 510 may be implemented as one or more waveguides that guide the propagation of the input optical signal 502. For example, the first branch 516 may be formed by a first waveguide, and the second branch 514 may be formed by a second waveguide. The microring 520 is optically coupled to one of the branches (e.g., the first branch 516 in this example), while the other branch includes a delay line 515. The delay line 515 may be provided as a length difference between the second branch 514 and the first branch 516, for example by increasing the length of the second branch 514 relative to the first branch 514 by ΔL. MZI .

[0045] The examples disclosed herein may include an optical splitter disposed at an input end 512 for coupling a portion of an input optical signal 502 into each branch 514 and 516. The optical splitter may be implemented as a Y-type optical splitter for splitting the input optical signal 502 into branches 514 and 516 according to the splitting ratio of the optical splitter. For example, the optical splitter may be operable to split the input optical signal 502 into a first optical signal propagating in a first branch 516 and a second optical signal propagating in a second branch 514. The relative optical powers of the first signal and the second signal depend on the splitting ratio of the optical splitter. In the illustrative example provided herein, the optical splitter may be a 50:50 Y-type optical splitter with a splitting ratio of 1, which means that the optical powers of the first signal and the second signal are approximately the same (e.g., 50% of the input optical signal 502 is directed into branch 514 and 50% is directed into branch 516). However, for a given application, any splitting ratio may be used as desired. An optical ground coupler may be provided at an output end 518, which may be similar to but opposite to the output end of the optical splitter. That is, for example, the optical coupler may be operable to combine the optical signals propagating in the second branch 514 and the first branch 516 onto the output end 518 as the output optical signal 506 .

[0046] The optical signal on the first branch 516 is optically coupled to the microring 520, for example, by evanescent coupling based on the coupling coefficient between the waveguide of the first branch 516 and the waveguide of the microring 520. According to various embodiments disclosed herein, the waveguide forming the MZI 510 and the waveguide forming the microring 520 are made of the same material (e.g., silicon). As described above, the microring 520 can be strongly coupled to the first branch 516, so that the coupling coefficient between the microring 520 and the first branch 516 is high (e.g., at least greater than 0.89).

[0047] In the microring 520, the phase shifter 522 can be controlled to tune the resonant frequency, thereby causing a phase shift applied to the optical signal resonating in the microring 520. As described above, the voltage bias (V ps ) is applied to the phase shifter 522, thereby modulating the electric field (E ps ). This change in the electric field then modulates the phase of the optical signal resonating in the microring 520.

[0048] The MZI 510 then converts the phase shift in the optical signal from the microring 520 into an intensity variation. For example, the optical signals propagating in the two branches 514 and 516 may combine at the output 518, producing constructive and / or destructive interference, resulting in intensity variations of the output optical signal 506. In the case where the over-coupled RAMZI 500 is implemented as one or more optical modulators 222 driven by a voltage bias from the O / E converter 214, the intensity variation of the output optical signal 506 over time may be decoded as elements of an output vector (e.g., output vector 224) that is the product of the optical signal 216 multiplied by the input optical signal 204.

[0049] For more details, refer to Figure 5A , the phase shifter 522 is coupled to the micro-ring 520, and the round-trip phase of the micro-ring 520 may be provided as:

[0050]

[0051] Where n is the refractive index of the waveguide forming the microring 520, Δn is the change in effective refractive index due to the phase shift caused by the phase shifter 522 (e.g., the phase change caused by tuning the phase shifter 522), and L r is the circumference of the phase shifter 522, and λ is the wavelength of the input optical signal 502. In addition, due to the difference in branch lengths (e.g., delay line 515) (excluding the circumference of the microring 520), there may be a mπ+π / 2 phase difference between the two branches 514 and 516. Therefore, the phase difference between the two branches 514 and 516 can be provided as:

[0052]

[0053] Where, ΔL MZI is the branch length difference between the lengths of the first branch and the second branch, and m is 0, 1, 2, 3, etc.

[0054] Thus, the output electric field of the output optical signal 506 of the over-coupled RAMZI 500 may be provided as:

[0055]

[0056] Among them, Ei is the input electric field, a is the round trip field transmission of the microring 520, and t is the field transmission of the coupling between the microring 520 and the first branch 516. For a lossless coupler, t 2 +κ 2 =1, where κ is the field coupling coefficient. For an illustrative strongly overcoupled microring 520, a may be 0.99 and κ may be between 0.89 and 0.94, and in an implementation example, κ may be 0.92.

[0057] Figure 5B The intensity of the output optical signal 506 (which may be provided as |E o | 2 ) changes with the phase shift Δφ. Figure 5B As shown, the linearity of the overcoupled RAMZI 500 can be obtained by linear regression (e.g. Figure 5B 530 in FIG. Figure 5B As shown, over-coupled RAMZI 500 provides a linear transformation (e.g., proportional) between the phase shift Δφ and the intensity of output optical signal 506. In addition, over-coupled RAMZI 500 exhibits a very small standard deviation σ of about 0.007, which corresponds to 7-bit accuracy. Due to the high accuracy, linearized over-coupled RAMZI 500 can be used to encode analog signals (e.g., from O / E converter 214) without any additional control circuitry.

[0058] Although an over-coupled RAMZI is described herein as being used for an optical modulator, the presently disclosed technology is not limited thereto. The optical modulator disclosed herein may be implemented using other technologies and devices, such as, but not limited to, ring resonators and conventional MZIs.

[0059] Figure 6 An example computing component that can be used to implement element-by-element optical multiplication according to various embodiments of the disclosed technology is illustrated. Figure 6 , the computing component 600 can be, for example, a server computer, a controller, or any other similar computing component capable of processing data. Figure 6 In the example implementation of FIG. 6 , computing component 600 includes a hardware processor 602 and a machine-readable storage medium 604 .

[0060] Hardware processor 602 may be one or more central processing units (CPUs), semiconductor-based microprocessors, and / or other hardware devices suitable for fetching and executing instructions stored in machine-readable storage medium 604. Hardware processor 602 may fetch, decode, and execute instructions such as instructions 606-612 to control a process or operation for element-by-element multiplication. As an alternative or in addition to fetching and executing instructions, hardware processor 602 may include one or more electronic circuits including electronic components for performing the functions of one or more instructions, such as a field programmable gate array (FPGA), an application specific integrated circuit (ASIC), or other electronic circuits.

[0061] A machine-readable storage medium, such as machine-readable storage medium 604, may be any electronic, magnetic, optical, or other physical storage device that contains or stores executable instructions. Thus, machine-readable storage medium 604 may be, for example, a random access memory (RAM), a non-volatile RAM (NVRAM), an electrically erasable programmable read-only memory (EEPROM), a storage device, an optical disk, or the like. In some embodiments, machine-readable storage medium 604 may be a non-transitory storage medium, wherein the term "non-transitory" does not encompass transient propagation signals. As described in detail below, machine-readable storage medium 604 may be encoded with executable instructions (e.g., instructions 606-612).

[0062] The hardware processor 602 can execute instructions 606 to encode the plurality of vectors into an input optical signal and one or more optical signals. Figure 1 , Figure 2 and Figure 5A As described, multiple vectors can be encoded into various optical signals for performing element-by-element multiplication.

[0063] The hardware processor 602 can execute instructions 508 to input the input optical signal into an input waveguide optically coupled to one or more optical modulators. Figure 1 and Figure 2 As described above, the input optical signal is injected into the input waveguide. Figure 1 and Figure 2 As described, the input waveguide may include one or more optical modulators optically coupled thereto.

[0064] The hardware processor 602 can execute instructions 610 to modulate the intensity of the input optical signal on the input waveguide by driving one or more optical modulators based on receiving one or more optical signals on one or more optical-to-electrical (O / E) converters. Figure 1 , Figure 2 as well as FIG. 4A to FIG. 5BAs described, one or more photocurrents may be generated based on one or more optical signals received at one or more O / E converters. One or more photocurrents may be converted into one or more voltage biases, which may be used to drive one or more optical modulators. One or more O / E converters may be electrically coupled to one or more linear transformation devices that convert one or more photocurrents into one or more voltage biases. One or more voltage biases may be applied to one or more phase shifters coupled to one or more optical modulators, and one or more optical modulators may modulate the intensity of the input optical signal based on the phase shift applied to the input optical signal by the one or more phase shifters. The input optical signal may be coupled to the one or more phase shifters via one or more microrings coupled to one or more Mach-Zehnder interferometers.

[0065] The hardware processor 602 can execute instructions 612 to output an output optical signal encoded with an output vector, wherein the output vector is an element-by-element product of multiple vectors. For example, as described above, the intensity of the input optical signal is modulated at instruction 610 to encode the element-by-element product of multiple vectors onto the input optical signal so that the output optical signal is encoded with the product.

[0066] Figure 7 A block diagram of an example computer system 700 is depicted in which various embodiments described herein may be implemented. The computer system 700 includes a bus 702 or other communication mechanism for transmitting information, and one or more hardware processors 704 coupled to the bus 702 for processing information. The hardware processor(s) 704 may be, for example, one or more general-purpose microprocessors. The computer system 700 may be implemented as Figure 6 The computing component 600 in.

[0067] The computer system 700 also includes a main memory 706, such as a random access memory (RAM), cache, and / or other dynamic storage device, coupled to the bus 702 for storing information and instructions to be executed by the processor 704. The main memory 706 may also be used to store temporary variables or other intermediate information during the execution of instructions to be executed by the processor 704. Such instructions, when stored in a storage medium accessible to the processor 704, make the computer system 700 a special-purpose machine customized to perform the operations specified in the instructions.

[0068] The computer system 700 further includes a read only memory (ROM) 708 or other static storage device coupled to the bus 702 for storing static information and instructions for the processor 704. A storage device 710, such as a magnetic disk, optical disk, or USB thumb drive (flash drive), is provided and coupled to the bus 702 for storing information and instructions.

[0069] The computer system 700 may be coupled to a display 712, such as a liquid crystal display (LCD) (or touch screen), via the bus 702 for displaying information to a computer user. An input device 714, including alphanumeric and other keys, is coupled to the bus 702 for communicating information and command selections to the processor 704. Another type of user input device is a cursor control 716, such as a mouse, trackball, or cursor direction keys, for communicating direction information and command selections to the processor 704 and for controlling cursor movement on the display 712. In some embodiments, the same direction information and command selections as a cursor control may be implemented via receiving touches on a touch screen without a cursor.

[0070] The computing system 700 may include a user interface module for implementing a GUI, which may be stored in a mass storage device as executable software code executed by the computing device(s). By way of example, this module and other modules may include components such as software components, object-oriented software components, class components, and task components, processes, functions, properties, programs, subroutines, program code segments, drivers, firmware, microcode, circuits, data, databases, data structures, tables, arrays, and variables.

[0071] Generally, the words "component", "engine", "system", "database", "data store", etc. as used herein may refer to logic implemented in hardware or firmware, or to a set of software instructions written in a programming language such as Java, C or C++, which may have entry points and exit points. Software components can be compiled and linked into executable programs, installed in dynamic link libraries, or written in interpreted programming languages ​​such as BASIC, Perl or Python. It should be understood that software components can be called from other components or from themselves, and / or can be called in response to detected events or interrupts. Software components configured for execution on a computing device can be provided on a computer-readable medium such as an optical disk, a digital video disk, a flash drive, a disk, or any other tangible medium, or can be provided as a digital download (and can be stored originally in a compressed format or an installable format that needs to be installed, decompressed or decrypted before execution). Such software code can be stored in part or in whole on a memory device of an executing computing device for execution by the computing device. Software instructions can be embedded in firmware such as an EPROM. It will be further understood that hardware components may include connected logic units such as gates and flip-flops, and / or may include programmable units such as programmable gate arrays or processors.

[0072] The computer system 700 may implement the techniques described herein using custom hard-wired logic, one or more ASICs or FPGAs, firmware, and / or program logic that is combined with the computer system to make the computer system 700 a special purpose machine or to program it as a special purpose machine. According to one embodiment, the techniques herein are performed by the computer system 700 in response to the processor(s) 704 executing one or more sequences of one or more instructions contained in the main memory 706. Such instructions may be read into the main memory 706 from another storage medium, such as the storage device 710. Execution of the sequences of instructions contained in the main memory 706 causes the processor(s) 704 to perform the process steps described herein. In alternative embodiments, hard-wired circuitry may be used in place of or in combination with software instructions.

[0073] As used herein, the term "non-transitory medium" and similar terms refer to any medium that stores data and / or instructions that cause a machine to operate in a particular manner. Such non-transitory media may include non-volatile media and / or volatile media. Non-volatile media include, for example, optical or magnetic disks, such as storage device 710. Volatile media include dynamic memory, such as main memory 706. Common forms of non-transitory media include, for example, floppy disks, floppy disks, hard disks, solid-state drives, tapes, or any other magnetic data storage media, CD-ROMs, any other optical data storage media, any physical media with a pattern of holes, RAM, PROMs and EPROMs, flash EPROMs, NVRAMs, any other memory chips or boxes, and networked versions of these media.

[0074] Non-transient media are distinct from transmission media but can be used in conjunction with transmission media. Transmission media participate in the transfer of information between non-transient media. For example, transmission media include coaxial cables, copper wire, and optical fiber, including the wires that comprise bus 702. Transmission media can also take the form of acoustic or light waves, such as those generated during radio wave and infrared data communications.

[0075] Computer system 700 also includes a communication interface 718 coupled to bus 702. Network interface 718 provides a two-way data communication coupled to one or more network links, and the one or more network links are connected to one or more local networks. For example, communication interface 718 can be an integrated services digital network (ISDN) card, a cable modem, a satellite modem, or a modem for providing a data communication connection with a corresponding type of telephone line. As another example, network interface 718 can be a local area network (LAN) card for providing a data communication connection with a compatible LAN (or a WAN component for communicating with a WAN). Wireless links can also be implemented. In any such implementation, network interface 718 sends and receives an electrical signal, an electromagnetic signal, or an optical signal carrying a digital data stream representing various types of information.

[0076] A network link typically provides data communication through one or more networks to other data devices. For example, a network link can provide a connection through a local network to a host computer or to data devices operated by an Internet Service Provider (ISP). The ISP, in turn, provides data communication services through a global packet data communication network now commonly referred to as the "Internet". Both local networks and the Internet use electrical, electromagnetic, or optical signals that carry digital data streams. Signals through various networks, as well as signals on network links and signals through communication interface 718 (which carry digital data to and from the computer system 700) are example forms of transmission media.

[0077] Computer system 700 can send messages and receive data, including program code, through the network(s), network links and communications interface 718. In the Internet example, a server can transmit the requested code for an application program through the Internet, an ISP, a local network and communications interface 718.

[0078] The received code may be executed by processor 704 as it is received, and / or stored in storage device 710 or other non-volatile storage for later execution.

[0079] Each of the processes, methods and algorithms described in the previous sections can be implemented in a code component executed by one or more computer systems or computer processors including computer hardware and automated in whole or in part by these code components. The one or more computer systems or computer processors can also operate to support the execution of related operations in a "cloud computing" environment, or operate as "software as a service (SaaS)". These processes and algorithms can be implemented in part or in whole in a dedicated circuit. The various features and processes described above can be used independently of each other, or can be combined in various ways. Different combinations and sub-combinations are intended to fall within the scope of the present disclosure, and some method boxes or process boxes can be omitted in some embodiments. The methods and processes described herein are also not limited to any particular order, and the boxes or states related to these methods and processes can be executed in other appropriate orders, or can be executed in parallel, or in some other way. Boxes or states can be added to or removed from the disclosed example embodiments. The execution of certain operations or processes can be distributed to computer systems or computer processors so that they do not reside only in a single machine, but are deployed across multiple machines.

[0080] As used herein, circuits can be implemented using any form of hardware, software or a combination thereof. For example, one or more processors, controllers, ASIC, PLA, PAL, CPLD, FPGA, logic components, software routines or other mechanisms can be implemented to form circuits. In an embodiment, various circuits described herein can be implemented as discrete circuits, or the functions and features described can be shared partially or entirely between one or more circuits. Although various features or functional elements can be described or advocated individually as separate circuits, these features and functions can also be shared among one or more common circuits, and this description should not require or imply that separate circuits are needed to implement such features or functions. In the case of using software to implement circuits in whole or in part, such software can be implemented to operate together with a computing system or processing system (such as computer system 700) that can perform the functions described about the software.

[0081] As used herein, the term "or" may be interpreted in an inclusive or exclusive sense. In addition, descriptions of resources, operations, or structures in the singular should not be understood to exclude the plural. Unless explicitly stated otherwise, or otherwise understood within the context as used, conditional language (such as, among others, "can," "could," "might," or "may") is generally intended to convey that certain embodiments include and other embodiments do not include certain features, elements, and / or steps.

[0082] Unless otherwise expressly stated, the terms and phrases used in this document and their variations should be interpreted as open-ended rather than restrictive. Adjectives such as "conventional", "traditional", "normal", "standard", "known" and terms of similar meaning should not be interpreted as limiting the items described to items available for a given time period or before a given time, but should be understood to cover traditional, conventional, normal or standard technologies that may be available or known at any time now or in the future. In some cases, the presence of broad words and phrases (such as "one or more", "at least", "but not limited to" or other similar phrases) should not be understood to mean that narrower situations are intended or required where such broad phrases may not be present.

Claims

1. An optical device, comprising: a waveguide for receiving an input optical signal encoded with a first vector and outputting an output optical signal; one or more optical-to-electrical (O / E) converters configured to receive one or more optical signals encoded with one or more vectors and generate one or more electrical signals based on the received one or more optical signals; one or more optical modulators optically coupled to the waveguide and electrically coupled to the one or more O / E converters, the one or more optical modulators modulating the intensity of the input optical signal on the waveguide based on the one or more electrical signals; and Wherein, the output optical signal is encoded by the product of the first vector and the one or more vectors.

2. The optical device according to claim 1, wherein: The one or more O / E converters are photodetectors.

3. The optical device according to claim 1, wherein: Each of the one or more optical modulators comprises: a phase shifter electrically coupled to a corresponding O / E converter; and A Mach-Zehnder interferometer (MZI) is optically coupled to the waveguide and the phase shifter.

4. The optical device of claim 3, further comprising: One or more linear transformation devices, the one or more linear transformation devices are electrically connected to the one or more O / E converters and the one or more optical modulators and are located between the one or more O / E converters and the one or more optical modulators, wherein the one or more linear transformation devices are configured to convert photocurrents generated by the one or more O / E converters into the one or more electrical signals, and wherein the phase shifter is configured to apply a phase change proportional to the one or more electrical signals.

5. The optical device of claim 4, wherein: The one or more linear transformation devices are one or more resistors.

6. The optical device according to claim 3, wherein: The MZI comprises a micro-ring assisted MZI (RAMZI), wherein the phase shifter is disposed on a micro-ring of the RAMZI.

7. The optical device of claim 6, wherein: The micro-ring is over-coupled with the MZI.

8. The optical device of claim 7, wherein: The coupling coefficient between the micro-ring and the MZI is between 0.89 and 0.

94.

9. The optical device of claim 3, wherein: The MZI includes a first branch, a second branch, and a branch length difference between lengths of the first branch and the second branch.

10. The optical device of claim 1, wherein: The one or more O / E converters include at least: a first O / E converter configured to receive a first optical signal encoded with a second vector, and a second O / E converter, the second O / E converter being configured to receive a second optical signal encoded with a third vector, The output optical signal is encoded by an element-by-element product of the first vector, the second vector and the third vector.

11. The optical device of claim 9, wherein: The one or more optical modulators include at least: a first optical modulator optically coupled to the waveguide and electrically coupled to the first O / E converter, and A second optical modulator is optically coupled to the waveguide and electrically coupled to the second O / E converter.

12. A method for element-by-element vector multiplication, the method comprising: encoding the plurality of vectors into an input optical signal and one or more optical signals; inputting the input optical signal into an input waveguide optically coupled to one or more optical modulators; modulating the intensity of the input optical signal on the input waveguide by driving the one or more optical modulators based on receiving the one or more optical signals on one or more optical-to-electrical (O / E) converters; as well as An output optical signal encoded with an output vector is output based on the modulated intensity, the output vector being an element-wise product of the plurality of vectors.

13. The method of claim 11, further comprising: generating one or more photocurrents based on the one or more optical signals received at the one or more O / E converters; converting the one or more photocurrents into one or more electrical signals; as well as The one or more optical modulators are driven based on the one or more electrical signals.

14. The method of claim 12, wherein: The one or more O / E converters are electrically coupled to one or more linear transformation devices that convert the one or more photocurrents into one or more electrical signals.

15. The method of claim 12, wherein: Driving the one or more optical modulators based on the one or more electrical signals comprises: applying the one or more electrical signals to one or more phase shifters coupled to the one or more optical modulators; Wherein the one or more optical modulators modulate the intensity of the input optical signal based on one or more phase shifts applied to the input optical signal by the one or more phase shifters, wherein the input optical signal is optically coupled to the one or more phase shifters.

16. The method of claim 14, wherein: The one or more optical modulators include one or more micro-rings coupled to one or more Mach-Zehnder interferometers, wherein the one or more phase shifters are coupled to the one or more micro-rings.

17. The method of claim 14, wherein: The one or more phase shifts are proportional to the one or more electrical signals, the one or more photocurrents, and the modulated intensity.

18. An element-by-element multi-vector optical multiplier, comprising: a waveguide configured to receive an input optical signal encoded with a first vector and output an output optical signal, the waveguide comprising a microring optically coupled to a Mach-Zehnder interferometer (MZI); a phase shifter coupled to the microring; as well as a photodiode electrically coupled to the phase shifter via a linear transformation device, the photodiode being configured to receive an optical signal encoded with a second vector, Wherein, the output optical signal is encoded by the element-by-element product of the first vector and the second vector.

19. The multi-vector optical multiplier of claim 18, wherein: The linear conversion device is a resistor electrically coupled to the phase shifter and the photodiode.

20. The multi-vector optical multiplier of claim 18, wherein: The coupling coefficient between the micro-ring and the MZI is between 0.89 and 0.94.

Images