GB2626540A

GB2626540A - Optical apparatus for the generation of multiple adjusted signals

Info

Publication number: GB2626540A
Application number: GB2301019.2A
Authority: GB
Inventors: Feldmann Johannes; Brückerhoff-Plückelmann Frank; Bente Ivonne; Wendland Daniel; Pernice Wolfram
Original assignee: Westfaelische Wilhelms Universitaet Muenster; Salience Labs Ltd
Current assignee: Westfaelische Wilhelms Universitaet Muenster; Salience Labs Ltd
Priority date: 2023-01-24
Filing date: 2023-01-24
Publication date: 2024-07-31
Also published as: EP4655659A1; WO2024156986A1; CN120752598A; GB202301019D0

Abstract

An optical apparatus 605 for multiply-accumulate operations in a matrix multiplication unit comprises a temporally incoherent light source 610 illuminating amplitude adjusters 641 to produce adjusted optical signals. A splitter 620 may split the temporally incoherent light into secondary signals. Each secondary signal may have a delay introduced by traveling a different distance. These distances may differ by at least the temporally incoherent light’s coherence length. The secondary signals may be demultiplexed 681 before the amplitude adjusters and multiplexed 681′ again afterwards. The temporally incoherent light source may be amplified spontaneous emission, a thermal source, a solid state source, or a white light source. The amplitude adjusters may be optical amplifiers or optical modulators, such as electro-optic modulators, phase change modulators, acousto-optic modulators, or mechanical modulators. Sets of amplitude adjusters may have their own corresponding temporally incoherent light source. Components may be coupled by integrated waveguides 631 or optical fibres, or the apparatus may be a free beam arrangement. The optical apparatus may be part of an integrated optical chip alongside a processor 650, such as an optical multiplication matrix, forming an optical computing system 600. The processor may add adjusted optical signals at the same wavelength while preventing interference.

Description

OPTICAL APPARATUS FOR THE GENERATION OF MULTIPLE ADJUSTED

SIGNALS

Technical Field

The present disclosure relates to an optical apparatus and corresponding method for generating a plurality of adjusted signals. In particular, the present disclosure relates to the generation of amplitude-adjusted signals based on a temporally incoherent light source. The present disclosure also relates to an optical computing system comprising such an optical apparatus.

Background

Optical processors rely on the effective amplitude adjustment (modulation or amplification) of optical signals. The modulation of an optical signal can be used to perform a multiplication operation of the optical signal by a predetermined coefficient. In turn the accumulation of modulated signals can be used to perform multiply-accumulate (MAC) operation.

Accumulation is carried out by superimposing signals over time. During the temporal overlay, interference effects may occur that affect the result of the computing operation and limit the ability to parallelize via wavelength multiplexing. It is an object of the disclosure to address one or more of the above mentioned limitations.

Summary

According to a first aspect of the disclosure, there is provided an optical apparatus for generating a plurality of adjusted optical signals, the optical apparatus comprising at least one temporally incoherent light source coupled to a plurality of amplitude adjusters, each amplitude adjuster being configured to attenuate or amplify an optical signal derived from the said at least one temporally incoherent light source to provide a corresponding adjusted optical signal.

Optionally, the said at least one temporally incoherent light source is adapted to provide a primary signal of temporally incoherent radiation, the optical apparatus further comprising a splitting device adapted to split the primary signal into a plurality of secondary signals.

For instance, the temporally incoherent light source may be a single source.

Optionally, wherein the splitting device is coupled to the plurality of amplitude adjusters via a plurality of channels for transmitting the secondary signals.

For instance, the channels may be waveguides such as integrated waveguides, or optical fibres.

Optionally, each channel has a different length for introducing a delay between different secondary signals.

Optionally, wherein the primary signal of temporally incoherent radiation has a coherence length and wherein the path length travelled by the secondary signals in each channel differ by at least the coherence length.

Optionally, wherein each channel is coupled to a demultiplexer and complementary multiplexer, wherein the demultiplexer is configured to demultiplex a spectrum of the secondary signal into a plurality of individual wavelengths channels.

Optionally, each individual wavelength channel comprises a corresponding amplitude adjuster, and wherein the multiplexer is configured to receive an adjusted signal from each amplitude adjuster.

Optionally, each amplitude adjuster is coupled to a corresponding temporally incoherent light source, or wherein the plurality of amplitude adjusters form sets of amplitude adjusters and wherein each set is coupled to a corresponding temporally incoherent light source.

Optionally, the said at least one temporally incoherent light source has a coherence time of less than about 10 nanoseconds or less than about 1 nanosecond, or less than about 100 picoseconds.

Optionally, the said at least one temporally incoherent light source comprises one or more of an amplified spontaneous emission source, a thermal source, a solid state light source, and a white light source.

For instance the solid state light source may be a light-emitting diode LED.

Optionally, the plurality of amplitude adjusters comprises optical modulators and/or optical amplifiers.

For instance the optical modulators may be electro-optic modulators (E0M5), phase change modulators (PCMs], acousto-optical modulators, polymer based modulators, thermal modulators or mechanical modulators.

According to a second aspect of the disclosure, there is provided an optical computing system comprising an optical apparatus according to the first aspect coupled to a processing device having a plurality of inputs for receiving the adjusted optical signals.

Optionally, the processing device comprises a combiner adapted to combine the adjusted optical signals.

Optionally, wherein the combiner is coupled to the plurality of amplitude adjusters via a set of optical channels, wherein each optical channel in the set has a same length.

Optionally, wherein the processing device comprises an optical multiplication matrix.

Optionally, the processing device is configured to perform addition of adjusted optical signals at a same wavelength while preventing interference between the adjusted optical signals.

According to a third aspect of the disclosure there is provided an integrated optical chip comprising an optical apparatus according to the first aspect or an optical computing system according to the second aspect.

The options described with respect to the first aspect of the disclosure are also common to the second and third aspects of the disclosure.

According to a fourth aspect of the disclosure there is provided a method of generating a plurality of adjusted signals, the method comprising providing at least one temporally incoherent light source coupled to a plurality of amplitude adjusters, and adjusting the amplitude of a signal derived from the said at least one temporally incoherent light source to provide a corresponding adjusted 25 signal.

Optionally the method further comprises adding adjusted optical signals at a same wavelength while preventing interference between the adjusted optical signals.

The fourth aspect may share features of the first and second aspects, as noted above and herein.

Description of the drawings

The disclosure is described in further detail below by way of example and with reference to the accompanying drawings, in which: figure 1 is a is a diagram of an optical system provided with a matrix multiplication unit; figure 2 is a flow chart of a method for generating a plurality of adjusted signals according to the disclosure; figure 3 is a diagram of a system comprising an optical apparatus for generating a plurality of modulated signals according to the method of figure 2; figure 4 is a diagram of an alternative implementation of the system figure 3; figure 5 is a diagram of an extended version of the system of figure 3; figure 6 is a diagram of another system comprising an optical apparatus for generating a plurality of modulated signals according to the method of figure 2; figure 7 is a schematic diagram of a multiplication matrix.

Description

Figure 1 is a diagram of an optical system provided with a matrix multiplication unit. The optical system 100 includes a matrix multiplication unit 110 coupled to a light source unit 120 and a sensor unit 130. The matrix multiplication unit 110 has a number N of input waveguides 112 (rows) coupled to the light source unit 120 and a number M of output waveguides 113 (columns) coupled to the sensor unit 130. It will be appreciated that the numbers N and M may vary. The system 100 uses a number L of channels and therefore L vectors are sent in parallel to the matrix multiplication unit 110.

The input and output waveguides 112, 113 are arranged to form a grid of multiplication unit cells 111. For each unit cell, input and output waveguides 112, 113 cross one another at a crossing point C. A coupler 114 is interposed between the input waveguide 112 and the associated output waveguide 113.The coupler 114 is provided with an optical modulator 115 The optical modulator 115 may be an electro-optic modulator EOM or a phase change modulator PCM.

The input waveguide 112 and the coupler 114 act as two directional couplers in the unit cell, with fixed transmission. The input waveguide 112 splits light from the input row so that part of the light is transmitted along 112 to the next cell and part of the light is sent to the modulator 115 to modulate /attenuate the light. The coupler 114 then adds the modulated light to the output waveguide 113.

The light source unit 120 has N subunits 121 adapted to generate an input beam for a corresponding input waveguide 112. Similarly, the detector unit has M subunits 131 each one having L detectors for receiving the output of the L wavelength channels from a corresponding output waveguide 113. Each subunit 121 includes L lasers coupled to a multiplexer MUX. The output of the multiplexer is coupled to a corresponding input waveguide 112. Each laser provides an output beam having a unique wavelength. A modulator 125 is provided at the output of each laser to modulate the output beam.

Each subunit 121 is therefore designed to generate an input beam having its own unique set of wavelengths so that every input waveguide in the matrix has its own unique set of wavelengths. This approach permits to avoid optical interference when combining multiple beams of light at a single waveguide. However, such a system has a limited scalability. For instance, in order to feed a 64 x 64 matrix unit with L vectors, L*64 different light sources (or laser lines) are required. To avoid beating and interference, each line must be spectrally separated, and a large optical bandwidth is required. For instance the first MUX gets range [Al to AL] and the second MUX gets [Al +A to AL+ A Al where the A A. is smaller than the difference between two neighbouring wavelengths (for instance aA <(A2-Al)). Assuming a 200 pm spacing between wavelengths, this would require an optical bandwidth of 100 nm.

Figure 2 is a flow chart of a method for generating a plurality of adjusted optical signals according to the disclosure. At step 210 at least one temporally incoherent light source is coupled to a plurality of amplitude adjusters. The amplitude adjusters may be optical modulators for attenuating an amplitude of an optical signal, or optical amplifiers for amplifying an amplitude of an optical signal or a combination of both. For instance combining optical modulators and optical amplifiers may be used to increase the dynamic range of the system.

The temporally incoherent light source is adapted to generate a beam of electromagnetic radiation. The beam of electromagnetic radiation may have a spectrum in a specific spectral region, which may be an infrared region such as near infrared N1R. The spectrum may also include the visible or near visible region.

At step 220 an amplitude of a signal derived from the said at least one temporally incoherent light source is adjusted to provide a corresponding adjusted signal.

A laser is a source of coherent light in which there is a fixed phase relationship between the electric field values at different locations (spatial coherence) and/or at different times (temporal coherence). Spatial coherence arises from the existence of resonator modes which define spatially correlated field patterns, providing beams with high spatial coherence. Lasers emitting multiple modes with different optical frequencies have a relatively low temporal coherence. When a laser operates with a single frequency and some stability features, it can have a high temporal coherence. Temporal coherence can be expressed by the coherence time which defines the time over which the field correlation decays, and the coherence length which defines the propagation distance over which the coherence significantly decays.

The source of temporally incoherent light may be implemented in different ways. A source of light has a low temporal coherence, and may be regarded as temporally incoherent, when its coherence time is relatively short. For instance a temporally incoherent light source may have a coherence time in the picosecond range and/or a coherence length in the mm range. For ID example the coherence time may be less than 100ps for instance between about lps-30 ps, depending on the width of the wavelengths channel.

The temporally incoherent light source and its coherence time may be selected depending on the speed of the system. For instance one could envisage running a processing element (for example a multiplication matrix) at relatively low speed (kHz or MHz instead of 10 GHz) to build a low power and lower throughput device. In this case a temporally incoherent light source with a relatively longer coherence time could be used, for instance a coherence time of less than about 10 ns or in the region of ins.

Various types of sources can be used to generate a radiation with low temporal coherence, including amplified spontaneous emission ASE sources, thermal light sources, solid state light sources such as LEDs or white-light sources. Other sources that do not include a resonator and emit light randomly could also be envisaged.

In a laser gain medium, the luminescence or fluorescence arising from spontaneous emission can be amplified to obtain an amplified spontaneous emission (ASE). The ASE emitted beam may cover a relatively broad range of wavelengths, for instance in the IR region the ASE may have a spectral width ranging from about 30 nm to about 100nm or more. This amplified radiation has a low temporal coherence but a relatively good spatial coherence. A good spatial coherence permits efficient optical coupling of light, for instance from the ASE source to a waveguide.

The temporally incoherent light source may also be a white-light source designed to generate a white-light spectrum. The white-light spectrum may have a width of several hundred of nanometres. Depending on the application, this approach can permit to use a single light source to provide all the wavelengths necessary for a given computational task.

Figure 3 is a diagram of a system comprising an optical apparatus for generating a plurality of adjusted signals according to the method of figure 2. The system 300 includes an optical apparatus 305 for generating a plurality of adjusted signals coupled to a detector 360 via a combiner 350. The optical apparatus 305 includes a single source of temporally incoherent light 310, in this example an ASE source, and a splitting device or splitter 320 coupled to a set of amplitude adjusters 341, 342 via two optical channels also referred to as waveguides 331, 332. The waveguides 331 and 332 have different lengths for introducing a delay between a signal transmitted through the waveguide 331 and a signal transmitted through the waveguide 332. The combiner 350 is coupled to the amplitude adjusters 341 and 342 via a pair of waveguides 371, 372 each having a same length. The splitter 320 and the combiner 350 may be implemented in different ways. For instance a directional coupler could be used such as a directional fibre coupler or an integrated photonic directional coupler. A first directional coupler may be used for the splitter, in one direction, and another directional coupler may be used for the combiner, in the opposite direction. In this example the amplitude adjusters 341 and 342 are provided by optical modulators. In another embodiment the amplitude adjusters may be optical amplifiers.

In operation, the temporally incoherent light source 310 generates a primary signal having an intensity lo. The splitter 320 divides the primary signal into two secondary signals. The splitter 320 may be designed to achieve different splitting ratios depending on the application. In this example, the splitter 320 provides a 50:50 splitting ratio, so that each secondary signal has an intensity of lo/ 2. The two secondary signals are transmitted to the modulators 341 and 342 via the waveguides 331 and 332, respectively. The waveguide 332 is longer than the waveguide 331. The difference in length is designed such that the path length travelled by the secondary signals differ by at least the coherence length of the source of temporally incoherent light 310. Therefore, the delay distances are chosen to be longer than the coherence time of the light source.

The modulators 341 and 342 are configured to modulate the signal intensity of the secondary signals by a predetermined coefficient. For instance the modulation may perform an attenuation of the signal with an attenuation coefficient (scalar factors "a" for 341 and "b" for 342). Each modulator can be controlled to adjust the attenuation coefficient which may range between 0 and 1. This results in a multiplication operation of the light intensity of the secondary signal. In this example the modulator 341 provides a modulated signal having an intensity of lo/ 2 *a, and the modulator 342 provides a modulated signal having an of lo/ 2 *b. When using optical amplifiers instead of optical modulators, the attenuation coefficient would be replaced by an amplification coefficient.

After multiplication in the modulators, the modulated signals are transmitted to the combiner 350 that combines or superimposes the modulated signals. The combiner 350 provides an output signal having an intensity equal to the sum of the intensities of the modulated signals. In this example the output signal has an intensity equal to lo/ 2 *a + Io/ 2 *b. By using a source of temporally incoherent light, multiple modulated signals having the same wavelength can be combined or overlapped without interference. The presence of interferences between signals having the same wavelength would corrupt the MAC operation as one could not assume that the signal intensities simply add up.

To ensure that the modulated signals do not experience a time offset, the path length of the waveguides 371 and 372 between the modulators and the combiner are designed to have a same length, that is a same optical path.

Since both signals are superimposed in time, the detector 360 can detect the accumulated intensity of the signals and the measurement signal is thus the result of the multiply-accumulate (MAC) operation.

Figure 4 is a diagram of an alternative implementation of the system figure 3. The system 400 is similar to the system 300 of figure 3 and the same components are represented with the same reference numerals. In this case, the splitter 320 has been removed and two temporally incoherent sources are used 310 and 310'. The waveguides 431, 432 do not need to introduce a delay, and therefore can have a same length.

For some applications the delay lines can become relatively long (depending on the coherence length and how many inputs are used) which increases the chip area and the cost of production. Such delay lines may also introduce optical losses that need to be accounted for. The system of figure 4 removes the need for delay lines. In addition, using individual sources also permits the use of more optical power, hence increasing the signal to noise ratio SNR.

Figure 5 is a diagram of an extended version of the system of figure 3. The system of figure 3 can be extended to perform the summation of any number of multiplication operations. For this purpose, the splitter 520 that divides the optical signal from the temporally incoherent light source 510 (primary signal) into the sub-signals (secondary signals) must be selected or designed in such a way that it emits the desired number N of secondary signals. These secondary signals are then forwarded to a corresponding number N of modulators labelled 541-54N via N channels or waveguides 531-53N.

The waveguide 531-5314 have different lengths of increasing values so that 532 is longer than 531, 533 is longer than 532 etc... The path difference between each one of these waveguides is designed to be at least equal to the coherence length of the temporally incoherent source 510. The combiner 550 is coupled to the modulators 541-54N via N waveguides 571-57N each having a same length (same transmission distances). The output of the combiner 550 is coupled to the detector 560. Optionally, additional processing elements (not shown) can be used after the combiner 550 and before the detector 560.

The operation of the system SOO is similar to the operation of the system 300. The combiner 550 superimposes the modulated signals received from modulators 541-54N and provides an output signal having an intensity equal to the sum of the intensities of the modulated signals. The output of the combiner 550 is then detected by the detector 560. It will be appreciated that the system of figure 5 could also be implemented using multiple sources following the example of figure 4.

Figure 6 is a diagram of another system comprising an optical apparatus for generating a plurality of modulated signals according to the method of figure 2. The system 600 includes an optical apparatus 605 for generating a plurality of modulated signals coupled to a processing device 650. The optical apparatus 605 includes a single temporally incoherent light source 610, in this example an ASE source, and a splitter 620 for splitting the primary signal from 610 into N secondary signals via N outputs. N pairs of complementary wavelength demultiplexer/multiplexer (681/681'- 68N/68N') are provided to perform wavelength division multiplexing.

The N pairs of demultiplexer/multiplexer are used to separate L individual wavelength channels and enable modulation of the vectors onto these channels. Each wavelength channel, among a total of L channels, carries one vector. The vector is provided as an input to the processing device 650. Each vector has N entries and is encoded via N modulators (with one wavelength per vector). For L channels, the system carries L vectors. Each vector is composed of its N entries (different input rows to the matrix). The N input rows have their own modulators, so that a single vector has N modulators, corresponding to its values. Looking at Figure 6: Modulator 64L and 64N get signals on the same wavelength but modulate the first and the last entry of the vector.

Each output of the splitter 620 is coupled to a corresponding demultiplexer (6 8 1-6 8N) via an optical waveguide or channel (631-63N). Like in previous embodiments the N waveguides 631-63N have different lengths for introducing a delay between signals transmitted through the different waveguides.

The temporally incoherent light source 610 provides a primary optical signal having a source spectrum. This source spectrum may be relatively broad. Each demultiplexer 681-68N is configured to receive a secondary signal having the source spectrum and to split the source spectrum into sub wavelength ranges to produce a plurality L of tertiary signals each having a specific wavelength range. This is achieved in parallel.

A plurality L of modulators (641-64L) is provided between each pair of demultiplexer/multiplexer (681-681') for modulating the L tertiary signals. The multiplexer 681' is then used to combine the L modulated signals and to transmit them in parallel to an input of the processing device 650. Since the L modulated signals arise from the same waveguide 631, they have the same delay.

Wavelength division multiplexing allows different signals encoded at different wavelengths to be routed simultaneously through the same channels without them interacting with each other.

The outputs of the optical apparatus 605 provide L*N signals in N channels. These signals can now be used as input signals for the processing device 650. The processing device 650 may be implemented in various fashions. For instance the processing device 650 may be an optical multiplication NxM matrix processor such as the matrix multiplication unit as described in figure 1.

The processing device 650 has a plurality of inputs and a plurality of outputs. Each input of the processing device 650 can receive the output of a corresponding multiplexer (681'-68N'). Each output of the processing device 650 is coupled to an output demultiplexer 691 coupled to a plurality of detectors 661-66M. Each output signal of the processing device 650 contains several wavelengths. Each output demultiplexer 691-69M is used to separate the wavelengths of a corresponding output signal of the processing device 650. The detectors 661-66M are used to detect the accumulated intensities at each individual wavelength. The detectors 661-66M may include L individual sub photodetectors.

The system 600 is therefore adapted to perform wavelength division multiplexing which enables a further increase in the number of MAC (multiply-accumulate) operations that can be carried out simultaneously. Using a source of temporally incoherent light with a broadband spectrum increases parallelisation and permits to perform matrix multiplications while avoiding optical interferences.

In the system of figure 6 multiple beams or signals can be added together at the same wavelength without signal degradation through interference. This reduces the number of wavelengths needed to one per vector. Considering a light source having a spectrum with a width of 100 nm, and a channel width of 200 pm, then 500 vectors could be sent in parallel.

Like in figure 4 multiple temporally incoherent light sources could be used instead of a single one. In this case one source could be used per matrix input. For instance, each DEMUX 681-68N could be provided with its own light source. This would permit to avoid difficult routing and waveguide crossings on the photonic chip and reduce the footprint of the chip as no optical delay lines would be needed. The coherence lengths of the multiple light sources may be selected depending on the speed of operation and path lengths on the chip.

Figure 7 is a schematic diagram of a multiplication matrix unit that may be used in the system of figure 6. The multiplication matrix unit 700 has four inputs waveguides 701-704. Each input waveguide is coupled to four different outputs (out1, out2, out3, out4) via four coupling waveguides. For example, the input waveguide 701 is coupled to the outputs 1-4 via the coupling waveguides 711, 712, 713, 714, respectively. Similarly, the input waveguide 702 is coupled to the outputs 1-4 via the coupling waveguides 721-724, respectively, etc... Each coupling waveguide is provided with a modulator labelled M1-M4 for modulating an optical signal transmitted through the waveguide.

The coupling between an input waveguide and a coupling waveguide is provided by an input coupler. For instance the input waveguide 701 has four input couplers Cla, Clb, C1c and C1d for coupling to the waveguides 711, 712, 713 and 714, respectively.

Each coupling waveguide is coupled to a dedicated output waveguide via an output coupler Cout. For instance the coupling waveguide 711 is coupled to the output waveguide 791 of output 1 (out1) via the output coupler Cout1.

The input couplers are used to split an input optical signal received at an input waveguide so that a portion of the signal is transmitted through the input waveguide while the other remaining portion is passed to the coupling waveguide.

The splitting ratio may be adjusted such that each column waveguide gets the same amount of optical signal. Assuming no losses, this would mean, that the first coupler C1a couples 1/4 of the input optical signal to the 711 waveguide, the second coupler C1b couples 1/3 of the remaining optical signal to 712, the third coupler C1c couples 1/2 of the remaining optical signal to 713 and the fourth coupler C1d couples all (1/1) of the remaining optical signal to the 714 waveguide.

The splitting ratios of the output couplers may be chosen to obtain an equal contribution from the inputs channels to the output signal provided at the output. For instance the splitting ratios of the output couplers Cout1, Cout2, Cout3 and Cout4, coupled to the first output channel 791 may be 1/4, 1/3, 1/2 and 1/1, respectively.

Depending on the device implementation, further adjustments may be required to consider optical losses at the crossing points between input channels and coupling channels as well as optical losses associated with the couplers. Overall ratios are adjusted so that each input contributes the same amount of light to each output.

The output couplers are used to combine a modulated signal transmitted through a coupling waveguide with other modulated signals transmitted through other coupling waveguides. The input waveguides may be made of silicon nitride, while the coupling waveguides may be made of silicon or both from the same material. It will be appreciated that the design of the multiplication matrix 700 may be extended to any number of inputs and outputs.

Compared with the multiplication matrix of figure 1, the matrix 700, adds flexibility and reduces footprint. Its design permits to use longer modulators which lead to better signal to noise ratio, as well as different types of modulators. For example Mach Zehnder modulators (MZMs) may be used instead of electro-absorption modulators (EAMs). Bigger matrix sizes can also be achieved, as the unit cell is smaller compared to the crossbar array, and the modulators can be more densely packed.

The proposed optical systems as described with reference to figures 3 to 7 permit to accumulate several optical signals and increase parallelisation.

This is achieved with a compact design. The systems of figures 3, 5 and 6 can be implemented using a single temporally incoherent light source or multiple temporally incoherent light sources.

It will be appreciated that the systems described with respect to figures 3 to 7 may be implemented in different ways. The systems of figures 3 to 7 may be implemented using an integrated optical circuit such as a photonic integrated circuit (PIC). The short coherence length of the light source (for instance in the mm range) permits to implement the system as an integrated optical circuit. In the case of an integrated chip implementation, the modulators may be electro optic modulators, for example, E0Ms based on Indium Phosphide InP, silicon germanium or lithium niobate. Other types of integrated modulators could also be envisaged such as optically controlled modulators, phase change modulators PCMs, and polymer based modulators to name a few. The light source or sources may be fibre coupled ASE source(s). The ASE could also be designed as part of the chip, for instance it could be implemented in Indium Phosphide.

Alternatively, the channels or waveguides used in the various embodiments of figures 3 to 7 may be implemented as optical fibres that can be optically coupled to various elements, including the light source, the splitter, the modulators, the multiplexers/ demultiplexers etc...

The systems of figures 3 to 7 may also be implemented in a free beam arrangement. The various components of the system may be selected depending on the type of implementation. For a fiber-coupled implementation or a free-beam implementation the modulators may be electro-optic modulators (E0Ms), phase change modulators (PCMs), acoustooptical modulators, spatial light modulators, or mechanical modulators, depending on what clock rates must be achieved. Although the amplitude adjusters have been described as optical modulators in the various embodiment, it will be appreciated that the optical modulators could be replaced by optical amplifiers.

A skilled person will therefore appreciate that variations of the disclosed arrangements are possible without departing from the disclosure.

Accordingly, the above description of the specific embodiments is made by way of example only and not for the purposes of limitation. It will be clear to the skilled person that minor modifications may be made without significant changes to the operation described.

Claims

CLAIMS1. An optical apparatus for generating a plurality of adjusted optical signals, the optical apparatus comprising at least one temporally incoherent light source coupled to a plurality of amplitude adjusters, each amplitude adjuster being configured to attenuate or amplify an optical signal derived from the said at least one temporally incoherent light source to provide a corresponding adjusted optical signal.
2. The optical apparatus as claimed in claim 1, wherein the said at least one temporally incoherent light source is adapted to provide a primary signal of temporally incoherent radiation, the optical apparatus further comprising a splitting device adapted to split the primary signal into a plurality of secondary signals.
3. The optical apparatus as claimed in claim 2, wherein the splitting device is coupled to the plurality of amplitude adjusters via a plurality of channels for transmitting the secondary signals.
4. The optical apparatus as claimed in claim 3, wherein each channel has a different length for introducing a delay between different secondary signals.
S. The optical apparatus as claimed in claim 4, wherein the primary signal of temporally incoherent radiation has a coherence length and wherein the path length travelled by the secondary signals in each channel differ by at least the coherence length.
6. The optical apparatus as claimed in claim 3, wherein each channel is coupled to a demultiplexer and complementary multiplexer, wherein the demultiplexer is configured to demultiplex a spectrum of the secondary signal into a plurality of individual wavelengths channels.
7. The optical apparatus as claimed in claim 6, wherein each individual wavelength channel comprises a corresponding amplitude adjuster, and wherein the multiplexer is configured to receive an adjusted signal from each amplitude adjuster.
8. The optical apparatus as claimed in claim 1, wherein each amplitude adjuster is coupled to a corresponding temporally incoherent light source, or wherein the plurality of amplitude adjusters form sets of amplitude adjusters and wherein each set is coupled to a corresponding temporally incoherent light source.
9. The optical apparatus as claimed in any of the preceding claims, wherein the said at least one temporally incoherent light source has a coherence time of less than about 10 nanoseconds or less than about 1 nanosecond, or less than about 100 picoseconds.
10. The optical apparatus as claimed in any of the preceding claims, wherein the said at least one temporally incoherent light source comprises one or more of an amplified spontaneous emission source, a thermal source, a solid state light source, and a white light source.
11. The optical apparatus as claimed in any of the preceding claims, wherein the plurality of amplitude adjusters comprises optical modulators and/or optical amplifiers.
12. An optical computing system comprising an optical apparatus as claimed in any of the preceding claims coupled to a processing device having a plurality of inputs for receiving the adjusted optical signals.
13. The optical computing system as claimed in claim 12, wherein the processing device comprises a combiner adapted to combine the adjusted optical signals.
14. The optical computing system as claimed in claim 13, wherein the combiner is coupled to the plurality of amplitude adjusters via a ID set of optical channels, wherein each optical channel in the set has a same length.
15. The optical computing system as claimed in claim 12, wherein the processing device comprises an optical multiplication matrix.
16. The optical computing system as claimed in any of the claims 12 to 15, wherein the processing device is configured to perform addition of adjusted optical signals at a same wavelength while preventing interference between the adjusted optical signals.
17. An integrated optical chip comprising an optical apparatus as claimed in any of the claims 1 to 11 or an optical computing system as claimed in any of the claims 12 to 16.
18. A method of generating a plurality of adjusted signals, the method comprising providing at least one temporally incoherent light source coupled to a plurality of amplitude adjusters, and adjusting the amplitude of a signal derived from the said at least one temporally incoherent light source to provide a corresponding adjusted signal.
19. The method as claimed in claim 17 further comprising adding adjusted optical signals at a same wavelength while preventing interference between the adjusted optical signals.