Always-on keyword detector
Provided herein is an integrated circuit including, in some embodiments, a special-purpose host processor, a neuromorphic co-processor, and a communications interface between the host processor and the co-processor configured to transmit information therebetween. The special-purpose host processor is operable as a stand-alone host processor. The neuromorphic co-processor includes an artificial neural network. The co-processor is configured to enhance special-purpose processing of the host processor through the artificial neural network. In such embodiments, the host processor is a keyword identifier processor configured to transmit one or more detected words to the co-processor over the communications interface. The co-processor is configured to transmit recognized words, or other sounds, to the host processor.
1. An integrated circuit for detecting keywords or sound, comprising:
a host processor configured to receive a data stream and compute frequency component signals for the data stream;
a co-processor configured to identify one or more desired keywords or sound within the computed frequency component signals;
wherein at least one of the host processor and co-processor is configured to remain in a low-power, always-on state that is continuously ready to detect the desired keyword or sound;
the co-processor comprising an enhancement network with cloud-updateable synaptic weights; and
an interface between the host processor and the co-processor configured to transmit information therebetween.
2. The integrated circuit of claim 1 , wherein the co-processor is a neuromorphic processor.
3. The integrated circuit of claim 2 , wherein the neuromorphic processor includes an artificial neural network.
4. The integrated circuit of claim 1 , wherein the co-processor includes a database comprised of known keywords whereby the one or more desired keywords may be utilized by the co-processor.
5. The integrated circuit of claim 4 , wherein additional keywords are added to the database.
6. The integrated circuit of claim 1 , wherein the integrated circuit is configured to operate on current leakage from a power source.
7. A method for detecting keywords within a data stream, comprising:
receiving, by a host processor, the data stream in the form of electrical signals;
computing frequency component signals for the data stream;
transmitting, to a co-processor, the computed frequency component signals;
remaining in a low-power, always-on state that is continuously ready to detect one or more desired keywords;
identifying the one or more desired keywords within the transmitted frequency component signals;
providing an enhancement network with cloud-updateable synaptic weights; and
transmitting, to the host processor, the one or more desired keywords.
8. The method of claim 7 , wherein the co-processor includes a database of known keywords utilized to identify electrical signals corresponding to one or more keywords.
9. The method of claim 7 , wherein the identifying one or more desired keywords is performed by an artificial neural network.
10. The method of claim 7 , wherein the one or more keywords are comprised of a plurality of predefined acoustic signals other than speech.
11. An integrated circuit, comprising:
a host processor;
a co-processor including an artificial neural network configured to compute frequency component signals for an audio stream,
wherein one or more desired keywords may be detected within the audio stream, the audio stream received in the form of electrical signals to the host processor,
wherein the co-processor is configured to remain in a low-power, always-on state that is continuously ready to detect the one or more desired keywords;
the co-processor comprising an enhancement network with cloud-updateable synaptic weights; and
an interface between the host processor and the co-processor configured to transmit information therebetween.
12. The integrated circuit of claim 11 , wherein the host processor is a pattern signal identifier processor configured to transmit the audio stream in the form of electrical signals to the co-processor over the interface, and wherein the co-processor is configured to enhance the processing of the host processor by providing one or more detected patterns within the pattern signal to the host processor over the interface.
13. The integrated circuit of claim 11 , wherein the interface between the host processor and the co-processor is a serial peripheral interface (“SPI”) bus or inter-integrated circuit (“I2C”) bus.
14. The integrated circuit of claim 11 , wherein the artificial neural network is disposed in an analog multiplier array of a number of two-quadrant multipliers in a memory sector of the integrated circuit.
15. The integrated circuit of claim 11 , wherein synaptic weights of the artificial neural network are stored in firmware of the integrated circuit.
16. The integrated circuit of claim 15 , wherein the firmware is configured to receive updates to the synaptic weights of the artificial neural network.
17. The integrated circuit of claim 11 , wherein the integrated circuit is configured to operate on battery power.
18. The integrated circuit of claim 11 , wherein the integrated circuit is configured to remain in a low-power, always-on state whereby the integrated circuit remains continuously ready to receive the audio stream.
19. The integrated circuit of claim 11 , wherein the host processor is configured to generate an output signal upon detection of one or more desired keywords.
20. The integrated circuit of claim 11 , wherein the integrated circuit is configured to remain in a fully aware state to receive the audio stream.