Techniques for personalization model transfer

1 . A method comprising:

during a plurality of beaconing sessions, collecting, with a network device positioned at a plurality of different locations during the plurality of beaconing sessions, information indicative of a plurality of patterns of wireless signals between the network device and a plurality of playback devices;

with the network device, training a first parameterized machine learning model to produce a trained model that identifies one or more playback devices proximal to the network device based on one or more features derived from the information;

transferring the trained model from the network device to a portable playback device;

during a subsequent beaconing session, collecting, with the portable playback device, data indicative of a pattern of wireless signals between the portable playback device and the plurality of playback devices;

with the portable playback device, applying the trained model to at least one feature derived from the data to identify at least one playback device of the plurality of playback devices that is proximal to the portable playback device; and

with the portable playback device, communicating at least one request to the at least one playback device.

2 . The method of claim 1 , further comprising:

during the plurality of beaconing sessions, collecting, with the network device, one or more reporting signals containing information indicative of patterns of beacon signals between each of the plurality of playback devices and the network device.

3 . The method of claim 1 , wherein applying the trained model comprises:

using the trained model to determine whether that the at least one playback device is a bonding target for the portable playback device; and

providing a recommendation to form a bonded group comprising the portable playback device and the at least one playback device.

4 . The method of claim 3 , further comprising:

detecting user feedback regarding the recommendation; and

retraining the trained model based on the user feedback.

5 . The method of claim 1 , further comprising:

using the trained model to predict that the at least one playback device is a bonding target for the portable playback device based; and

automatically forming a bonded group comprising the portable playback device and the at least one playback device.

6 . The method of claim 1 , wherein collecting the data indicative of a pattern of wireless signals between the portable playback device and the plurality of playback devices comprises collecting received signal strength indicator (RSSI) values of the wireless signals and a count of the wireless signals.

7 . A portable playback device comprising:

a wireless communication interface configured to support communication of data via at least one network protocol;

at least one processor; and

at least one non-transitory computer readable medium storing program instructions that are executable by the at least one processor to cause the portable playback device to

initiate a beaconing session based on determining that a position of the portable playback device is unknown,

detect, via the wireless communication interface and during the beaconing session, information indicative of a pattern of wireless signals between the portable playback device and a plurality of other playback devices,

detect, via the wireless communication interface, information from an external device correlating a baseline pattern of wireless signals with a location,

compare the pattern of wireless signals with the baseline pattern, and

based on a level of similarity between the pattern of wireless signals and the baseline pattern transgressing a predetermined threshold value, infer that the position of the portable playback device corresponds to the location.

8 . The portable playback device of claim 7 , wherein the baseline pattern is determined by a network device by, during one or more beaconing sessions, collecting, with the network device positioned at a plurality of locations during the one or more beaconing sessions, information indicative of a plurality of patterns of wireless signals between the network device and the plurality of other playback devices, each pattern of the plurality of patterns of wireless signals corresponding to a respective location of the plurality of locations of the network device.

9 . The portable playback device of claim 7 , wherein the at least one non-transitory computer readable medium further stores program instructions that are executable by the at least one processor to cause the portable playback device to:

based on inferring that the position of the portable playback device corresponds to the location, apply one or more personalization settings selected from a plurality of personalization settings.

10 . The portable playback device of claim 9 , wherein the plurality of personalization settings includes volume settings, a power management schedule, and a grouping status of the portable playback device with one or more other playback devices of the plurality of other playback devices.

11 . The portable playback device of claim 7 , wherein the at least one non-transitory computer readable medium further stores program instructions that are executable by the at least one processor to cause the portable playback device to:

based on inferring that the position of the portable playback device corresponds to the location, identify a proposed target playback device for receiving an instruction from the portable playback device.

12 . The portable playback device of claim 11 , wherein the instruction is one of (i) an instruction to form a bonded group comprising the portable playback device and the proposed target playback device, or (ii) an instruction to begin playback of audio content.

13 . The portable playback device of claim 7 , wherein to detect the information indicative of the pattern of wireless signals, the at least one non-transitory computer readable medium stores program instructions that are executable by the at least one processor to cause the portable playback device to:

detect a plurality of beacon signals emitted by the plurality of other playback devices.

14 . The portable playback device of claim 13 , wherein to detect the information indicative of the pattern of wireless signals, the at least one non-transitory computer readable medium further stores program instructions that are executable by the at least one processor to cause the portable playback device to:

determine, for each beacon signal, a received signal strength indicator (RSSI) value and a standard deviation of a signal strength of the detected beacon signal relative to a median signal strength of the plurality of beacon signals; and

determine a first count of the plurality of beacon signals detected during a predetermined collection time period.

15 . The portable playback device of claim 14 , wherein the at least one tangible non-transitory computer readable medium further stores program instructions that are executable by the at least one processor to cause the portable playback device to:

transmit, via the wireless communication interface, a beacon signal.

16 . The portable playback device of claim 14 , wherein each of the plurality of beacon signals includes a sequence of tones and a transmission identifier; and

wherein at least one of the sequence of tones or the transmission identifier identifies a playback device that is a source of the respective beacon signal.

17 . A portable playback device comprising:

a wireless communication interface configured to support communication of data via at least one network protocol;

at least one processor; and

at least one non-transitory computer readable medium storing program instructions that are executable by the at least one processor to cause the portable playback device to

detect, via the wireless communication interface, a plurality of beacon signals emitted by a plurality of other playback devices,

determine characteristics for each beacon signal of the plurality of beacon signals, the characteristics including a received signal strength indicator (RSSI) value and a standard deviation of a signal strength of the beacon signal relative to a median signal strength of the plurality of beacon signals,

determine a count of the plurality of beacon signals detected during a predetermined collection time period,

based on the characteristics of the plurality of beacon signals and the count, determine a pattern of wireless signals between the portable playback device and the plurality of other playback devices,

detect, via the wireless communication interface, information containing one or more location-based personalization settings,

based on the pattern of wireless signals, infer a location of the portable playback device, and

use the location of the portable playback device and at least one of the one or more location-based personalization settings to train a parameterized machine learning model to predict a setting of the portable playback device.

18 . The portable playback device of claim 17 , wherein the one or more location-based personalization settings include one or more of a volume setting, a power management schedule, or a grouping status; and

wherein the setting of the portable playback device includes one or more of the volume setting, the power-management schedule, or the grouping status.

19 . The portable playback device of claim 17 , wherein the parameterized machine learning model includes one of a logistic regression model or a Gaussian Process model.