darwin phones: the evolution of sensing and inference on mobile phones emiliano miluzzo *, cory t....

135
Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo * , Cory T. Cornelius * , Ashwin Ramaswamy * , Tanzeem Choudhury * , Zhigang Liu ** , Andrew T. Campbell * * CS Department – Dartmouth College ** Nokia Research Center – Palo Alto

Upload: austen-offord

Post on 11-Dec-2015

212 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

Darwin Phones: the Evolution of Sensing and Inference on

Mobile Phones

Emiliano Miluzzo*, Cory T. Cornelius*, Ashwin Ramaswamy*, Tanzeem Choudhury*, Zhigang Liu**,

Andrew T. Campbell*

* CS Department – Dartmouth College** Nokia Research Center – Palo Alto

Page 2: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

Page 3: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

evolution of sensing and inferenceon mobile phones

Page 4: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

Emiliano Miluzzo

PR time

[email protected]

Page 5: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

Page 6: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

Page 7: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

Page 8: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

Page 9: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

Page 10: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

ok… so what ??

[email protected] Miluzzo

Page 11: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

density

Page 12: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

accelerometer

digital compass

microphone

WiFi/bluetooth GPS

….

light sensor/camera

sensing

Page 13: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

accelerometer

digital compass

microphone

WiFi/bluetooth GPS

light sensor/camera

gyroscope

air quality /pollution sensor

sensing….

Page 14: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

- free SDK

- multitasking

programmability

Page 15: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

- 600 MHz CPU

- up to 1GB application memory

hardware

computation capability is increasing

Page 16: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

application distribution

Page 17: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

application distribution

deploy apps onto millions of phones at

the blink of an eye

Page 18: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

application distribution

collect huge amount of data for research

purposes

deploy apps onto millions of phones at

the blink of an eye

Page 19: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

cloud infrastructure

cloud - backend support

Page 20: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

cloud infrastructure

cloud - backend support

Page 21: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

cloud infrastructure

cloud - backend support

we want to push intelligence to the

phone

Page 22: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

cloud infrastructure

cloud - backend support

preserve the phone user experience

(battery lifetime, ability to make calls, etc.)

Page 23: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

cloud infrastructure

cloud - backend support

- sensing

- run machine learning algorithms locally

(feature extraction + inference)

Page 24: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

cloud infrastructure

cloud - backend support

- sensing

- run machine learning algorithms locally

(feature extraction + inference)

run machine learningalgorithms (learning)

Page 25: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

cloud infrastructure

cloud - backend support

store and crunch big data(fusion)

run machine learningalgorithms (learning)

- sensing

- run machine learning algorithms locally

(feature extraction + inference)

Page 26: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

cloud infrastructure

cloud - backend support

run machine learningalgorithms (learning)

store and crunch big data(fusion)

3 to 5 years from now our phones will be as powerful as a - sensing

- run machine learning algorithms locally

(feature extraction + inference)

Page 27: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

cloud infrastructure

cloud - backend support

run machine learningalgorithms (learning)

store and crunch big data(fusion)

3 to 5 years from now our phones will be as powerful as a - sensing

- run machine learning algorithms locally

(feature extraction + inference)

Page 28: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

cloud infrastructure

cloud - backend support

run machine learningalgorithms (learning)

store and crunch big data(fusion)

3 to 5 years from now our phones will be as powerful as a - sensing

- run machine learning algorithms locally

(feature extraction + inference)

Page 29: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

cloud infrastructure

cloud - backend support

- Sensing

- run machine learning algorithms locally

(feature extraction + learning + inference)

run machine learningalgorithms (learning)

store and crunch big data(fusion)

3 to 5 years from now our phones will be as powerful as a

Page 30: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

sensingprogrammability

cloud infrastructure

Page 31: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

sensingprogrammability

cloud infrastructure

??

Page 32: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

societal scale sensing

global mobile sensor network

reality mining using mobile phones

will play a big role in the future

Page 33: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

end of PR – now darwin

Emiliano Miluzzo [email protected]

Page 34: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

a small building block towards the big vision

Emiliano Miluzzo [email protected]

Page 35: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

from motes to mobile phones

Page 36: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

evolution of sensing and inferenceon mobile phones

from motes to mobile phones

Page 37: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

evolution of sensing and inferenceon mobile phones

from motes to mobile phones

darwin

- classification model evolution

- classification model pooling

- collaborative inference

Page 38: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

microphone

camera

GPS/WiFi/cellular

air quality pollution

sensing apps

social context

audio / pollution / RF fingerprinting

image / video manipulation

darwin applies distributed computing and collaborative inference concepts to

mobile sensing systems

darwin

- classification model evolution

- classification model pooling

- collaborative inference

Page 39: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

why darwin?

[email protected] Miluzzo

mobile phone sensing today

Page 40: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

why darwin?

[email protected] Miluzzo

train classification model X in the lab

mobile phone sensing today

Page 41: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

why darwin?

[email protected] Miluzzo

deploy classifier X

mobile phone sensing today

train classification model X in the lab

Page 42: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

why darwin?

[email protected] Miluzzo

train classification model X in the lab deploy classifier X

train classification model X’ in the lab

mobile phone sensing today

Page 43: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

why darwin?

[email protected] Miluzzo

deploy classifier X

deploy classifier X’

mobile phone sensing today

train classification model X’ in the lab

train classification model X in the lab

Page 44: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

why darwin?

[email protected] Miluzzo

train classification model X in the lab deploy classifier X

deploy classifier X’

a fully supervised approach doesn’t

scale!

mobile phone sensing today

train classification model X’ in the lab

Page 45: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

why darwin? a same classifier does not scale to multiple

environments (e.g., quiet and noisy env)

[email protected] Miluzzo

Page 46: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

why darwin? a same classifier does not scale to multiple

environments (e.g., quiet and noisy env)

[email protected] Miluzzo

Page 47: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

why darwin? a same classifier does not scale to multiple

environments (e.g., quiet and noisy env)

[email protected] Miluzzo

Page 48: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

why darwin? a same classifier does not scale to multiple

environments (e.g., quiet and noisy env)

[email protected] Miluzzo

darwin creates new classification models transparently from the user

(classification model evolution)

Page 49: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

why darwin?

ability for an application torapidly scale to many devices

Page 50: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

why darwin?

ability for an application torapidly scale to many devices

darwin re-uses classification models when possible

(classification model pooling)

Page 51: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

why darwin?

leverage the large ensemble of in-situ resources

Page 52: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

why darwin?

leverage the large ensemble of in-situ resources

darwin exploits spatial diversity and co-operate to alleviate the “sensing context”

problem(collaborative inference)

Page 53: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

darwin design

Page 54: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

speaker recognition (subject to audio noise, sensing context, etc.)

Page 55: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

darwin phases

Page 56: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

darwin phases

initial training (derive model seed)supervised

Page 57: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

darwin phases

initial training (derive model seed)

classification model evolution

supervised

unsupervised

Page 58: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

darwin phases

initial training (derive model seed)

classification model evolution

classification model pooling

supervised

unsupervised

Page 59: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

darwin phases

initial training (derive model seed)

classification model evolution

classification model pooling

collaborative inference

supervised

unsupervised

Page 60: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

classification model training

sensed event

Page 61: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

classification model training

sensed event

filtering (silence suppression +

voicing)

Page 62: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

classification model training

sensed event

filtering (silence suppression +

voicing)

featureextraction(MFCC)

Page 63: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

classification model training

filtering (silence suppression +

voicing)

featureextraction(MFCC)

modeltraining(GMM)

model

baseline

sensed event

send model + baseline back to phone

send MFCC tobackend to train the model

backend

Page 64: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

classification model training

phone: feature extraction(low

computation)

backend

backend: model training (high

computation)

Page 65: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

classification model evolution

phone: determines when to evolve

Page 66: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

classification model evolution

phone: determines when to evolve

training

Page 67: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

classification model evolution

phone: determines when to evolve

training sampled

Page 68: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

classification model evolution

phone: determines when to evolve

match?

YES

do not evolve

Page 69: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

classification model evolution

phone: determines when to evolve

match?

NO

evolve(train new model using

backend as before)

Page 70: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

classification model evolution

new speaker voice model

training

Page 71: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

classification model evolution

new speaker voice model

training

Page 72: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

classification model evolution

new speaker voice model

training

Page 73: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

classification model pooling

Page 74: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

classification model pooling

Speaker A’s model

Phone A Phone B

Phone C

Speaker C’s model

Speaker B’s modelSpeaker B’s model

Speaker C’s model

Page 75: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

classification model pooling

Speaker A’s model

Phone A Phone B

Phone C

Speaker C’s model

Speaker B’s modelSpeaker B’s model

Speaker C’s model

we have two options

1. train a new classifier for each speaker (costly for power, inference delay)

Page 76: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

classification model pooling

Speaker A’s model

Phone A Phone B

Phone C

Speaker C’s model

Speaker B’s modelSpeaker B’s model

Speaker C’s model

we have two options

1. train a new classifier for each speaker (costly for power, inference delay)

2. re-use already available classifiers

Page 77: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

classification model pooling

Speaker A’s model

Phone A Phone B

Phone C

Speaker C’s model

Speaker B’s modelSpeaker B’s model

Speaker C’s model

we have two options

1. train a new classifier for each speaker (costly for power, inference delay)

2. re-use already available classifiers

Page 78: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

classification model pooling

Speaker A’s model

Phone A Phone B

Phone C

Speaker C’s model

Speaker B’s modelSpeaker B’s model

Speaker C’s model

Page 79: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

classification model pooling

Speaker A’s model

Phone A Phone B

Phone C

Speaker B’s model

Speaker C’s model

Speaker C’s model

Speaker B’s model

Speaker A’s model

Speaker C’s model

Page 80: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

classification model pooling

Speaker A’s model

Phone A Phone B

Phone C

Speaker B’s model

Speaker C’s model

Speaker C’s model

Speaker B’s model

Speaker A’s model

Speaker C’s model

Speaker A’s model

Speaker B’s model

Page 81: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

classification model pooling

Speaker A’s model

Phone A Phone B

Phone C

Speaker B’s model

Speaker C’s model

Speaker C’s model

Speaker A’s model

Speaker B’s model

Speaker B’s model

Speaker A’s model

Speaker C’s model

Page 82: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

classification model pooling

Speaker A’s model

Phone A Phone B

Phone C

Speaker B’s model

Speaker C’s model

Speaker C’s model

Speaker A’s model

Speaker B’s model

Speaker B’s model

Speaker A’s model

Speaker C’s model

ready to run the collaborative inference algorithm

- local inference first- final inference later

Page 83: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

collaborative inference

two phases

Page 84: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

collaborative inference

1. local inference (running independently in parallel on each mobile phone)

two phases

Page 85: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

collaborative inference

1. local inference (running independently in parallel on each mobile phone)

two phases

2. final inference (after collecting Local Inference results, to get better confidence about the final classification result)

Page 86: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

local inference (LI)

[email protected] Miluzzo

collaborative inference

Phone A Phone B

Phone C

Page 87: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

collaborative inference

Phone A Phone B

Phone C

speaker A speaking!!!local inference (LI)

Page 88: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

collaborative inference

Phone A Phone B

Phone C

speaker A speaking!!!local inference (LI)

A’s LI results:Prob(A speaking) = 0.65Prob(B speaking) = 0.25Prob(C speaking) = 0.10

C’s LI results:Prob(A speaking) = 0.30Prob(B speaking) = 0.67Prob(C speaking) = 0.03

B’s LI results:Prob(A speaking) = 0.79Prob(B speaking) = 0.11Prob(C speaking) = 0.10

Page 89: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

collaborative inference

Phone A Phone B

Phone C

speaker A speaking!!!local inference (LI)

A’s LI results:Prob(A speaking) = 0.65Prob(B speaking) = 0.25Prob(C speaking) = 0.10

C’s LI results:Prob(A speaking) = 0.30Prob(B speaking) = 0.67Prob(C speaking) = 0.03

B’s LI results:Prob(A speaking) = 0.79Prob(B speaking) = 0.11Prob(C speaking) = 0.10

Page 90: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

collaborative inference

Phone A Phone B

Phone C

speaker A speaking!!!

A’s LI results:Prob(A speaking) = 0.65Prob(B speaking) = 0.25Prob(C speaking) = 0.10

local inference (LI)

C’s LI results:Prob(A speaking) = 0.30Prob(B speaking) = 0.67Prob(C speaking) = 0.03

B’s LI results:Prob(A speaking) = 0.79Prob(B speaking) = 0.11Prob(C speaking) = 0.10

Page 91: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

collaborative inference

Phone A Phone B

Phone C

speaker A speaking!!!

A’s LI results:Prob(A speaking) = 0.65Prob(B speaking) = 0.25Prob(C speaking) = 0.10

local inference (LI)

C’s LI results:Prob(A speaking) = 0.30Prob(B speaking) = 0.67Prob(C speaking) = 0.03

B’s LI results:Prob(A speaking) = 0.79Prob(B speaking) = 0.11Prob(C speaking) = 0.10

Page 92: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

collaborative inference

Phone A Phone B

Phone C

speaker A speaking!!!local inference (LI)

A’s LI results:Prob(A speaking) = 0.65Prob(B speaking) = 0.25Prob(C speaking) = 0.10

C’s LI results:Prob(A speaking) = 0.30Prob(B speaking) = 0.67Prob(C speaking) = 0.03

B’s LI results:Prob(A speaking) = 0.79Prob(B speaking) = 0.11Prob(C speaking) = 0.10

Page 93: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

collaborative inference

Phone A Phone B

Phone C

speaker A speaking!!!local inference (LI)

A’s LI results:Prob(A speaking) = 0.65Prob(B speaking) = 0.25Prob(C speaking) = 0.10

C’s LI results:Prob(A speaking) = 0.30Prob(B speaking) = 0.67Prob(C speaking) = 0.03

B’s LI results:Prob(A speaking) = 0.79Prob(B speaking) = 0.11Prob(C speaking) = 0.10

individual classification can be misleading!

Page 94: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

final inference (FI)

[email protected] Miluzzo

collaborative inference

Phone A Phone B

Phone C

each phone gathers LI results

A’s LI results

C’s LI results

B’s LI results

A’s LI results A’s LI results

C’s LI results C’s LI results

B’s LI resultsB’s LI results

Page 95: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

final inference (FI)

[email protected] Miluzzo

collaborative inferenceon each phone

A’s LI results:Prob(A speaking) = 0.65Prob(B speaking) = 0.25Prob(C speaking) = 0.10

C’s LI results:Prob(A speaking) = 0.30Prob(B speaking) = 0.67Prob(C speaking) = 0.03

B’s LI results:Prob(A speaking) = 0.79Prob(B speaking) = 0.11Prob(C speaking) = 0.10

Page 96: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

A’s LI results:Prob(A speaking) = 0.65Prob(B speaking) = 0.25Prob(C speaking) = 0.10

C’s LI results:Prob(A speaking) = 0.30Prob(B speaking) = 0.67Prob(C speaking) = 0.03

B’s LI results:Prob(A speaking) = 0.79Prob(B speaking) = 0.11Prob(C speaking) = 0.10

xxx

xxx

final inference (FI)

collaborative inferenceon each phone

Page 97: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

A’s LI results:Prob(A speaking) = 0.65Prob(B speaking) = 0.25Prob(C speaking) = 0.10

C’s LI results:Prob(A speaking) = 0.30Prob(B speaking) = 0.67Prob(C speaking) = 0.03

B’s LI results:Prob(A speaking) = 0.79Prob(B speaking) = 0.11Prob(C speaking) = 0.10

xxx

xxx

FI results (normalized):Confidence (A speaking) = 1 Confidence (B speaking) =

0.12Confidence (C speaking) =

0.002

=

final inference (FI)

collaborative inferenceon each phone

Page 98: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

A’s LI results:Prob(A speaking) = 0.65Prob(B speaking) = 0.25Prob(C speaking) = 0.10

C’s LI results:Prob(A speaking) = 0.30Prob(B speaking) = 0.67Prob(C speaking) = 0.03

B’s LI results:Prob(A speaking) = 0.79Prob(B speaking) = 0.11Prob(C speaking) = 0.10

xxx

xxx

=FI results (normalized):Confidence (A speaking) = 1 Confidence (B speaking) =

0.12Confidence (C speaking) =

0.002

final inference (FI)

collaborative inferenceon each phone

Page 99: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

A’s LI results:Prob(A speaking) = 0.65Prob(B speaking) = 0.25Prob(C speaking) = 0.10

C’s LI results:Prob(A speaking) = 0.30Prob(B speaking) = 0.67Prob(C speaking) = 0.03

B’s LI results:Prob(A speaking) = 0.79Prob(B speaking) = 0.11Prob(C speaking) = 0.10

xxx

xxx

=

collaborative inference compensates the inaccuracies of individual

inferences

FI results (normalized):Confidence (A speaking) = 1 Confidence (B speaking) =

0.12Confidence (C speaking) =

0.002

final inference (FI)

collaborative inferenceon each phone

Page 100: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

evaluation

Page 101: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

evaluation

C/C++ &

implemented on Nokia N97 andiPhone in support of a speaker

recognition app

Page 102: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

evaluation

C/C++ &

unix server

implemented on Nokia N97 andiPhone in support of a speaker

recognition app

Page 103: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

evaluation

C/C++ &

unix server

lightweight reliable protocol to transfer models from the server

and between phones

implemented on Nokia N97 andiPhone in support of a speaker

recognition app

Page 104: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

evaluation

C/C++ &

UDP multicast protocol to distribute

local inference results between phones

implemented on Nokia N97 andiPhone in support of a speaker

recognition app

Page 105: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

experimental scenarios

up to eight people in conversation in three different scenarios (quiet indoor, down the

street, in a restaurant)

Page 106: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

some numerical results

Page 107: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

need for evolution

train indoor, evaluate outdoor

Page 108: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

need for evolution

accuracy improvement after evolution

accuracy

Page 109: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

indoor quiet scenario

8 people talking around a table

Page 110: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

indoor quiet scenario

8 people talking around a table

Page 111: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

indoor quiet scenario

8 people talking around a table

Page 112: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

indoor quiet scenario

8 people talking around a table

Page 113: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

indoor quiet scenario

8 people talking around a table

Page 114: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

indoor quiet scenario

8 people talking around a table

Page 115: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

indoor quiet scenario

8 people talking around a table

collaborative inference + classification model evolution

boost the performance of a mobile sensing app

Page 116: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

impact of the number of mobile phones

Page 117: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

impact of the number of mobile phones

Page 118: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

impact of the number of mobile phones

Page 119: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

impact of the number of mobile phones

Page 120: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

impact of the number of mobile phones

the larger the number of mobile phones collaborating, the better the final inference result

Page 121: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

battery lifetime Vs inference responsiveness

Page 122: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

battery lifetime Vs inference responsiveness

Page 123: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

high responsiveness

battery lifetime Vs inference responsiveness

Page 124: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

short battery life

battery lifetime Vs inference responsiveness

Page 125: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

longer battery duration

battery lifetime Vs inference responsiveness

Page 126: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

low responsiveness

battery lifetime Vs inference responsiveness

Page 127: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

battery lifetime Vs inference responsiveness

smart duty-cycling techniques and machine learning algorithms with better performance in

terms of energy usage on mobile phones need to be identified

Page 128: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

a quick recap

smartphone’s are everywhere, let’s exploit their collective sensing and

computation capabilities

Page 129: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

a quick recapsmartphone’s are everywhere – let’s exploit

their collective sensing and computation capabilities

smartphone sensing opens up new frontiers: applications can be spread and

big data collected at unprecedented scale enabling endless research

opportunities

Page 130: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

a quick recapsmartphone’s are everywhere – let’s exploit

their collective sensing and computation capabilities

continuous sensing is still challenging; efficient mobile sensing requires to

preserve the phone user experience (need for energy efficient ML algorithms

and smart duty-cycling techniques)

smartphone sensing opens up new frontiers: applications can be spread and big data collected at

unprecedented scale enabling endless research opportunities

Page 131: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

a quick recapsmartphone’s are everywhere – let’s exploit

their collective sensing and computation capabilities

continuous sensing is still challenging; efficient mobile sensing requires to preserve the phone user

experience (need for energy efficient ML algorithms and smart duty-cycling techniques)

ML algorithms should perform reliably in the wild

smartphone sensing opens up new frontiers: applications can be spread and big data collected at

unprecedented scale enabling endless research opportunities

Page 132: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

a quick recapsmartphone’s are everywhere – let’s exploit

their collective sensing and computation capabilities

continuous sensing is still challenging; efficient mobile sensing requires to preserve the phone user

experience (need for energy efficient ML algorithms and smart duty-cycling techniques)

ML algorithms should perform reliably in the wild

smartphone sensing opens up new frontiers: applications can be spread and big data collected at

unprecedented scale enabling endless research opportunitiesok I think I’m done…

Page 133: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

a quick recapsmartphone’s are everywhere – let’s exploit

their collective sensing and computation capabilities

continuous sensing is still challenging; efficient mobile sensing requires to preserve the phone user

experience (need for energy efficient ML algorithms and smart duty-cycling techniques)

ML algorithms should perform reliably in the wild

smartphone sensing opens up new frontiers: applications can be spread and big data collected at

unprecedented scale enabling endless research opportunitiesbut please bear in mind…

Page 134: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

[email protected] Miluzzo

Mobile Phone Sensing is the Next Big Thing!

Page 135: Darwin Phones: the Evolution of Sensing and Inference on Mobile Phones Emiliano Miluzzo *, Cory T. Cornelius *, Ashwin Ramaswamy *, Tanzeem Choudhury *,

Thank you!!

[email protected] Miluzzo

Mobile Sensing Grouphttp://sensorlab.cs.dartmouth.edu