SlideShare a Scribd company logo
1 of 92
Download to read offline
Saarland University, July 6th, 2023
Breaking the Boundaries of
Human-in-the-Loop Optimization
Yi-Chi Liao. Ph.D. Candidate, UI Group @ Aalto University
Website: http://yichiliao.com | Twitter: @yichiliao | Email: yi-chi.liao@aalto.
f
1
• Ph.D. dissertation:
• Human-in-the-Loop Design Optimization
• Current research group:
• User Interfaces Group at Aalto University, led by Prof Antti Oulasvirta
• Research interests:
• Optimization (inference-free); human-in-the-loop optimization, simulation-based optimization
• User modelling via reinforcement learning
• My tools:
• Bayesian optimization, reinforcement learning, meta-learning, and physics simulation.
• Am joining in November!
Yi-Chi Liao
2
Three goals
• Introduce human-in-the-loop optimization (HILO)
• Introduce the recent advancements in HILO
• Turn you into a Bayesian optimization user
3
Let's play a game! Here's how it works:
You can choose any floating-point number between 0 and 1.2.
Each number corresponds to a specific amount of money in return.
The challenge is to find the number that yields the highest amount of money!
4
Design optimization
0 1
0
3
What is the best location to place the optical sensor?
x ~ [0,1]
y ~ [0,3]
[Optimal Sensor Position for a Computer Mouse, CHI ’20]
5
Design optimization
0 1
0
3
6
Design optimization
0 1
0
3
X := design space
x := a design instance
7
x ~ [0,1]
y ~ [0,3]
Design optimization
0 1
0
3
X := design space
x := a design instance
Optimal design
8
Design optimization
0 1
0
3
9
f is unknown function!
f is expensive to evaluate!
[Caeses (https://www.caeses.com/)]
Design optimization
[Human-in-the-loop optimization of exoskeleton assistance during walking, 2017]
10
Design space (X) Objective-function space (Y)
x1
x2
x3
x4
x5
y1
y2
y3
y4
y5
Black-box function,
y = f(x)
Design instance, x Objective function value, y
Design optimization
11
Manual
optimization
Human-in-the-loop optimization
(HILO)
Beyond current
HILO
12
Manual
optimization
Human-in-the-loop optimization
(HILO)
Beyond current
HILO
13
Manual optimization
Design and
prototype
Analysis User testing
14
Manual optimization
[ThirdHand: Wearing a Robotic Arm to Experience Rich Force Feedback, Liao et al., 2015]
What is the optimal torque value(s) that maximizes the gaming experience?
15
Manual optimization
What is the optimal way to split the strokes?
[EdgeVib: E
ff
ective Alphanumeric Character Output Using a Wrist-Worn Tactile Display, Liao et al., 2015]
16
Manual optimization
[Outside-In: Visualizing Out-of-Sight Regions-of-Interest in a 360 Video Using Spatial Picture-in-Picture Preview, Lin et al., 2017]
What is the optimal position, size, and tilt angle of the window?
17
Manual optimization
Design and
prototype
Analysis User testing
• Pros:
• It is easy and feasible
• Cons:
• It does not guarantee the result
• The developers may introduce bias
• Can not adapt instantly
18
“Don’t ever make the mistake [of thinking] that you can
design something better than what you get from ruthless
massively parallel trial-and-error with a feedback cycle.”
— Linus Torvalds
19
Manual
optimization
Human-in-the-loop optimization
(HILO)
Beyond current
HILO
20
Human-in-the-loop optimization (HILO)
Interactive system
paremeterized by x
Computational
optimizer
User
Design candidate,
x ∈ X
Objective-function
value, y ∈ Y
Interaction
Designer
Design space, X
Objective space, Y
Optimal design, x
21
But which optimizer?
22
Bayesian Optimization (BO)!
23
Will BO solve this?
24
Bayesian optimization —
1.Surrogate model of the target
problem: a Gaussian Process
regression
2.Acquisition function: a
function estimates the utility
of selecting a design choice
25
26
1.Able to (roughly) predict the
outcome of a design choice
2.Able to identify the design
choice that may yield the most
improvements within a huge
design space
Bayesian optimization —
1.Surrogate model of the target
problem: a Gaussian Process
regression
2.Acquisition function: a
function estimates the utility
of selecting a design choice
27
True function
Surrogate model
28
Observations
(Evaluations)
Surrogate model
29
Predicted mean value
Surrogate model
Predicted con
fi
dence interval
i.e., uncertainty
30
Acquisition function
Acquisition value
31
32
33
34
35
36
Surrogate model: Gaussian Process
• GP is a generalized form of multivariate Gaussian (normal)
distribution.
• Covariance function (kernel):
• White noise kernel
• Exponentiated quadratic kernel
• Periodic kernel
37
Selection of kernels
Exponentiated quadratic kernel Periodic kernel
38
Acquisition function
• AF quantifies the utility, potential, or worth of different x points based
on the GP surrogate model.
• Computing AF is way cheaper than actual evaluation!
39
Acquisition function
• AF quantifies the utility, potential, or worth of different x points based
on the GP surrogate model.
• Computing AF is way cheaper than actual evaluation!
• Common acquisition functions:
• Expected Improvements (EI)
• Probability of Improvements (PI)
• Upper Confidence Bound (UCB)
40
Why is BO so good (mathematically)?
41
• BO is a general tool with the least assumption of the problem.
• BO is clearly explainable and transparent.
• BO is a computational framework allows for easy customization & extension.
Applications of Bayesian Optimization
42
43
arxiv.org/abs/2303.14291
BO in engineering
BO in natural science
BO in social/
economical/political
decision-making
[Bayesian optimization explains human active search, Borji and Itti, 2013]
[Generalization guides human exploration in vast decision spaces, Wu et al., 2017]
44
BO is the best model/framework to explain human’s
parametric decision-making!
When Bayesian optimization performs poorly?
45
• GP will not work well when dealing with
non-continuous function
• More sample points will be needed in high-
dimensional (input/output) function
• Gets too expensive when handling with huge
amount of data: GP
fi
tting time ~ O(m^3)
where m is number of observations
Bayesian optimization for human-involved
interactive systems
[Human-in-the-loop optimization of exoskeleton assistance during walking, 2017]
46
[A Bayesian Interactive Optimization Approach to Procedural Animation Design, Siggraph SCA ’10]
47
Bayesian optimization for human-involved
interactive systems
48
[Designing Engaging Games Using Bayesian Optimization, Khajah et al., 2016]
Bayesian optimization for human-involved
interactive systems
Interested in BO for HCI? Also check out Yuki’s works!!!
49
https://koyama.xyz/
Is HILO / BO perfect?
There are boundaries!
• HILO only addresses single-objective problems
• HILO reduces the designer’s agency and ownership in the process
• HILO only optimizes for one user/task at a time
• HILO only works for prototype-free interactions (software interfaces)
• HILO is still not efficient enough!
50
Manual
optimization
Human-in-the-loop optimization
(HILO)
Beyond current
HILO
51
From a single objective to multiple objectives
52
From a single objective to multiple objectives
53
Can we come up with a new
acquisition function?
From a single objective to multiple objectives
54
[Single- and Multiobjective Evolutionary Design Optimization Using
Gaussian Random Field Metamodels, Emmerich, 2005]
[Pareto frontier learning with expensive correlated objectives,
Shah & Ghahramani, 2016]
Short duration,
high amplitude
Long duration,
high amplitude
Long duration,
low amplitude
Short duration,
low amplitude
Short duration,
high amplitude
Long duration,
high amplitude
Long duration,
low amplitude
Short duration,
low amplitude
What is the optimal set of tacton (tactile icon)?
55
[Interaction Design With Multi-Objective Bayesian Optimization, Liao et al., 2023]
Short duration,
high amplitude
Long duration,
high amplitude
Long duration,
low amplitude
Short duration,
low amplitude
Short duration,
high amplitude
Long duration,
high amplitude
Long duration,
low amplitude
Short duration,
low amplitude
What is the optimal set of tacton (tactile icon)?
Length indicates duration
Set 1
56
[Interaction Design With Multi-Objective Bayesian Optimization, Liao et al., 2023]
Short duration,
high amplitude
Long duration,
high amplitude
Long duration,
low amplitude
Short duration,
low amplitude
Short duration,
high amplitude
Long duration,
high amplitude
Long duration,
low amplitude
Short duration,
low amplitude
What is the optimal set of tacton (tactile icon)?
Length indicates duration
Color indicates vibration amplitude
Set 1
Set 2
57
[Interaction Design With Multi-Objective Bayesian Optimization, Liao et al., 2023]
Short duration,
high amplitude
Long duration,
high amplitude
Long duration,
low amplitude
Short duration,
low amplitude
Short duration,
high amplitude
Long duration,
high amplitude
Long duration,
low amplitude
Short duration,
low amplitude
What is the optimal set of tacton (tactile icon)?
Length indicates duration
Color indicates vibration amplitude
Set 1
Set 2
58
Set 3
[Interaction Design With Multi-Objective Bayesian Optimization, Liao et al., 2023]
Short duration,
high amplitude
Long duration,
high amplitude
Long duration,
low amplitude
Short duration,
low amplitude
Short duration,
high amplitude
Long duration,
high amplitude
Long duration,
low amplitude
Short duration,
low amplitude
[Interaction Design With Multi-Objective Bayesian Optimization, Liao et al., 2023]
What are the design objectives?
59
a b
60
[Interaction Design With Multi-Objective Bayesian Optimization, Liao et al., 2023]
[Investigating Positive and Negative Qualities of Human-in-the-Loop Optimization for Designing Interaction Techniques, Chan et al., 2022]
61
[Investigating Positive and Negative Qualities of Human-in-the-Loop Optimization for Designing Interaction Techniques, Chan et al., 2022]
What is the optimal design?
Imagine there are just 10 discrete levels in each parameter,
the number of possible combinations is already 10^4!
62
[Investigating Positive and Negative Qualities of Human-in-the-Loop Optimization for Designing Interaction Techniques, Chan et al., 2022]
What are the design objectives?
63
[Investigating Positive and Negative Qualities of Human-in-the-Loop Optimization for Designing Interaction Techniques, Chan et al., 2022]
What are the design objectives?
E
ffi
ciency & accuracy
64
[Investigating Positive and Negative Qualities of Human-in-the-Loop Optimization for Designing Interaction Techniques, Chan et al., 2022]
65
[Investigating Positive and Negative Qualities of Human-in-the-Loop Optimization for Designing Interaction Techniques, Chan et al., 2022]
Does multi-objective BO really perform better
than human designers?
66
How do designers perceive BO as a tool?
[Interaction Design With Multi-Objective Bayesian Optimization, Liao et al., 2023]
67
[Investigating Positive and Negative Qualities of Human-in-the-Loop Optimization for Designing Interaction Techniques, Chan et al., 2022]
Shortcomings of HILO?
68
Collaborative Bayesian optimization
69
Human’s strength
Quickly determine whether a
design is good or bad
AI (BO)’s strength
More principled exploration
+
A cooperative system where designers can steer the
design when needed
70
[Cooperative Multi-Objective Bayesian Design Optimization, Mo et al., submitted to TiiS 2023]
A cooperative system where designers can steer the
design when needed
71
[Cooperative Multi-Objective Bayesian Design Optimization, Mo et al., submitted to TiiS 2023]
Results:
The Cooperative BO reaches
comparable user performances
with less full evaluations (more
e
ffi
ciently).
The designers felt more engaging
in the process than fully relying on
BO.
Boundaries of current HILO
• HILO only addresses single-objective problems
• HILO reduces the designer’s agency and ownership in the process
• HILO only optimizes for one user/task at a time
• HILO only works for prototype-free interactions (software interfaces)
• HILO is still not efficient enough!
72
Boundaries of current HILO
• HILO only addresses single-objective problems
• HILO reduces the designer’s agency and ownership in the process
• HILO only optimizes for one user/task at a time
• HILO only works for prototype-free interactions (software interfaces)
• HILO is still not efficient enough!
73
From a single user to a population of users
74
[Practical Approaches to Group-Level Multi-Objective Bayesian Optimization in Interaction Technique Design, Liao et al., submitted 2023]
Aggregate
Sample over
design space
Global Pareto-optimal
performances
y2
y1
Mean prediction of y1
Mean prediction of y2
x
y
P1
y
x
x
y
x
y
Global GP
y
x
P2
P3 P4
Boundaries of current HILO
• HILO only addresses single-objective problems
• HILO reduces the designer’s agency and ownership in the process
• HILO only optimizes for one user/task at a time
• HILO only works for prototype-free interactions (software interfaces)
• HILO is still not efficient enough!
75
From prototyping to physical emulation
76
[Button Simulation and Design via FDVV Models, Liao et al., 2020]
From prototyping to physical emulation
77
[Button Simulation and Design via FDVV Models, Liao et al., 2020]
Boundaries of current HILO
• HILO only addresses single-objective problems
• HILO reduces the designer’s agency and ownership in the process
• HILO only optimizes for one user/task at a time
• HILO only works for prototype-free interactions (software interfaces)
• HILO is still not efficient enough!
78
Time-efficiency issue
• Bayesian optimization is mainly seen as a design tool. That is, it is
used to derive the optimal design parameter prior to deploying the
interface/interaction on end-users.
79
Time-efficiency issue
• Bayesian optimization is mainly seen as a design tool. That is, it is
used to derive the optimal design parameter prior to deploying the
interface/interaction on end-users.
• Why?
• It takes ~ 10 x “NO of design parameters” iterations to converge.
This takes 40 - 60 mins
80
But why? Isn’t BO the best optimization tool?
• Yes! It is the upper bound of optimization methods!
• But it suffers from the cold-start problem
81
Can we boost the efficiency of BO?
82
Can we boost the efficiency of BO?
Yes! By augmenting BO with prior information!
83
84
[Practical Approaches to Group-Level Multi-Objective Bayesian Optimization in Interaction Technique Design, Liao et al., submitted 2023]
Extract
representative
points
x
y
All observations across users
Initialize
BO
Mean prediction of y1
Mean prediction of y2
Observed y1
Observed y2
x
y
Warm-Start GP BO with a new user
x
y
New samples
Warm-start Bayesian optimization
85
Current HILO
86
Current HILO
Single-
objective
a b
Pareto frontier
learning
87
Current HILO
Single-
objective Lack agency
Cooperative
AI
Pareto frontier
learning
88
Current HILO
Single-
objective Lack agency
Cooperative
AI
Only for
1 user/task
Population
model
Aggregate
Sam
desig
Mean prediction of y1
Mean prediction of y2
x
y
P1
y
x
x
y
x
y
Global GP
y
x
P2
P3 P4
Pareto frontier
learning
89
Current HILO
Single-
objective Lack agency
Cooperative
AI
Only for
1 user/task
Population
model
Only GUI
Physical
emulation
Pareto frontier
learning
90
Current HILO
Single-
objective
Pareto frontier
learning
Lack agency
Cooperative
AI
Only for
1 user/task
Population
model
Only GUI
Physical
emulation
Low e
ffi
ciency
Constructing
prior
Extract
representative
points
x
y
All observations across users
Initialize
BO
Mean prediction of y1
Mean prediction of y2
Observed y1
Observed y2
x
y
Warm-Start GP BO with a new user
x
y
New samples
My advisor -
Prof. Antti Oulasvirta
Prof. Sunjun Kim
Dr. Kashyap Todi Dr. John Dudley Prof. Per Ola Kristensson Prof. Byungjoo Lee
Prof. Liwei Chan Dr. Aditya Acharya Antti Keurulainen Prof. Andrew Howes
Dr. Aakar Gupta Dr. Ruta Desai
George B. Mo
Dr. Tanya Jonker Dr. Hrvoje Benko Alec Pierce
Aini Putkonen Dr. Hee-Seung Moon Aleksi Ikkala
Amazing people who
made my Ph.D. studies
a wonderful journey!
91
92
How to get Yi-Chi
http://yichiliao.com
Twitter: @yichiliao
Breaking the Boundaries of Human-in-the-Loop Optimization

More Related Content

What's hot

What's hot (20)

強化学習と逆強化学習を組み合わせた模倣学習
強化学習と逆強化学習を組み合わせた模倣学習強化学習と逆強化学習を組み合わせた模倣学習
強化学習と逆強化学習を組み合わせた模倣学習
 
Learning to Rank for Recommender Systems - ACM RecSys 2013 tutorial
Learning to Rank for Recommender Systems -  ACM RecSys 2013 tutorialLearning to Rank for Recommender Systems -  ACM RecSys 2013 tutorial
Learning to Rank for Recommender Systems - ACM RecSys 2013 tutorial
 
Soft Actor Critic 解説
Soft Actor Critic 解説Soft Actor Critic 解説
Soft Actor Critic 解説
 
DeNA TechCon2018 ゲーム体験を支えるための強化学習
DeNA TechCon2018 ゲーム体験を支えるための強化学習DeNA TechCon2018 ゲーム体験を支えるための強化学習
DeNA TechCon2018 ゲーム体験を支えるための強化学習
 
Joint Multisided Exposure Fairness for Search and Recommendation
Joint Multisided Exposure Fairness for Search and RecommendationJoint Multisided Exposure Fairness for Search and Recommendation
Joint Multisided Exposure Fairness for Search and Recommendation
 
From Rapid Prototypes to an end-to-end Model Deployment: an AI Hedge Fund Use...
From Rapid Prototypes to an end-to-end Model Deployment: an AI Hedge Fund Use...From Rapid Prototypes to an end-to-end Model Deployment: an AI Hedge Fund Use...
From Rapid Prototypes to an end-to-end Model Deployment: an AI Hedge Fund Use...
 
サイバーエージェントにおけるMLOpsに関する取り組み at PyDataTokyo 23
サイバーエージェントにおけるMLOpsに関する取り組み at PyDataTokyo 23サイバーエージェントにおけるMLOpsに関する取り組み at PyDataTokyo 23
サイバーエージェントにおけるMLOpsに関する取り組み at PyDataTokyo 23
 
Outracing champion Gran Turismo drivers with deep reinforcement learning
Outracing champion Gran Turismo drivers with deep reinforcement learningOutracing champion Gran Turismo drivers with deep reinforcement learning
Outracing champion Gran Turismo drivers with deep reinforcement learning
 
ECCV2020 オーラル論文完全読破 (2/2)
ECCV2020 オーラル論文完全読破 (2/2) ECCV2020 オーラル論文完全読破 (2/2)
ECCV2020 オーラル論文完全読破 (2/2)
 
Introduction to Prioritized Experience Replay
Introduction to Prioritized Experience ReplayIntroduction to Prioritized Experience Replay
Introduction to Prioritized Experience Replay
 
金融時系列解析入門 AAMAS2021 著者発表会
金融時系列解析入門 AAMAS2021 著者発表会金融時系列解析入門 AAMAS2021 著者発表会
金融時系列解析入門 AAMAS2021 著者発表会
 
AI技術の現状と課題 〜aiboおよび自然言語処理〜
AI技術の現状と課題 〜aiboおよび自然言語処理〜AI技術の現状と課題 〜aiboおよび自然言語処理〜
AI技術の現状と課題 〜aiboおよび自然言語処理〜
 
【DL輪読会】"A Generalist Agent"
【DL輪読会】"A Generalist Agent"【DL輪読会】"A Generalist Agent"
【DL輪読会】"A Generalist Agent"
 
From mcmc to sgnht
From mcmc to sgnhtFrom mcmc to sgnht
From mcmc to sgnht
 
深層強化学習による自動運転車両の経路探索に関する研究
深層強化学習による自動運転車両の経路探索に関する研究深層強化学習による自動運転車両の経路探索に関する研究
深層強化学習による自動運転車両の経路探索に関する研究
 
【DL輪読会】Spectral Normalisation for Deep Reinforcement Learning: An Optimisatio...
【DL輪読会】Spectral Normalisation for Deep Reinforcement Learning: An Optimisatio...【DL輪読会】Spectral Normalisation for Deep Reinforcement Learning: An Optimisatio...
【DL輪読会】Spectral Normalisation for Deep Reinforcement Learning: An Optimisatio...
 
Entrenamiento Intermitente FUTBOL (Aplicable al Futsal y al Futbol 11)
Entrenamiento Intermitente FUTBOL (Aplicable al Futsal y al Futbol 11)Entrenamiento Intermitente FUTBOL (Aplicable al Futsal y al Futbol 11)
Entrenamiento Intermitente FUTBOL (Aplicable al Futsal y al Futbol 11)
 
第5回NIPS読み会・関西発表資料
第5回NIPS読み会・関西発表資料第5回NIPS読み会・関西発表資料
第5回NIPS読み会・関西発表資料
 
[DL輪読会]LightTrack: A Generic Framework for Online Top-Down Human Pose Tracking
[DL輪読会]LightTrack: A Generic Framework for Online Top-Down Human Pose Tracking[DL輪読会]LightTrack: A Generic Framework for Online Top-Down Human Pose Tracking
[DL輪読会]LightTrack: A Generic Framework for Online Top-Down Human Pose Tracking
 
大富豪に対する機械学習の適用 + α
大富豪に対する機械学習の適用 + α大富豪に対する機械学習の適用 + α
大富豪に対する機械学習の適用 + α
 

Similar to Breaking the Boundaries of Human-in-the-Loop Optimization

Elg 5100 project report anurag & jayanshu
Elg 5100 project report   anurag & jayanshuElg 5100 project report   anurag & jayanshu
Elg 5100 project report anurag & jayanshu
Anurag Das
 
PSU 2023 Final Presentation ISSIP_AI_Collab.pptx
PSU 2023 Final Presentation ISSIP_AI_Collab.pptxPSU 2023 Final Presentation ISSIP_AI_Collab.pptx
PSU 2023 Final Presentation ISSIP_AI_Collab.pptx
International Society of Service Innovation Professionals
 
DSD-INT 2014 - OpenMI Symposium - Federated modelling of Critical Infrastruct...
DSD-INT 2014 - OpenMI Symposium - Federated modelling of Critical Infrastruct...DSD-INT 2014 - OpenMI Symposium - Federated modelling of Critical Infrastruct...
DSD-INT 2014 - OpenMI Symposium - Federated modelling of Critical Infrastruct...
Deltares
 
Exploring UX Practices 4 Product Development Agile2012
Exploring UX Practices 4 Product Development Agile2012Exploring UX Practices 4 Product Development Agile2012
Exploring UX Practices 4 Product Development Agile2012
Ariadna Font Llitjos
 
Exploring ux practices 4 product development agile2012
Exploring ux practices 4 product development agile2012Exploring ux practices 4 product development agile2012
Exploring ux practices 4 product development agile2012
drewz lin
 

Similar to Breaking the Boundaries of Human-in-the-Loop Optimization (20)

When UX (guy) Meets Operations
When UX (guy) Meets OperationsWhen UX (guy) Meets Operations
When UX (guy) Meets Operations
 
Model-Based User Interface Optimization: Part IV: ADVANCED TOPICS - At SICSA ...
Model-Based User Interface Optimization: Part IV: ADVANCED TOPICS - At SICSA ...Model-Based User Interface Optimization: Part IV: ADVANCED TOPICS - At SICSA ...
Model-Based User Interface Optimization: Part IV: ADVANCED TOPICS - At SICSA ...
 
Yuandong Tian at AI Frontiers : Planning in Reinforcement Learning
Yuandong Tian at AI Frontiers : Planning in Reinforcement LearningYuandong Tian at AI Frontiers : Planning in Reinforcement Learning
Yuandong Tian at AI Frontiers : Planning in Reinforcement Learning
 
Autonomous Machines with Project Bonsai
Autonomous Machines with Project BonsaiAutonomous Machines with Project Bonsai
Autonomous Machines with Project Bonsai
 
Elg 5100 project report anurag & jayanshu
Elg 5100 project report   anurag & jayanshuElg 5100 project report   anurag & jayanshu
Elg 5100 project report anurag & jayanshu
 
Agile practices and benefits
Agile practices and benefitsAgile practices and benefits
Agile practices and benefits
 
PSU 2023 Final Presentation ISSIP_AI_Collab.pptx
PSU 2023 Final Presentation ISSIP_AI_Collab.pptxPSU 2023 Final Presentation ISSIP_AI_Collab.pptx
PSU 2023 Final Presentation ISSIP_AI_Collab.pptx
 
the application of machine lerning algorithm for SEE
the application of machine lerning algorithm for SEEthe application of machine lerning algorithm for SEE
the application of machine lerning algorithm for SEE
 
Ux prototyping
Ux prototypingUx prototyping
Ux prototyping
 
Beyond TensorBoard: AutoML을 위한 interactive visual analytics 서비스 개발 경험 공유
Beyond TensorBoard: AutoML을 위한 interactive visual analytics 서비스 개발 경험 공유Beyond TensorBoard: AutoML을 위한 interactive visual analytics 서비스 개발 경험 공유
Beyond TensorBoard: AutoML을 위한 interactive visual analytics 서비스 개발 경험 공유
 
Bluemix Paris Meetup - Optimization on Cloud (DOcloud) - 14 octobre 2015
Bluemix Paris Meetup -  Optimization on Cloud (DOcloud) - 14 octobre 2015Bluemix Paris Meetup -  Optimization on Cloud (DOcloud) - 14 octobre 2015
Bluemix Paris Meetup - Optimization on Cloud (DOcloud) - 14 octobre 2015
 
Design optimization of BOP for fatigue and strength in HPHT environment using...
Design optimization of BOP for fatigue and strength in HPHT environment using...Design optimization of BOP for fatigue and strength in HPHT environment using...
Design optimization of BOP for fatigue and strength in HPHT environment using...
 
Parents
ParentsParents
Parents
 
DSD-INT 2014 - OpenMI Symposium - Federated modelling of Critical Infrastruct...
DSD-INT 2014 - OpenMI Symposium - Federated modelling of Critical Infrastruct...DSD-INT 2014 - OpenMI Symposium - Federated modelling of Critical Infrastruct...
DSD-INT 2014 - OpenMI Symposium - Federated modelling of Critical Infrastruct...
 
Exploring UX Practices 4 Product Development Agile2012
Exploring UX Practices 4 Product Development Agile2012Exploring UX Practices 4 Product Development Agile2012
Exploring UX Practices 4 Product Development Agile2012
 
Exploring ux practices 4 product development agile2012
Exploring ux practices 4 product development agile2012Exploring ux practices 4 product development agile2012
Exploring ux practices 4 product development agile2012
 
Tuning for Systematic Trading: Talk 2: Deep Learning
Tuning for Systematic Trading: Talk 2: Deep LearningTuning for Systematic Trading: Talk 2: Deep Learning
Tuning for Systematic Trading: Talk 2: Deep Learning
 
User Experience Design + Agile: The Good, The Bad, and the Ugly
User Experience Design + Agile: The Good, The Bad, and the UglyUser Experience Design + Agile: The Good, The Bad, and the Ugly
User Experience Design + Agile: The Good, The Bad, and the Ugly
 
[SIGGRAPH 2020] Sequential Gallery for Interactive Visual Design Optimization
[SIGGRAPH 2020] Sequential Gallery for Interactive Visual Design Optimization[SIGGRAPH 2020] Sequential Gallery for Interactive Visual Design Optimization
[SIGGRAPH 2020] Sequential Gallery for Interactive Visual Design Optimization
 
Building an mvp that works for users
Building an mvp that works for users Building an mvp that works for users
Building an mvp that works for users
 

Recently uploaded

Spellings Wk 4 and Wk 5 for Grade 4 at CAPS
Spellings Wk 4 and Wk 5 for Grade 4 at CAPSSpellings Wk 4 and Wk 5 for Grade 4 at CAPS
Spellings Wk 4 and Wk 5 for Grade 4 at CAPS
AnaAcapella
 

Recently uploaded (20)

OS-operating systems- ch05 (CPU Scheduling) ...
OS-operating systems- ch05 (CPU Scheduling) ...OS-operating systems- ch05 (CPU Scheduling) ...
OS-operating systems- ch05 (CPU Scheduling) ...
 
What is 3 Way Matching Process in Odoo 17.pptx
What is 3 Way Matching Process in Odoo 17.pptxWhat is 3 Way Matching Process in Odoo 17.pptx
What is 3 Way Matching Process in Odoo 17.pptx
 
Details on CBSE Compartment Exam.pptx1111
Details on CBSE Compartment Exam.pptx1111Details on CBSE Compartment Exam.pptx1111
Details on CBSE Compartment Exam.pptx1111
 
Beyond_Borders_Understanding_Anime_and_Manga_Fandom_A_Comprehensive_Audience_...
Beyond_Borders_Understanding_Anime_and_Manga_Fandom_A_Comprehensive_Audience_...Beyond_Borders_Understanding_Anime_and_Manga_Fandom_A_Comprehensive_Audience_...
Beyond_Borders_Understanding_Anime_and_Manga_Fandom_A_Comprehensive_Audience_...
 
Tatlong Kwento ni Lola basyang-1.pdf arts
Tatlong Kwento ni Lola basyang-1.pdf artsTatlong Kwento ni Lola basyang-1.pdf arts
Tatlong Kwento ni Lola basyang-1.pdf arts
 
Graduate Outcomes Presentation Slides - English
Graduate Outcomes Presentation Slides - EnglishGraduate Outcomes Presentation Slides - English
Graduate Outcomes Presentation Slides - English
 
FICTIONAL SALESMAN/SALESMAN SNSW 2024.pdf
FICTIONAL SALESMAN/SALESMAN SNSW 2024.pdfFICTIONAL SALESMAN/SALESMAN SNSW 2024.pdf
FICTIONAL SALESMAN/SALESMAN SNSW 2024.pdf
 
dusjagr & nano talk on open tools for agriculture research and learning
dusjagr & nano talk on open tools for agriculture research and learningdusjagr & nano talk on open tools for agriculture research and learning
dusjagr & nano talk on open tools for agriculture research and learning
 
Unit 3 Emotional Intelligence and Spiritual Intelligence.pdf
Unit 3 Emotional Intelligence and Spiritual Intelligence.pdfUnit 3 Emotional Intelligence and Spiritual Intelligence.pdf
Unit 3 Emotional Intelligence and Spiritual Intelligence.pdf
 
How to Create and Manage Wizard in Odoo 17
How to Create and Manage Wizard in Odoo 17How to Create and Manage Wizard in Odoo 17
How to Create and Manage Wizard in Odoo 17
 
UGC NET Paper 1 Unit 7 DATA INTERPRETATION.pdf
UGC NET Paper 1 Unit 7 DATA INTERPRETATION.pdfUGC NET Paper 1 Unit 7 DATA INTERPRETATION.pdf
UGC NET Paper 1 Unit 7 DATA INTERPRETATION.pdf
 
On National Teacher Day, meet the 2024-25 Kenan Fellows
On National Teacher Day, meet the 2024-25 Kenan FellowsOn National Teacher Day, meet the 2024-25 Kenan Fellows
On National Teacher Day, meet the 2024-25 Kenan Fellows
 
Spellings Wk 4 and Wk 5 for Grade 4 at CAPS
Spellings Wk 4 and Wk 5 for Grade 4 at CAPSSpellings Wk 4 and Wk 5 for Grade 4 at CAPS
Spellings Wk 4 and Wk 5 for Grade 4 at CAPS
 
OSCM Unit 2_Operations Processes & Systems
OSCM Unit 2_Operations Processes & SystemsOSCM Unit 2_Operations Processes & Systems
OSCM Unit 2_Operations Processes & Systems
 
Python Notes for mca i year students osmania university.docx
Python Notes for mca i year students osmania university.docxPython Notes for mca i year students osmania university.docx
Python Notes for mca i year students osmania university.docx
 
COMMUNICATING NEGATIVE NEWS - APPROACHES .pptx
COMMUNICATING NEGATIVE NEWS - APPROACHES .pptxCOMMUNICATING NEGATIVE NEWS - APPROACHES .pptx
COMMUNICATING NEGATIVE NEWS - APPROACHES .pptx
 
Economic Importance Of Fungi In Food Additives
Economic Importance Of Fungi In Food AdditivesEconomic Importance Of Fungi In Food Additives
Economic Importance Of Fungi In Food Additives
 
AIM of Education-Teachers Training-2024.ppt
AIM of Education-Teachers Training-2024.pptAIM of Education-Teachers Training-2024.ppt
AIM of Education-Teachers Training-2024.ppt
 
Wellbeing inclusion and digital dystopias.pptx
Wellbeing inclusion and digital dystopias.pptxWellbeing inclusion and digital dystopias.pptx
Wellbeing inclusion and digital dystopias.pptx
 
Accessible Digital Futures project (20/03/2024)
Accessible Digital Futures project (20/03/2024)Accessible Digital Futures project (20/03/2024)
Accessible Digital Futures project (20/03/2024)
 

Breaking the Boundaries of Human-in-the-Loop Optimization

  • 1. Saarland University, July 6th, 2023 Breaking the Boundaries of Human-in-the-Loop Optimization Yi-Chi Liao. Ph.D. Candidate, UI Group @ Aalto University Website: http://yichiliao.com | Twitter: @yichiliao | Email: yi-chi.liao@aalto. f 1
  • 2. • Ph.D. dissertation: • Human-in-the-Loop Design Optimization • Current research group: • User Interfaces Group at Aalto University, led by Prof Antti Oulasvirta • Research interests: • Optimization (inference-free); human-in-the-loop optimization, simulation-based optimization • User modelling via reinforcement learning • My tools: • Bayesian optimization, reinforcement learning, meta-learning, and physics simulation. • Am joining in November! Yi-Chi Liao 2
  • 3. Three goals • Introduce human-in-the-loop optimization (HILO) • Introduce the recent advancements in HILO • Turn you into a Bayesian optimization user 3
  • 4. Let's play a game! Here's how it works: You can choose any floating-point number between 0 and 1.2. Each number corresponds to a specific amount of money in return. The challenge is to find the number that yields the highest amount of money! 4
  • 5. Design optimization 0 1 0 3 What is the best location to place the optical sensor? x ~ [0,1] y ~ [0,3] [Optimal Sensor Position for a Computer Mouse, CHI ’20] 5
  • 7. Design optimization 0 1 0 3 X := design space x := a design instance 7 x ~ [0,1] y ~ [0,3]
  • 8. Design optimization 0 1 0 3 X := design space x := a design instance Optimal design 8
  • 9. Design optimization 0 1 0 3 9 f is unknown function! f is expensive to evaluate!
  • 10. [Caeses (https://www.caeses.com/)] Design optimization [Human-in-the-loop optimization of exoskeleton assistance during walking, 2017] 10
  • 11. Design space (X) Objective-function space (Y) x1 x2 x3 x4 x5 y1 y2 y3 y4 y5 Black-box function, y = f(x) Design instance, x Objective function value, y Design optimization 11
  • 15. Manual optimization [ThirdHand: Wearing a Robotic Arm to Experience Rich Force Feedback, Liao et al., 2015] What is the optimal torque value(s) that maximizes the gaming experience? 15
  • 16. Manual optimization What is the optimal way to split the strokes? [EdgeVib: E ff ective Alphanumeric Character Output Using a Wrist-Worn Tactile Display, Liao et al., 2015] 16
  • 17. Manual optimization [Outside-In: Visualizing Out-of-Sight Regions-of-Interest in a 360 Video Using Spatial Picture-in-Picture Preview, Lin et al., 2017] What is the optimal position, size, and tilt angle of the window? 17
  • 18. Manual optimization Design and prototype Analysis User testing • Pros: • It is easy and feasible • Cons: • It does not guarantee the result • The developers may introduce bias • Can not adapt instantly 18
  • 19. “Don’t ever make the mistake [of thinking] that you can design something better than what you get from ruthless massively parallel trial-and-error with a feedback cycle.” — Linus Torvalds 19
  • 21. Human-in-the-loop optimization (HILO) Interactive system paremeterized by x Computational optimizer User Design candidate, x ∈ X Objective-function value, y ∈ Y Interaction Designer Design space, X Objective space, Y Optimal design, x 21
  • 24. Will BO solve this? 24
  • 25. Bayesian optimization — 1.Surrogate model of the target problem: a Gaussian Process regression 2.Acquisition function: a function estimates the utility of selecting a design choice 25
  • 26. 26 1.Able to (roughly) predict the outcome of a design choice 2.Able to identify the design choice that may yield the most improvements within a huge design space Bayesian optimization — 1.Surrogate model of the target problem: a Gaussian Process regression 2.Acquisition function: a function estimates the utility of selecting a design choice
  • 29. 29 Predicted mean value Surrogate model Predicted con fi dence interval i.e., uncertainty
  • 31. 31
  • 32. 32
  • 33. 33
  • 34. 34
  • 35. 35
  • 36. 36
  • 37. Surrogate model: Gaussian Process • GP is a generalized form of multivariate Gaussian (normal) distribution. • Covariance function (kernel): • White noise kernel • Exponentiated quadratic kernel • Periodic kernel 37
  • 38. Selection of kernels Exponentiated quadratic kernel Periodic kernel 38
  • 39. Acquisition function • AF quantifies the utility, potential, or worth of different x points based on the GP surrogate model. • Computing AF is way cheaper than actual evaluation! 39
  • 40. Acquisition function • AF quantifies the utility, potential, or worth of different x points based on the GP surrogate model. • Computing AF is way cheaper than actual evaluation! • Common acquisition functions: • Expected Improvements (EI) • Probability of Improvements (PI) • Upper Confidence Bound (UCB) 40
  • 41. Why is BO so good (mathematically)? 41 • BO is a general tool with the least assumption of the problem. • BO is clearly explainable and transparent. • BO is a computational framework allows for easy customization & extension.
  • 42. Applications of Bayesian Optimization 42
  • 43. 43 arxiv.org/abs/2303.14291 BO in engineering BO in natural science BO in social/ economical/political decision-making
  • 44. [Bayesian optimization explains human active search, Borji and Itti, 2013] [Generalization guides human exploration in vast decision spaces, Wu et al., 2017] 44 BO is the best model/framework to explain human’s parametric decision-making!
  • 45. When Bayesian optimization performs poorly? 45 • GP will not work well when dealing with non-continuous function • More sample points will be needed in high- dimensional (input/output) function • Gets too expensive when handling with huge amount of data: GP fi tting time ~ O(m^3) where m is number of observations
  • 46. Bayesian optimization for human-involved interactive systems [Human-in-the-loop optimization of exoskeleton assistance during walking, 2017] 46
  • 47. [A Bayesian Interactive Optimization Approach to Procedural Animation Design, Siggraph SCA ’10] 47 Bayesian optimization for human-involved interactive systems
  • 48. 48 [Designing Engaging Games Using Bayesian Optimization, Khajah et al., 2016] Bayesian optimization for human-involved interactive systems
  • 49. Interested in BO for HCI? Also check out Yuki’s works!!! 49 https://koyama.xyz/
  • 50. Is HILO / BO perfect? There are boundaries! • HILO only addresses single-objective problems • HILO reduces the designer’s agency and ownership in the process • HILO only optimizes for one user/task at a time • HILO only works for prototype-free interactions (software interfaces) • HILO is still not efficient enough! 50
  • 52. From a single objective to multiple objectives 52
  • 53. From a single objective to multiple objectives 53 Can we come up with a new acquisition function?
  • 54. From a single objective to multiple objectives 54 [Single- and Multiobjective Evolutionary Design Optimization Using Gaussian Random Field Metamodels, Emmerich, 2005] [Pareto frontier learning with expensive correlated objectives, Shah & Ghahramani, 2016]
  • 55. Short duration, high amplitude Long duration, high amplitude Long duration, low amplitude Short duration, low amplitude Short duration, high amplitude Long duration, high amplitude Long duration, low amplitude Short duration, low amplitude What is the optimal set of tacton (tactile icon)? 55 [Interaction Design With Multi-Objective Bayesian Optimization, Liao et al., 2023]
  • 56. Short duration, high amplitude Long duration, high amplitude Long duration, low amplitude Short duration, low amplitude Short duration, high amplitude Long duration, high amplitude Long duration, low amplitude Short duration, low amplitude What is the optimal set of tacton (tactile icon)? Length indicates duration Set 1 56 [Interaction Design With Multi-Objective Bayesian Optimization, Liao et al., 2023]
  • 57. Short duration, high amplitude Long duration, high amplitude Long duration, low amplitude Short duration, low amplitude Short duration, high amplitude Long duration, high amplitude Long duration, low amplitude Short duration, low amplitude What is the optimal set of tacton (tactile icon)? Length indicates duration Color indicates vibration amplitude Set 1 Set 2 57 [Interaction Design With Multi-Objective Bayesian Optimization, Liao et al., 2023]
  • 58. Short duration, high amplitude Long duration, high amplitude Long duration, low amplitude Short duration, low amplitude Short duration, high amplitude Long duration, high amplitude Long duration, low amplitude Short duration, low amplitude What is the optimal set of tacton (tactile icon)? Length indicates duration Color indicates vibration amplitude Set 1 Set 2 58 Set 3 [Interaction Design With Multi-Objective Bayesian Optimization, Liao et al., 2023]
  • 59. Short duration, high amplitude Long duration, high amplitude Long duration, low amplitude Short duration, low amplitude Short duration, high amplitude Long duration, high amplitude Long duration, low amplitude Short duration, low amplitude [Interaction Design With Multi-Objective Bayesian Optimization, Liao et al., 2023] What are the design objectives? 59
  • 60. a b 60 [Interaction Design With Multi-Objective Bayesian Optimization, Liao et al., 2023]
  • 61. [Investigating Positive and Negative Qualities of Human-in-the-Loop Optimization for Designing Interaction Techniques, Chan et al., 2022] 61
  • 62. [Investigating Positive and Negative Qualities of Human-in-the-Loop Optimization for Designing Interaction Techniques, Chan et al., 2022] What is the optimal design? Imagine there are just 10 discrete levels in each parameter, the number of possible combinations is already 10^4! 62
  • 63. [Investigating Positive and Negative Qualities of Human-in-the-Loop Optimization for Designing Interaction Techniques, Chan et al., 2022] What are the design objectives? 63
  • 64. [Investigating Positive and Negative Qualities of Human-in-the-Loop Optimization for Designing Interaction Techniques, Chan et al., 2022] What are the design objectives? E ffi ciency & accuracy 64
  • 65. [Investigating Positive and Negative Qualities of Human-in-the-Loop Optimization for Designing Interaction Techniques, Chan et al., 2022] 65
  • 66. [Investigating Positive and Negative Qualities of Human-in-the-Loop Optimization for Designing Interaction Techniques, Chan et al., 2022] Does multi-objective BO really perform better than human designers? 66
  • 67. How do designers perceive BO as a tool? [Interaction Design With Multi-Objective Bayesian Optimization, Liao et al., 2023] 67
  • 68. [Investigating Positive and Negative Qualities of Human-in-the-Loop Optimization for Designing Interaction Techniques, Chan et al., 2022] Shortcomings of HILO? 68
  • 69. Collaborative Bayesian optimization 69 Human’s strength Quickly determine whether a design is good or bad AI (BO)’s strength More principled exploration +
  • 70. A cooperative system where designers can steer the design when needed 70 [Cooperative Multi-Objective Bayesian Design Optimization, Mo et al., submitted to TiiS 2023]
  • 71. A cooperative system where designers can steer the design when needed 71 [Cooperative Multi-Objective Bayesian Design Optimization, Mo et al., submitted to TiiS 2023] Results: The Cooperative BO reaches comparable user performances with less full evaluations (more e ffi ciently). The designers felt more engaging in the process than fully relying on BO.
  • 72. Boundaries of current HILO • HILO only addresses single-objective problems • HILO reduces the designer’s agency and ownership in the process • HILO only optimizes for one user/task at a time • HILO only works for prototype-free interactions (software interfaces) • HILO is still not efficient enough! 72
  • 73. Boundaries of current HILO • HILO only addresses single-objective problems • HILO reduces the designer’s agency and ownership in the process • HILO only optimizes for one user/task at a time • HILO only works for prototype-free interactions (software interfaces) • HILO is still not efficient enough! 73
  • 74. From a single user to a population of users 74 [Practical Approaches to Group-Level Multi-Objective Bayesian Optimization in Interaction Technique Design, Liao et al., submitted 2023] Aggregate Sample over design space Global Pareto-optimal performances y2 y1 Mean prediction of y1 Mean prediction of y2 x y P1 y x x y x y Global GP y x P2 P3 P4
  • 75. Boundaries of current HILO • HILO only addresses single-objective problems • HILO reduces the designer’s agency and ownership in the process • HILO only optimizes for one user/task at a time • HILO only works for prototype-free interactions (software interfaces) • HILO is still not efficient enough! 75
  • 76. From prototyping to physical emulation 76 [Button Simulation and Design via FDVV Models, Liao et al., 2020]
  • 77. From prototyping to physical emulation 77 [Button Simulation and Design via FDVV Models, Liao et al., 2020]
  • 78. Boundaries of current HILO • HILO only addresses single-objective problems • HILO reduces the designer’s agency and ownership in the process • HILO only optimizes for one user/task at a time • HILO only works for prototype-free interactions (software interfaces) • HILO is still not efficient enough! 78
  • 79. Time-efficiency issue • Bayesian optimization is mainly seen as a design tool. That is, it is used to derive the optimal design parameter prior to deploying the interface/interaction on end-users. 79
  • 80. Time-efficiency issue • Bayesian optimization is mainly seen as a design tool. That is, it is used to derive the optimal design parameter prior to deploying the interface/interaction on end-users. • Why? • It takes ~ 10 x “NO of design parameters” iterations to converge. This takes 40 - 60 mins 80
  • 81. But why? Isn’t BO the best optimization tool? • Yes! It is the upper bound of optimization methods! • But it suffers from the cold-start problem 81
  • 82. Can we boost the efficiency of BO? 82
  • 83. Can we boost the efficiency of BO? Yes! By augmenting BO with prior information! 83
  • 84. 84 [Practical Approaches to Group-Level Multi-Objective Bayesian Optimization in Interaction Technique Design, Liao et al., submitted 2023] Extract representative points x y All observations across users Initialize BO Mean prediction of y1 Mean prediction of y2 Observed y1 Observed y2 x y Warm-Start GP BO with a new user x y New samples Warm-start Bayesian optimization
  • 87. 87 Current HILO Single- objective Lack agency Cooperative AI Pareto frontier learning
  • 88. 88 Current HILO Single- objective Lack agency Cooperative AI Only for 1 user/task Population model Aggregate Sam desig Mean prediction of y1 Mean prediction of y2 x y P1 y x x y x y Global GP y x P2 P3 P4 Pareto frontier learning
  • 89. 89 Current HILO Single- objective Lack agency Cooperative AI Only for 1 user/task Population model Only GUI Physical emulation Pareto frontier learning
  • 90. 90 Current HILO Single- objective Pareto frontier learning Lack agency Cooperative AI Only for 1 user/task Population model Only GUI Physical emulation Low e ffi ciency Constructing prior Extract representative points x y All observations across users Initialize BO Mean prediction of y1 Mean prediction of y2 Observed y1 Observed y2 x y Warm-Start GP BO with a new user x y New samples
  • 91. My advisor - Prof. Antti Oulasvirta Prof. Sunjun Kim Dr. Kashyap Todi Dr. John Dudley Prof. Per Ola Kristensson Prof. Byungjoo Lee Prof. Liwei Chan Dr. Aditya Acharya Antti Keurulainen Prof. Andrew Howes Dr. Aakar Gupta Dr. Ruta Desai George B. Mo Dr. Tanya Jonker Dr. Hrvoje Benko Alec Pierce Aini Putkonen Dr. Hee-Seung Moon Aleksi Ikkala Amazing people who made my Ph.D. studies a wonderful journey! 91
  • 92. 92 How to get Yi-Chi http://yichiliao.com Twitter: @yichiliao Breaking the Boundaries of Human-in-the-Loop Optimization