34
#SIRACon14 When Algorithms Are Our Co- Pilots Allison Miller President, SIRA

When Algorithms Are Our Co-Pilots

Embed Size (px)

DESCRIPTION

Effective risk management in large, user-facing systems (like platforms) requires strategic and effective use of data. Typically we discuss data in the context of behavioral analysis within the systems, where it is used to better profile threats and vulnerabilities. In this talk we will discuss how data-driven controls are incorporated into the systems themselves, and typical areas where risk managers will need to focus their attention and algorithms in the future: defenses, dev/ops, & devices. We will explore what this means for risk control designers working on closed systems versus open systems, and also review proper care and feeding of algorithms.

Citation preview

Page 1: When Algorithms Are Our Co-Pilots

#SIRACon14

When Algorithms Are Our Co-Pilots

Allison MillerPresident, SIRA

Page 2: When Algorithms Are Our Co-Pilots

#SIRACon14 @selenakyle

Outline

❖ What are algorithms?

❖ OMG, I LOVE Algorithms!

❖ Algorithms as the foundation of information risk control systems

❖ Algorithms VS (Adversity + Adversaries)

❖ Control vs Complexity

Page 3: When Algorithms Are Our Co-Pilots

#SIRACon14 @selenakyle

#TBT

Page 4: When Algorithms Are Our Co-Pilots

#SIRACon14 @selenakyle

Namshub

❖Tie Your Shoes

❖Common core

❖$15.51

Page 5: When Algorithms Are Our Co-Pilots

1,000,000,000,000

-1

Are you a good algorithm or a bad algorithm?

#SIRACon14 @selenakyle

ed

Page 6: When Algorithms Are Our Co-Pilots

#SIRACon14 @selenakyle

Namshub

❖Tie Your Shoes

❖Common core❖$15.51

Page 7: When Algorithms Are Our Co-Pilots

#SIRACon14 @selenakyle

#InQuantsWeTrust

Page 8: When Algorithms Are Our Co-Pilots

#SIRACon14 @selenakyle

w(e) <3 Math❖ Gauss: What’s Normal Anyway

❖ Bernoulli: Making Heads or Tails of things

❖ Plenty of Poisson in the Sea

❖ Euler: Graph theory, topology

❖ All about that Bayes

❖ Groves of Decision Trees, Nets of Neurons, Totally Random Forests

Page 9: When Algorithms Are Our Co-Pilots

#SIRACon14 @selenakyle

#WWAD?

Page 10: When Algorithms Are Our Co-Pilots

#SIRACon14 @selenakyle

What Algorithms Do

❖Simplify❖Match❖Price❖Automate❖Predict❖Compete

Page 11: When Algorithms Are Our Co-Pilots

#SIRACon14 @selenakyle

#FTW

Page 12: When Algorithms Are Our Co-Pilots

#SIRACon14 @selenakyle

The Player is the Game

❖ Tic Tac Toe❖ Chess❖ Trivia❖ Poker❖ Tit for Tat

Kasparov vs X3D Fritz

Page 13: When Algorithms Are Our Co-Pilots

#SIRACon14 @selenakyle

Discussing Games

Mechanics of decision trees

UP

DOWN

CIRCLE

RED

BLUE

MARIO

LUIGI

KIRBY

GIZMO

10, 3

2, 10

10

2, 5

-3, 3A

B

B

A

A

A

Page 14: When Algorithms Are Our Co-Pilots

#SIRACon14 @selenakyle

#WTF(What, Thou Failest?)

* Es tu, Bot — eh?

Page 15: When Algorithms Are Our Co-Pilots

#SIRACon14 @selenakyle

Algorithms Gone Wild

❖Bidding/auctions❖Targeting❖Too Many Likes

Page 16: When Algorithms Are Our Co-Pilots

Why?

Page 17: When Algorithms Are Our Co-Pilots

#SIRACon14 @selenakyle

Speaking of $$$

❖Black Scholes❖Automation of trading

❖HFT: Faster than the speed of light

Page 18: When Algorithms Are Our Co-Pilots

#SIRACon14 @selenakyle

#YMMV

Page 19: When Algorithms Are Our Co-Pilots

#SIRACon14 @selenakyle

Managing DecisionsRisk Management = Decision Management Algorithms are useful for “outsourcing” decisions to our

systems

Systems enforce risk strategies

Defenders and operators design control systems that make decisions Where risks/issues manifest in observable behavior

They can make moves/counter-moves depending on the context and understanding of an actor’s identity or intent

Where system or individual costs/payoffs depend on the outcome of an actor’s actions

Decision strategies can be optimized

Page 20: When Algorithms Are Our Co-Pilots

UP

DOWN

CIRCLE

RED

BLUE

MARIO

LUIGI

KIRBY

GIZMO

10, 3

2, 10

2, 5

-3, 3A

B

B

A

A

A

Payoffs!

Page 21: When Algorithms Are Our Co-Pilots

Optimization!

% of populationC

ost

Number of Defects Produced

Cost of Control

Cost of Defects

CV

Total CostCV

Operations

Risk/Loss

Page 22: When Algorithms Are Our Co-Pilots

#SIRACon14 @selenakyle

Algorithms Rule

❖Defenses❖DevOps❖Devices

Page 23: When Algorithms Are Our Co-Pilots

#SIRACon14 @selenakyle

Operating a Learning System

Disposition&

Time

EmailCC#ItemsTotal

Submit

Maybe

No!

Yes!! SuOutcomeSuAttempt

Black & Whitelists

Machine Learning

Velocity & Spend caps

Geo & IP Logic

Linking

Data

• Reporting• Metrics• Analysis• Modeling

GoodBadIndeterminate

Page 24: When Algorithms Are Our Co-Pilots

#SIRACon14 @selenakyle

The Better Mousetrap

Automates defensive action x-platform

- Fast

- Accurate

- Cheap

In Real TimeIn Time to Minimize

LossReasonable False PositivesAs good as a human specialist

Reduces Reduction > Cost CreatedCheaper than Manual

intervention

Page 25: When Algorithms Are Our Co-Pilots

#SIRACon14 @selenakyle

TheJoyof… #If/Then

Page 26: When Algorithms Are Our Co-Pilots

#SIRACon14 @selenakyle

The first rule of any technology

used in a business is that automation

applied to an efficient operation

will magnify the efficiency.

The second is that automation

applied to an inefficient

operation will magnify the inefficiency.–Bill Gates

Page 27: When Algorithms Are Our Co-Pilots

#SIRACon14 @selenakyle

Managing Social Systems

“…all policy systems…pose incentives that are reacted to by groups of agents acting in their own interest, and often those reactions are unexpected

and act counter to the policy’s intentions…”

— Brian Arthur, All Systems will be Gamed: Exploitive Behavior in Economic and Social Systems

Page 28: When Algorithms Are Our Co-Pilots

#SIRACon14 @selenakyle

Send in the Bots

❖ Convenience vs dependence

❖ Transparency vs obfuscation

❖ Simplicity vs complexity

Page 29: When Algorithms Are Our Co-Pilots

#SIRACon14 @selenakyle

#TheInterdependenceOf(AllThe)Things

Page 30: When Algorithms Are Our Co-Pilots

#SIRACon14 @selenakyleAsimov’s 3 Laws of Robotics

1. A robot may not injure a human being or, through inaction, allow a human being to come to harm.

2. A robot must obey the orders given to it by human beings, except where such orders would conflict with the First Law.

3. A robot must protect its own existence as long as such protection does not conflict with the First or Second Law.

Page 31: When Algorithms Are Our Co-Pilots

#SIRACon14 @selenakyle

SIRAtonin 3 Bot Guidelines

1.Bots are thoughtless but purposeful

2.Bots compete but will also collude

3.The behavior of a network of bots is more complex than the behavior of a single bot

Page 32: When Algorithms Are Our Co-Pilots

#SIRACon14 @selenakyle

The Complexity Dimension

Embedded system controls are part of the system

Instrument your instrumentation

Learning systems are hungry

Intelligent agents are dumb

Fatal feedback is fatal

Risk hides in complexity

Page 33: When Algorithms Are Our Co-Pilots

https://www.youtube.com/watch?v=G1t4M2XnIhI#action=share

Photo: Michael Rubenstein/Harvard University

Page 34: When Algorithms Are Our Co-Pilots

Allison Miller

@selenakyle