Considering high-impact emerging technologies (AI machine learning and blockchain) together suggests the emergence of a new class of global computational infrastructure: smart networks. Smart networks are intelligent self-operating computation networks such as deep learning neural nets, blockchains, UAV fleets, industrial robotics cloudminds
How AI, OpenAI, and ChatGPT impact business and software.
Smart Networks: Blockchain, Deep Learning, and Quantum Computing
1. Melanie Swan
Purdue University
Emerging Technologies
shaping the future of
Fraud Detection, Banking, and Finance
Association of Certified Fraud Examiners
Indianapolis IN, August 8, 2019
Slides: http://slideshare.net/LaBlogga
2. 8 Aug 2019
EmergingTech 1
Melanie Swan, Technology Theorist
Philosophy Department, Purdue University,
Indiana, USA
Founder, Institute for Blockchain Studies
Singularity University Instructor; Institute for Ethics and
Emerging Technology Affiliate Scholar; EDGE
Essayist; FQXi Advisor
Traditional Markets Background Economics and Financial
Theory Leadership
New Economies research group
Source: http://www.melanieswan.com, http://blockchainstudies.org/NSNE.pdf, http://blockchainstudies.org/Metaphilosophy_CFP.pdf
https://www.facebook.com/groups/NewEconomies
3. 8 Aug 2019
EmergingTech
Smart Network Thesis
2
Considering high-impact emerging
technologies (AI machine learning and
blockchain) together suggests the emergence
of a new class of global computational
infrastructure: smart networks
(Smart networks: intelligent self-operating computation
networks such as deep learning neural nets,
blockchains, UAV fleets, industrial robotics cloudminds)
4. 8 Aug 2019
EmergingTech
Agenda
Digital Transformation
Deep Learning Neural Networks
Blockchain Technology
Implications for Fraud
3
5. 8 Aug 2019
EmergingTech
Top Job Growth Areas
Top job machine learning and data analysis supplanted
by blockchain in 2018
1,775 US blockchain-related job openings August 2018
300 percent annual increase
Median salary: $84,884/year ($52,461 US avg)
4
Source: Glassdoor’s August 2018 Local Pay Report. https://www.glassdoor.com/research/rise-in-bitcoin-jobs/
6. 8 Aug 2019
EmergingTech
Digital Transformation Journey
Digital transformation: digitizing information and
processes in all enterprise and government functions
$3.8 trillion global IT spend 2019 (Gartner)
$1.3 trillion Digital Transformation Technologies (IDC)
5
Source: https://www.gartner.com/en/newsroom/press-releases/2019-01-28-gartner-says-global-it-spending-to-reach--3-8-trillio,
https://www.idc.com/getdoc.jsp?containerId=prUS43381817
Digital transformation
Technology used to make
existing work more
efficient, now technology is
transforming the work itself
Convergence of
blockchain, IoT, AI, Cloud
technologies
7. 8 Aug 2019
EmergingTech
Exascale supercomputing 2021e
Exabyte global data volume 2020e: 40 EB
Scientific, governmental, corporate, and personal
Big Data ≠ Smart Data
Sources: http://www.oyster-ims.com/media/resources/dealing-information-growth-dark-data-six-practical-steps/,
https://www.theverge.com/2019/3/18/18271328/supercomputer-build-date-exascale-intel-argonne-national-laboratory-energy
6
Only 6% data protected, only
42% companies say they know
how to extract meaningful
insights from the data available
to them (Oxford Economics
Workforce 2020)
8. 8 Aug 2019
EmergingTech
Why do we need Learning Technologies?
7
Big data is not smart data (i.e. usable)
New data science methods needed for data growth,
older learning algorithms under-performing
Source: http://blog.algorithmia.com/introduction-to-deep-learning-2016
9. 8 Aug 2019
EmergingTech
Artificial Intelligence (AI)
Artificial intelligence is using
computers to preform cognitive
work (physical or mental) that
usually requires a human
Deep Learning/Machine Learning
is the biggest area in AI
8
Source: Swan, M. Philosophy of Deep Learning Networks: Reality Automation Modules.
Ke Jie vs. AlphaGo AI Go player, Future of
Go Summit, Wuzhen China, May 2017
10. 8 Aug 2019
EmergingTech
Progression in AI Learning Machines
9
Single-purpose AI:
Hard-coded rules
Multi-purpose AI:
Algorithm detects rules,
reusable template
Question-answering AI:
Natural-language processing
Deep Learning prototypeHard-coded AI machine Deep Learning machine
Deep Blue, 1997 Watson, 2011 AlphaGo, 2016
11. 8 Aug 2019
EmergingTech 10
Conceptual Definition:
Deep learning is a computer program that can
identify what something is (physical or digital)
Technical Definition:
Deep learning is a class of machine learning
algorithms in the form of a neural network that
uses a cascade of layers of processing units to
extract features from data sets in order to make
predictive guesses about new data
Source: Extending Jann LeCun, http://spectrum.ieee.org/automaton/robotics/artificial-intelligence/facebook-ai-director-yann-lecun-
on-deep-learning
What is Deep Learning?
12. 8 Aug 2019
EmergingTech
How are AI and Deep Learning related?
11
Source: Machine Learning Guide, 9. Deep Learning
Artificial intelligence:
Using computers to do cognitive work
that usually requires a human
Machine learning:
A statistical method in which
computers perform tasks by relying
on information patterns and inference
as opposed to explicit instructions
Neural network:
A computer system modeled on the
human brain and nervous system
Deep learning:
Program that can recognize objects
Deep
Learning
Neural Nets
Machine Learning
Artificial Intelligence
Computer Science
Within the Computer Science
discipline, in the field of Artificial
Intelligence, Deep Learning is a
class of Machine Learning
algorithms, that are in the form
of a Neural Network
13. 8 Aug 2019
EmergingTech
What is a Neural Network?
12
Intuition: create an Artificial Neural Network to solve
problems in the same way as the human brain
14. 8 Aug 2019
EmergingTech
Agenda
Digital Transformation
Deep Learning Neural Networks
Blockchain Technology
Implications for Fraud
13
15. 8 Aug 2019
EmergingTech
Why is it called “Deep” Learning?
Hidden layers of processing (2-20 intermediary layers)
“Deep” networks (3+ layers) versus “shallow” (1-2 layers)
Basic deep learning network: 5 layers; GoogleNet: 22 layers
14
Sandwich Architecture:
visible Input and Output layers
with hidden processing layers
GoogleNet:
22 layers
16. 8 Aug 2019
EmergingTech
Why Deep “Learning”?
System is “dumb” (i.e. mechanistic)
“Learns” by making trial-and-error guesses about the data it
receives to log the relevant features in order to identifying
similar examples
Usual AI argument: big enough data is what makes a
difference (“simple” algorithms run over large data sets)
15
Input: Big Data (e.g.;
many examples)
Method: Trial-and-error
guesses to adjust node weights
Output: system identifies
new examples
17. 8 Aug 2019
EmergingTech
Sample task: is that a Car?
Create an image recognition system that determines
which features are relevant (at increasingly higher levels
of abstraction) and correctly identifies new examples
16
Source: Jann LeCun, http://www.pamitc.org/cvpr15/files/lecun-20150610-cvpr-keynote.pdf
18. 8 Aug 2019
EmergingTech
Two classes of Learning Systems
Supervised and Unsupervised Learning
Supervised
Classify already-
labeled data
Unsupervised
Find patterns in
unlabeled data
17
Source: https://www.slideshare.net/ThomasDaSilvaPaula/an-introduction-to-machine-learning-and-a-little-bit-of-deep-learning
19. 8 Aug 2019
EmergingTech
Early success in Supervised Learning (2011)
YouTube: user-classified data
perfect for Supervised Learning
18
Source: Google Brain: Le, QV, Dean, Jeff, Ng, Andrew, et al. 2012. Building high-level features using large scale unsupervised
learning. https://arxiv.org/abs/1112.6209
20. 8 Aug 2019
EmergingTech
2 main kinds of Deep Learning neural nets
19
Source: Yann LeCun, CVPR 2015 keynote (Computer Vision ), "What's wrong with Deep Learning" http://t.co/nPFlPZzMEJ
Convolutional Neural Nets
Image recognition
Convolve: roll up to higher
levels of abstraction to identify
feature sets
Recurrent Neural Nets
Speech, text, audio recognition
Recur: iterate over sequential
inputs with a memory function
LSTM (Long Short-Term
Memory) remembers
sequences and avoids
gradient vanishing
21. 8 Aug 2019
EmergingTech
Image Recognition and Computer Vision
20
Source: Quoc Le, https://arxiv.org/abs/1112.6209; Yann LeCun, NIPS 2016,
https://drive.google.com/file/d/0BxKBnD5y2M8NREZod0tVdW5FLTQ/view
Marv Minsky, 1966
“summer project”
Jeff Hawkins, 2004, Hierarchical
Temporal Memory (HTM)
Quoc Le, 2011, Google
Brain cat recognition
Convolutional net for autonomous driving, http://cs231n.github.io/convolutional-networks
History
Current state of
the art - 2019
22. 8 Aug 2019
EmergingTech
Image Classification
21
Source: https://cs.stanford.edu/people/karpathy/deepimagesent/?hn
Human-level image recognition and captioning (2018)
23. 8 Aug 2019
EmergingTech
Machine “Understanding” of Concepts
22
Source: https://cs.stanford.edu/people/karpathy/deepimagesent/?hn
“Understanding” is the system’s three-step process
Image -> internal representation -> text
Labels “tennis racket” = concepts
Machine learning: Kantian-level object recognition, not Hegelian
25. 8 Aug 2019
EmergingTech
Modular Processing Units
24
Source: http://deeplearning.stanford.edu/tutorial
1. Input 2. Hidden layers 3. Output
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Unit: processing unit, logit (logistic
regression unit), perceptron, artificial neuron
26. 8 Aug 2019
EmergingTech
Image Recognition
Digitize Input Data into Vectors
25
Source: Quoc V. Le, A Tutorial on Deep Learning, Part 1: Nonlinear Classifiers and The Backpropagation Algorithm, 2015, Google
Brain, https://cs.stanford.edu/~quocle/tutorial1.pdf
27. 8 Aug 2019
EmergingTech
Image Recognition
Log features and trial-and-error test
26
1. Input 2. Hidden layers 3. Output
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Source: http://deeplearning.stanford.edu/tutorial; MNIST dataset: http://yann.lecun.com/exdb/mnist
Mathematical methods used to update the weights
Linear algebra: matrix multiplications of input vectors
Statistics: logistic regression units (Y/N (0,1)), probability weighting
and updating, inference for outcome prediction
Calculus: optimization (minimization), gradient descent in back-
propagation to avoid local minima with saddle points
Feed-forward pass (0,1)
1.5
Backward pass to update probabilities per correct guess
.5.5
.5.5.5
1
10
.75
.25
Inference
Guess
Actual
Feature 1
Feature 2
Feature 3
28. 8 Aug 2019
EmergingTech
Learning process
27
Source: http://neuralnetworksanddeeplearning.com/chap2.html
Vary the weights and biases
for improved outcome
Repeat until the net correctly
classifies the data
Edge
Input value = 4
Edge
Input value = 16
Edge
Output value = 20
Node
Operation =
Add
Input Values have
Weights w
Nodes have a
Bias bw1* x1
w2*x2
N+b
.25*4=1
.75*16=12
13+2 15
Input Processing Output Variable Weights and
Biases
Basic Node Structure (fixed) Basic Node with Weights and Bias (variable)
29. 8 Aug 2019
EmergingTech
Image Recognition
Levels of Abstraction Object Recognition
28
Source: Jann LeCun, http://www.pamitc.org/cvpr15/files/lecun-20150610-cvpr-keynote.pdf
Layer 1: Log all features (line, edge, unit of sound)
Layer 2: Identify more complicated features (jaw line,
corner, combination of speech sounds)
Layer 3+: Push features to higher levels of abstraction
until full objects can be recognized
30. 8 Aug 2019
EmergingTech
Image Recognition
Higher Abstractions of Feature Recognition
29
Source: https://adeshpande3.github.io/The-9-Deep-Learning-Papers-You-Need-To-Know-About.html
31. 8 Aug 2019
EmergingTech
Example: NVIDIA Facial Recognition
30
Source: NVIDIA
First hidden layer extracts all possible low-level features
from data (lines, edges, contours); next layers abstract
into more complex features of possible relevance
32. 8 Aug 2019
EmergingTech
Deep Brain Face and Cat Recognition
31
Source: Quoc V. Le et al, Building high-level features using large scale unsupervised learning, 2011, https://arxiv.org/abs/1112.6209
Google
image net
33. 8 Aug 2019
EmergingTech
Speech, Text, Audio Recognition
Sequence-to-sequence Recognition + LSTM
32
Source: Andrew Ng
LSTM: Long Short Term Memory
Technophysics technique: each subsequent layer remembers
data for twice as long (fractal-type model)
The “grocery store” not the “grocery church”
37. 8 Aug 2019
EmergingTech
Loss function optimization
Backpropagation
Problem: Combinatorial complexity
Inefficient to test all possible parameter variations
Solution: Backpropagation (1986 Nature paper)
Optimization method used to calculate the error
contribution of each neuron after a batch of data is
processed
36
Source: http://neuralnetworksanddeeplearning.com/chap2.html
38. 8 Aug 2019
EmergingTech
Gradient Descent
Gradient: derivative to find the minimum of a function
Gradient descent: optimization algorithm to find the
biggest errors (minima) most quickly
Error = MSE, log loss, cross-entropy; e.g.; least correct
predictions to correctly identify data
Technophysics methods: spin glass, simulated
annealing
37
Source: http://briandolhansky.com/blog/2013/9/27/artificial-neural-networks-backpropagation-part-4
39. 8 Aug 2019
EmergingTech
Agenda
Digital Transformation
Deep Learning Neural Networks
Applications
Blockchain Technology
Implications for Fraud
38
40. 8 Aug 2019
EmergingTech
Applications: Cats to Cancer to Cognition
39
Source: Yann LeCun, CVPR 2015 keynote (Computer Vision ), "What's wrong with Deep Learning" http://t.co/nPFlPZzMEJ
Computational imaging: Machine learning for 3D microscopy
https://www.nature.com/nature/journal/v523/n7561/full/523416a.html
41. 8 Aug 2019
EmergingTech
Radiology: Tumor Image Recognition
40
Source: https://www.nature.com/articles/srep24454
Computer-Aided
Diagnosis with
Deep Learning
Breast tissue
lesions in images
Pulmonary nodules
in CT Scans
42. 8 Aug 2019
EmergingTech
Melanoma Image Recognition
41
Source: Nature volume542, pages115–118 (02 February 2017
http://www.nature.com/nature/journal/v542/n7639/full/nature21056.html
2017
43. 8 Aug 2019
EmergingTech
Melanoma Classification
42
Source: https://www.techemergence.com/machine-learning-medical-diagnostics-4-current-applications/
Diagnose skin cancer using deep learning CNNs
Algorithm trained to detect skin cancer (melanoma)
using 130,000 images of skin lesions representing over
2,000 different diseases
44. 8 Aug 2019
EmergingTech
DIY Image Recognition: use Contrast
43
Source: https://developer.clarifai.com/modelshttps://developer.clarifai.com/models
How many orange pixels?
Apple or Orange? Melanoma risk or healthy skin?
Degree of contrast in photo colors?
45. 8 Aug 2019
EmergingTech
Deep Learning World Clinic
WHO estimates 400 million people without access to
essential health services
Earlier stage diagnosis, personalized health clinic
Smartphone-based diagnostic tools with AI for optical
detection and EVA (enhanced visual assessment)
44
Source: http://www.who.int/mediacentre/news/releases/2015/uhc-report/en/
46. 8 Aug 2019
EmergingTech
Agenda
Digital Transformation
Deep Learning Neural Networks
Blockchain Technology
Implications for Fraud
45
48. 8 Aug 2019
EmergingTech 47
Conceptual Definition:
Blockchain is a software protocol;
just as SMTP is a protocol for
sending email, blockchain is a
protocol for sending money
Source: http://www.amazon.com/Bitcoin-Blueprint-New-World-Currency/dp/1491920491
What is Blockchain/Distributed Ledger Tech?
49. 8 Aug 2019
EmergingTech 48
Technical Definition:
A blockchain is a distributed data structure
that is an immutable, cryptographic,
consensus-driven ledger
Blockchain is the tamper-resistant distributed ledger
software underlying cryptocurrencies such as Bitcoin, for
transferring money, financial property, and real estate titles
via the internet without third-party intermediaries
Source: http://www.amazon.com/Bitcoin-Blueprint-New-World-Currency/dp/1491920491
What is Blockchain/Distributed Ledger Tech?
50. 8 Aug 2019
EmergingTech
Blockchain Technology: What is it?
49
Blockchain technology is the secure distributed ledger
software that underlies cryptocurrencies like Bitcoin
Skype is an app for phone calls via Internet without POTS;
Bitcoin is an app for money transfer via Internet without banks
Internet
(decentralized network)
Blockchain
Bitcoin
Source: http://www.amazon.com/Bitcoin-Blueprint-New-World-Currency/dp/1491920491
Application
Layer
Protocol
Layer
Infrastructure
Layer
SMTP
Email
VoIP
Phone
calls
OSI Protocol Stack:
51. 8 Aug 2019
EmergingTech
change.
50
“…financial institutions…face the risk that
payment processing and other services could
be disrupted by technologies, such as
cryptocurrencies, that require no intermediation”
10K, Mar 2018
52. 8 Aug 2019
EmergingTech
Trustless multi-party exchange with software
51
Source: Santander
Institutional functions relocated to execution by
software, not human-based organizations
Blockchain software replaces intermediaries
53. 8 Aug 2019
EmergingTech
internet traffic.
52
information.
email.
voice.
video.
money.
point cloud SLAM.
SLAM: simultaneous localization and mapping, point cloud data captures 3D positioning information about entities (humans, robots,
objects) in the context of their surroundingsc
54. 8 Aug 2019
EmergingTech
How does Bitcoin work?
Use eWallet app to submit transaction
53
Source: https://www.youtube.com/watch?v=t5JGQXCTe3c
Scan recipient’s address
and submit transaction
$ appears in recipient’s eWallet
Wallet has keys not money
Creates PKI Signature address pairs A new PKI signature for each transaction
55. 8 Aug 2019
EmergingTech
P2P network confirms & records transaction
54
Source: https://www.youtube.com/watch?v=t5JGQXCTe3c
Transaction computationally confirmed
Ledger account balances updated
Peer nodes maintain distributed ledger
Transactions submitted to a pool and miners assemble
new batch (block) of transactions each 10 min
Each block includes a cryptographic hash of the last
block, chaining the blocks, hence “Blockchain”
56. 8 Aug 2019
EmergingTech
How robust is the Bitcoin p2p network?
55
p2p: peer to peer; Source: https://bitnodes.21.co, https://github.com/bitcoin/bitcoin
9,501 global nodes run full Bitcoind (7/31/19); 160 gb
Run the software yourself:
58. 8 Aug 2019
EmergingTech
What is Bitcoin mining?
57
Mining is the accounting function to record
transactions (automated and fee-based)
Mining software constantly makes nonce
(number used once) guesses
Rate of 2^32 (4 billion) hashes (guesses)/second
One machine at random guesses a winning
answer
Winning machine confirms and records the
transactions, and collects the rewards
Other nodes confirm the result and append the
new block to their copy of the distributed ledger
“Wasteful” effort deters malicious players
Run the software yourself:
Fast because ASICs
represent the hashing
algorithm as hardware
59. 8 Aug 2019
EmergingTech
How does Bitcoin mining work?
• Problem: Create an internet economic system with untrusted parties
• Solution: Use software based on cryptography, game theory, and
economic incentives to produce trustworthy behavior
• Nodes running the mining software are called "miners"
• Automatically validate and package outstanding transactions into blocks
• Mining software guesses answers to a cryptographic puzzle per
known parameters (part of the Bitcoin software)
• The winning answer is a number that, when combined with the data in the block
and passed through a hash function, produces a result that is within a certain
range (for Bitcoin, an integer between 0 and 4,294,967,296)
• The resulting hash has to start with a pre-established number of zeroes
• Cannot predict a winning number, consecutive integers give different results
First miner to guess within the desired range announces victory
Other miners confirm the answer and add the new block to the chain
58
Source: https://www.coindesk.com/information/how-bitcoin-mining-works
Run the software yourself:
60. 8 Aug 2019
EmergingTech 59
How does Bitcoin mining
work?
https://blockexplorer.com/block/0000000000000000002274a2b1f93c85a489c5d75895e9250ac40f06268fafc0
Difficulty – system set computational number
involving floating point operations,
exponents, integrals
Bitcoin nonce: an
integer between 0 and
4,294,967,296
The Bitcoin hash is created by running the SHA-256 algorithm on six pieces of data:
1. The Bitcoin version number. 2. The previous block hash. 3. The Merkle Root of all the
transactions selected to be in that block. 4. The timestamp. 5. The difficulty target. 6. The Nonce.
Winning
nonce:
869666145
61. 8 Aug 2019
EmergingTech 60
public chains. private chains.
trustless. mined.
p2p software.
trusted. not-mined.
enterprise software.
63. 8 Aug 2019
EmergingTech
Blockchain Applications Areas
62
Source: http://www.blockchaintechnologies.com
Smart Property
Cryptographic
Asset Registries
Smart Contracts
IP Registration
Money, Payments,
Financial Clearing
Identity
Confirmation
Impacting all industries
because allows secure
value transfer in four
application areas
64. 8 Aug 2019
EmergingTech
Agenda
Digital Transformation
Deep Learning Neural Networks
Blockchain Technology
Applications
Implications for Fraud
63
65. 8 Aug 2019
EmergingTech
Global Trade: Maersk 15-carrier blockchain
Digitization = BPR 2.0 for secure
transfer of money and information
15.8% of the world's global shipping
fleet traffic ($236bn value, 628 ships)
On average, 30 people/organizations
involved in the shipment of a product
using a shipping container
Over 200 separate interactions, each
requiring a new set of documents
IBM-Maersk shipping blockchain with
15 carriers (Hyperledger)
Pilot project: Relocate empty containers
to available nearby ships
64
Source: https://www.coindesk.com/worlds-largest-shipping-company-tracking-cargo-blockchain/,
https://www.coindesk.com/ibm-maersk-shipping-blockchain-gains-steam-with-15-carriers-now-on-board
66. 8 Aug 2019
EmergingTech
Supply chain custody and traceability
Traceability system for
materials and products
Bring verified information
from supply chain to point
of sale
Used by 200 suppliers
Convergence of IoT,
mobile, and blockchain
Example
Smart tags used to track
fish caught by fishermen
with verified social
sustainability claims
65
Source: Provenance (NL)
67. 8 Aug 2019
EmergingTech 66
Concept: global inventories of high-value
items: jewels, controlled substances
Mechanism: registered with digital certificate
Diamond supply chain projects:
Everledger (2015)
Records and tracks the immutable provenance of
an asset with blockchain, IoT, smart contracts
TrustChain Initiative (2018)
IBM, precious metals refiner Asahi Refining,
jewelry retailer Helzberg Diamonds, precious
metals supplier LeachGarner, jewelry
manufacturer The Richline Group, and
independent verification service UL
Source: https://diamonds.everledger.io/; https://cointelegraph.com/news/ibm-and-jewelry-industry-leaders-to-use-blockchain-to-
trace-origin-of-diamonds
High-value tracking
68. 8 Aug 2019
EmergingTech
Enterprise Blockchains
Single shared business processes with private views
across the industry value chain
Controlled-use credentials and read/write access
67
Source: Swan, M. (2017.) Anticipating the Economic Benefits of Blockchain. Technology Innovation Management Review.
7(10): 6-13. https://timreview.ca/article/1109
69. 8 Aug 2019
EmergingTech
Enterprise Blockchains: trade finance
Transparency, immutability, auditability, safety
All parties using the same software infrastructure
prevents fraudulent (duplicate) invoices
68
Source: Swan, M. 2018. Blockchain Economics: 'Ripple for ERP' integrated blockchain supply chain ledgers. European Financial
Review. Feb-Mar: 24-7. http://www.europeanfinancialreview.com/?p=21755
70. 8 Aug 2019
EmergingTech
Counterfeit Airbags
Business case
30% global airbags sold and installed are counterfeit
Estimated 3.3% goods sold in the EC are counterfeit
69
Source: https://www.oecd.org/newsroom/trade-in-fake-goods-is-now-33-of-world-trade-and-rising.htm
Solution
Single shared
process for airbag
registration and
lookup
Used by
manufacturers,
vendors, repair
shops, end users
71. 8 Aug 2019
EmergingTech
Health and Pharmaceutical
Electronic Medical Records (EMRs)
Smart contract-based consent
Digital health wallet
Identity credentials + EMR + health
insurance + payment information
Health insurance claims
Automated claims billing
Multi-party value chain
Genomic research
Files too large (20-40 Gb) for centralized
research repositories
Require secure validated access
70
Digital health wallet
Use Case: Factom health
insurance claims billing
• Automated claims billing,
validation, payment, and
settlement
• Multi-party value chain:
patient, service provider,
billing agent, insurance
company, payor,
government, collections
72. 8 Aug 2019
EmergingTech
Agenda
Digital Transformation
Deep Learning Neural Networks
Blockchain Technology
Implications for Fraud
71
73. 8 Aug 2019
EmergingTech 72
the farther future: better horse is a car.
new technology.
better horse “horseless carriage” => car
74. 8 Aug 2019
EmergingTech
risks.
tech: scalability.
political: regulation.
social: adoption.
Rapid
Adoption
Unfavorable
Regulation
Favorable
Regulation
Slow
Adoption
Future Scenarios
73
Status
Quo
Tech
Cold
War
Trustful
Privacy
Regulatory
Arbitrage
75. 8 Aug 2019
EmergingTech
Quantum Computing
When will it be possible to break existing RSA
cryptography standards with quantum computers?
Estimated unlikely within 10 years however methods are
constantly improving
US National Academies of Sciences 2019 report: “highly
unexpected that a quantum computer can compromise
RSA 2048 within the next decade”
74
Source: Quantum Computing: Progress and Prospects (2019), The National Academies Press,
https://www.nap.edu/catalog/25196/quantum-computing-progress-and-prospects
Status: quantum computers commercially
available from IBM, D-WAVE Systems, Rigetti
76. 8 Aug 2019
EmergingTech
Fraud
Law enforcement argument: criminality deploys in new
technologies and so too must law enforcement
Example: Silk road
75
77. 8 Aug 2019
EmergingTech
Fraud Detection
Corruption
Transparent process, private data
Cross-border trade error and
malfeasance
Single-shared ledger, business
processes
Counterfeiting and product
traceability
Anomaly detection with
statistical distributions
Machine learning
Quantum computing
76
78. 8 Aug 2019
EmergingTech
Fraud detection leading the path ahead
77
Expertise in organizational,
computational, behavioral,
and psychological cues
Global reach, sophisticated
business, technologically-
intense solutions, real-time
detection methods
Challenge is to envision and
continue modernizing the
way forward with fraud
detection strategies
79. 8 Aug 2019
EmergingTech
Conclusion
• Deep learning is not merely an
AI technique or a software
program, but a new class of
smart network information
technology that is changing the
concept of the modern
technology project by offering
real-time engagement with
reality
• Deep learning is a data
automation method that
replaces hard-coded software
with a capacity, in the form of a
learning network that is trained
to perform a task
78
Conclusion
Deep learning is a class of
machine learning algorithms in
the form of a neural network
that uses a cascade of layers of
processing units to extract
features from data sets in order
to make predictive guesses
about new data
A blockchain is a distributed
data structure that is an
immutable, cryptographic,
consensus-driven ledger
80. 8 Aug 2019
EmergingTech
Smart Network Thesis
79
Considering high-impact emerging
technologies (AI machine learning and
blockchain) together suggests the emergence
of a new class of global computational
infrastructure: smart networks
(Smart networks: intelligent self-operating computation
networks such as deep learning neural nets,
blockchains, UAV fleets, industrial robotics cloudminds)
81. Melanie Swan
Purdue University
Emerging Technologies
shaping the future of
Fraud Detection, Banking, and Finance
Association of Certified Fraud Examiners
Indianapolis IN, August 8, 2019
Slides: http://slideshare.net/LaBlogga
Thank you!
Questions?