Center for Future Banking Network Economies Research Review November 10th, 2008 Confidential 1057 blogs 2007 French Presidential Election Video links: MS future vision on retail banking ING Living Tomorrow - Retail Banking 2015 Banking Retrospective Presentation by: Ray Garcia
Research Review on Network Economies. Includes an overview of the research agenda for CFB and explanation of innovation.
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Information flow is the signal versus and noise of information science. The value is an increase or decrease in signal. Information is the lowest atomic unit of measure for our research. The flows form interdependent chains or graph of relationships. The flows of value are not exclusive to money and include any convertible value.
Behavioral Economics
Includes any behaviors that happen before and after the decision of a consumer or producer. Behavior is observed, modeled, anticipated, projected, predicted as well as all the vagaries of the human condition. The unit of measure is at human scale and includes many uncontrolled factors.
Identity, Trust Privacy, Security
Identity includes concepts of privacy or public disclosure. Trust implies the concepts of gradations of security or no security if full trust is granted. Both included measures of credibility and honesty or value systems that are human. The concepts span system processes and human interaction.
NetworkEconomies
Networks are the systems that connect and the people acting in social interactions. Information flows, social behavior, identity and trust are aggregated into economic interactions within a network.
Social Responsibility
Includes individual actions representing corporations as well their household and raises questions of ethics. This aggregates the network economies into defined groups that care for direct and indirect impact and consequences of decisions.
* joined by a team of colleagues, researchers, students, support staff, bank line of business, committee
How will the next generation of information clouds emerge?
How do we de-mystify financial management?
Living Lab / Flagship store
Retail Banking Analytics
Sensing Stores / Speechome Video
Ambient Intelligence / Surface Computing/ Siftables /Word play
Affective Computing
Information flow is the signal versus and noise of information science. The value is an increase or decrease in signal. Information is the lowest atomic unit of measure for our research. The flows form interdependent chains or graph of relationships. The flows of value are not exclusive to money and include any convertible value.
Includes any behaviors that happen before and after the decision of a consumer or producer. Behavioral is observed, modeled, anticipated, projected, predicted as well as all the vagaries of the human condition. The unit of measure is at human scale and includes many uncontrolled factors.
What are the new economic models for trust and security?
Living Labs / New deal for Data
Identity and Identification Models
Personalized Customer Service
Trusted Investment Advisor
Face Reader for affective responsive computing
Identity includes concepts of privacy or public disclosure or fabricated personas which are both public and private. Trust implies the concepts of gradations of security or no security if full trust is granted. Both included measures of credibility and honesty or value systems that are human. The concepts span system processes and human interaction.
Networks are the systems that connect and the people acting in social interactions. Information flows, social behavior, identity and trust are aggregated into economic interactions within a network.
Financial Life Time household Literacy and decision making
Account Ability and sensible consumerism
Next Billion in our Neighborhood / Living Labs
Includes individual actions representing corporations as well their household and raises questions of ethics. This aggregates the network economies into defined groups that care for direct and indirect impact and consequences of decisions.
CFB Approach: Design Research and Business Model Development
InformationValue chain
Network Economies
Customer Needs Observed (IDEO & Design Research)
BehavioralEconomics
Identity/TrustPrivacy/Security
Social Responsibility
TANGIBLE COMMON SENSE CLARITY•help people grasp the bigger financial picture•Go beyond a menu to a recipe; “I need expert financial guidance”•Make financial information more understandable and tangible instead of using abstract concepts.
WISDOM OF CROWDS•Enable people to use collective knowledge of fellow customers•Foster reassurance by connecting like-minded customers
CULTURE OF ACHIEVING•Create moments of pride and mementos of achievement. “Account alerts are embarrassing – they make me feel poor” •Set an atmosphere of encouragement
MY PACE, MY ROUTINE•Infer preferences: “Amazon knows me better than my bank” •Know and accommodate people’s schedules and daily patterns.
TransformativeBusiness Model
-------------------------Outside-In
Customer Analysis / Research
Inside-Out Competitive Analysis
Compelling Offer
Finance ModelDISTRUST•Trust is difficult to attain, fragile, and can be easily eroded. When trust is broken, we are often the last to know.
How do networks emerge and scale? How are physical and virtual different?
What is the inherent value system and constraints under which they form?
How is knowledge shared and problems solved within social networks.
Do SmartMobs behave predictably? Can they be controlled?
How would augmented reality retail emerge with social networks?
Network Economic Models
What are the micro-economic forces at play?
How do new currencies impact group decisions?
New currencies could be coupons, loyalty point systems, barter systems, time trading systems, money within games, micro/nano payments, one time use codes, media capture and exchanges, peer to peer value exchanges.
How do networks behave in physical spaces when using mobile Location based services and can form spontaneous connections?
What are Network Economies Value exchanges based on relationships that transcend a single transaction Frictionless matching of producers and consumers, freescale growth Communities of Practice with convertible social capital Rebalancing of power asymmetries and disequilibrium
Enablers of Network Economies Internet and ICT Mobility, cell phones, laptops Software that affords self organizing activities
US Industry Examples using Social Networking Google, EBay, Amazon, Paypal, Skype, Joost Intuit Small Business Community Facebook - Visa Business Network Facebook - Fiserv MyMoney program Prosper Peer to Peer Lending
Network Economies – How does it work? a research history
6 (5.5 – 6.6) Degrees of separation
In 1960’s S. Milgram (psychologist) demonstrates 6 degrees using US postal mail
In 1973 M. Granovetter (sociologist) theory “Strength of Weak Ties” spread of info in networks
In 1994 R. Reynolds paper on Cultural Algorithms using computational models
Online 3 degrees of separation is more likely
A. Barabasi (physics) demonstrates preferential attachment model, nodes with lots of links have higher probability of acquiring more links.
Yahoo research discovers growth in members and connections are needed for healthy networks to thrive. If either stall the network dies.
R. Dunbar (anthropologist) argues a maximum of 150 social relationships can be maintained which is the historical size of farming villages.
The world is highly clustered geographically and socially, enabling short paths traversals within networks and across groups
Social Contagion, diffusion of innovation, people imitate each other 0 1 2 effect where the probability of imitation is great once 2 friends have done something
Jared Diamond posits a theory for Societal Collapse inferring social networking theory
Network Economies – How does it work? Current research finding Nov. ‘08
mathematical model “hidden metric space”
may explain the “small-world phenomenon”
relates to man-made and natural networks
human language to gene regulation
neural networks connecting neurons to organs and muscles within our bodies.
Natural world routing only uses local knowledge and not global knowledge of the network which is how communications routing works today.
Natural networks transmit information very efficiently without any single node having knowledge of the structure of the entire network
Many complex networks share similar shapes that maximize their communications
Implications range from cancer research on gene therapy to more efficient routing on the internet.
How the hidden metric space guides communication. If node A wants to reach node F, it checks the hidden distances between F and its two neighbors B and C. Distance CF (green dashed line) is smaller than BF (red dashed line), therefore A forwards information to C. Node C then performs similar calculations and selects its neighbor D as the next hop on the path to F. Node D is directly connected to F. The result is path ACDF shown by green edges in the observable topology.
Inquiry – Examination of associate interactions with peers, managers and customers, and activity analysis for predicting group behavior by including all available communications traces.
Inputs – Frontline call center interaction, work flows, email and phone data, including socio-meter instrumentation.
Outputs - group behavior analytics, models for successful interactions across multiple channels, define behavioral characteristics that are critical to driving success.
Academic Team
Faculty - Sandy Pentland
Students - Ben Waber, Coco Krumme, Anmol Madan, Iolanthe Chronis
Oct 21st Design Charette, 70 people, Faculty, Students, BAC, City of Boston, guest speakers
Event stimulated research possibilities not previously considered
Video Documentaries available for Leadership Learning.
Highlights of Projects Initiated:
Living Labs project initiated with Mobile Initiative, Sandy Pentland
City of Boston engaged with CFB, MIT, BAC in Living Labs project
SBOC Data analysis – discovery of internal competencies in data and text analysis prior to engaging MIT students.
SBOC used in the EpiCenter project – connecting the virtual with the physical for richer engagement with customers.
Peer to Peer Lending Analysis – pattern recognition class doing analysis, possible academic paper with significant finding pending. Ray Garcia with Dawei Shen and Hyungil Ahn
Team and Inquiry – Information Economics of Network Knowledge
Bank Strategic Champions – Laurie Readhead and Lance Drummond
Business Tactical Leaders – Margaret Weichert, Ross Feldman, Beverly Ladley, Indur Koul
Innovators Committee – CFB - Hans Schumacher, Todd Inskeep; BAC - Marc Keller, Matt Calman
Quant Committee – David Joffe and David Joa
Audience – Quant doing Informatics and business leaders needing to understand information economics, network and emergence theory.
Business Inquiry –
How might data and information captured within Bank systems be converted to knowledge that informs a Financial Ecology and Network Relationships such that they can be leveraged as residual products.
How would the Economics of Information assess value or reveal insights to be discovered to inform service innovations, generate revenue, reduce risk, or improve operations?
How does value emerge from a network and how would Quants use Informatics to continuously discover emerging opportunities and threats given the severe challenges of data quality management.
Research Inquiry –
How do personal, social, commercial, knowledge networks emerge?
What is the inherent value system and constraints under which networks form?
How does using information economics to show how fairness contributes to efficiency and innovation?
How is knowledge elicited, assessed, shared and problems solved within networks?
What are the micro-economic forces within information intensive service businesses?
What experimental methods are used to model, simulate, predict, inform, improve, knowledge products?
What are the learning models, cognitive models, and affective models that influence knowledge economies?
Research Design - Information Economics of Network Knowledge
Theoretical Basis – Information Economics, Social Science, Computation, Information Science, Ontology
Business Basis – Financial Ecology, Human Dynamics, Knowledge Management, Social Media
Methods – Qualitative Research methods, quantitative analysis, ethnographies
Techniques - data processing, statistical analysis, pattern recognition, text processing, machine learning
Business Inputs – Bank transaction data, data from information network experimentation, marketing research databases, external market data sources, financial data feeds, expert annotation
Research Inputs – MIT prototypes in; common sense knowledge, social media, cognitive/affective markets, knowledge markets
Experimental Platform – communications market borrowing from several MIT projects and creating a system for on-going long term research.
Informatics Toolkit – an assembly of open source, or low cost, tools used in the analysis and visualizations of data.
Key Outputs –
Valuation of latent Network Knowledge factoring system dynamics, content, membership, relations between nodes.
Knowledge elicitation, discovery, and dissemination, modeling, simulation and prediction of human systems including cultural algorithms which inform market dynamics.
A formulation of the research discipline in Service Science merging techniques from Qualitative and Quantitative research borrowing from various practices. This Service Science would determine how to convert research findings into innovative services that can be formulated, tested, disseminated within a large scale global company.
The Network Knowledge and information economics would include a synergistic co-generated service innovation between consumers and producers. This may include a communications marketplace that is both internal and external and reconciles the need for knowledge elicitation and dissemination.
Quant tools, techniques, and methods for applied information economics against bank data.
Academic Team (proposed)
Research Scientist – Marshall Van Alstyne (Sloan School), Kelly Hewett (BAC Market Researcher)
Related Research to Information Economics of Network Knowledge
Economic Choice Theory and Group Decision Making
Problem Statement – How do people learn to make important economic decisions that impact their personal finances and/or company value from the social engagements and information sources they have available? What are the cultural forces at work? What methods could be used to experiment, model, simulate, predict, inform, improve, group decision making within tribes, co-workers, and communities of practice? How does economic literacy impact other important decisions such as health, home, children, work, and consumer behavior? How does economic literacy change throughout ones life and how is it inter-generational? How is economic literacy taught and learned. What are the educational psychological models and cognitive models that help inform personal and group decisions and how do social values systems influence the choices and expectations.
Scaling the Service Innovation Process to create breakthrough new products (Cooperative Collective Intelligence Elicitation and Dissemination)
Problem Statement – Service companies lack the experimental research activities that are prevalent in product companies and therefore do not have a science of service to create and build a body of knowledge. The research methods and process for converting research findings into innovative services requires a discipline to be formulated, tested, and disseminated to be effective within a large scale global company. Can these service innovations be co-generated between the consumer and producer such that they evolve into a synergistic relationship? Using Social Media to elicit knowledge and applying methods for evolving the knowledge into innovations while opening communications to a free exchange of ideas between consumer and producers is a challenge for industries where confidentiality and proprietary information is consider a risk to the business and therefore all communications is controlled. How to balance the economic benefits of a free communications with the desire for control requires research in the economics of information to devise techniques to resolve these conflicting goals.
Personal Augmented Reality in a Retail Environment
Inventors: Ray Garcia and Stephen Miles
An intelligent personalized agent running on a mobile wireless connected device that monitors and advises a user in recommendation-making process in a retail setting through both private and public information services, whereby the identification of the mobile device and rfid transponder impacts both the personal display and the public display
Conceptual Ideas under development
Multi-function card associate with a mobile device that acts as a dual factor authentication using a transparent card, embedded image that is reveal through matching with the display on the phone. The card is anonymous with no markers. The information on the card includes dual magstripes, barcode, optical storage, contactless rfid with a coil.
Communications market which is a hyper network of knowledge and affective associations of constructed information amongst a large group of participants. The knowledge is hyper connect to the people, context of situation and circumstance, and has a financial ecology that expands in a free market with minimal regulatory rules. Small world scale free networks of efficient information economies are allowed to emerge.
Pattern Recognition Analysis of Peer to Peer Lending data from Prosper
From Sergio, Rahul, and Aithne, they are exploring the effects of social capital:
1. Social profile Friend: number of 1st-degree friends, number of 2nd-degree friends Endorsement: Endorsement number Group: Group leader reward rate, Group Rating, Group Size
2. Social Interaction - Bids from friends and group members
3. Classifiers were built to determine whether a list could turn into a loan.
4. Social factors are not determinant, but they do increase the chance of getting a loan when all other financial features are similar. This is demonstrated clearly by using clustering.
5. Users do not have time to maintain many social profiles, it's beneficial to use existing social networks as a foundation.
From Charlie, Matt, Ernesto, and Coco:
1. A Bayes belief network and a decision tree are formulated for classification. The unique benefit of Bayes network and decision tree is that it graphically presents some good practice for users if they try to increase their odds to get a loan.
2. Textual information, such as words used in endorsement, description, etc. is analyzed and proved to be very influential.
3. A risk assessment model built on Hidden Markov Model is formulated, and works very well. It can effectively predict a loan carrier's financial health, and predict potential delinquency.
4. Coco starts some interesting games with the image posted by users.
Correlation of speed of Information Diffussion via social networks with productivity
Internal knowledge markets - Information economics applied to quantifying the effects of knowledge management using price theory, information asymmetry, and network theory. The goal is to discover information behaviors that predict success.
Anti-spam & Malware Research - A formal proof of information economics impact on eliminating spam without using a filter.
Problem Description: A next generation social network that is distributed and provides users complete ownership over their data and social connections. Contrary to centralized social networking systems like facebook, each user maintains their personal information and their list of friends on their own private web space. The list of their friends are pointers to other web spaces, while public key infrastructure is used to provide different levels of access to different friends, while providing minimum or no information to strangers.
Experience Sharing Market for Forecasting Marketplace Success
Team: Hyungil Ahn
Problem Description: We develop a novel market game that harnesses people's collective perceptions and experience sharing to forecast the success or failure of new items (products / services / UI designs, etc). Companies can register their new items on this market (as a test bed) to ask people's collective opinion. In each trial session, a participant makes his or her own best prediction on other people's overall opinion about the new items to get incentives (e.g., real opportunities to experience the items) and have fun in gambling-like games. As a participants guess (or portfolio) approaches the collective guess of all participants, he or she has a greater chance of winning an incentive. Participants improve the accuracy of their next prediction by sharing their experiences. As participants have more trial sessions, their collective prediction converges into one common opinion (forecasting the success or failure of new items).
Funk 2
Team: Bo Morgan
Problem Description: Funk2 is a novel process description language that keeps track of everything that it does. Remembering these causal execution traces allows parallel threads to reflect, recognize, and react to the history and status of other threads. Novel forms of complex, adaptive, nonlinear control algorithms can be written in the Funk2 programming language. Currently, Funk2 is implemented to take advantage of distributed grid processors consisting of a heterogeneous network of computers, so that hundreds of thousands of parallel threads can be run concurrently, each using many gigabytes of memory. Funk2 is inspired by Marvin Minsky's Critic-Selector theory of human cognitive reflection, and is the foundation for the Neural Models of Mind project.
MIT Projects to License for use in future research
Selectricity
Team: Chris Csikszentmihalyi, Alyssa Wright, Benjamin Mako Hill
Problem Description: Selectricity is a web-based voting system that supports anonymous and voter-verifiable balloting, and includes an election-methods library that implements a variety of election techniques, includeing several preferential systems. Unlike most voting projects, Selectricity does not attempt to address the issues raised in mainstream political elections. Instead, it provides a simple set of tools that small groups and organizations can use to incorporate computationally complex decision-making into new areas, and for purposes where they ordinarily would find such decision-making into new areas, and for purposes where they ordinarily would find such decision-making prohibitively complex. By supporting a variety of election methods, it provides a way for users to explore and compare the effects of different voting systems and, ultimately, come to better decisions.
Common Sense Reasoning
Team: Henry Leiberman, Catherine Havasi, Robert Spear, Dustin Smith, Jayant Krishnamurthy
Collecting Common Sense – the open mind common sense project, acquiring knowledge from untrained people through the use of on-line interfaces and games.
Common Sense Recommendations – that are more user friendly than collaborative filtering systems. Uses tools to build intelligent recommendation agents and effective product exploration tools
Not-So-Common Sense – infusing data sets with common sense
Perspective Space – discovering distinct communities of people with jargon and belief structures from simple ratings
Analogy Space
Common Sense Investing
Common consensus: a game for collecting commonsense goals
MIT Project techniques to re-build in future research for new inquiry
Behavior Capture from Thousands of People On-line
Team: Jeff Orkin, Deb Roy
Restaurant game simulation of patron to host interaction to capture dialogs and sequences of common actions. The analysis is statistical and produces dialog that mimics frequent occurrences of phrases.
New Media Medicine
Team: Frank Moss, John More
Collaborhythm – collaborative decision making between doctor and patient
Collective Discovery – a massive collection of “everyday experiments”
HealthMap – real-time disease outbreak tracking and visualization system
I’m Listening – pre-visit interviews to help categorize patient symptoms for efficient diagnosis.
Emerging context-enriched services will use location, presence, social attributes, and other environmental information to anticipate an end-user's immediate needs, offering more-sophisticated, situation-aware and usable functions.
By 2012 estimates:
More than 7.3 billion networked devices worldwide
298 million subscribers of location-based services
>75% of new search installations will include social search element
• $150 billion of $1.8 trillion global telecom spending will shift from services to applications
• Global market potential that context-aware computing can impact: $215 billion
MVE algorithm plot on citywide activitySource: Sense Networks
Big Company (statements heard in my first 90 days at BAC)
STAGNANT “We own the market we are just that big!”CYNICAL “It will never happen!”CONFUSED “What is going on in the market?”TOO LATE “How did that happen?”
Future Proof Company (why I am here)
THINK “Service as a Science”LISTEN “Co-create the future with the consumer”SERVE “Mass preference and configuration”SCALE “Human dynamics scale”ADAPT “Grow, split, shift, evolve”