Ross Gayler

Credit scoring:
Applied statistical modeller working on operational prediction of customer behaviour, especially in consumer finance. Employed since 1989 by vendor and client organisations for project work and R&D of products and services. Particular expertise in profitability modelling, customer responses to lender actions, and optimisation of lender policies. I use a range of non-standard modelling techniques but always hold pragmatism as more important than technical virtuosity when applied to systems that make millions of automated decisions.

Cognitive science:
Independent researcher developing a practical, implementable, connectionist architecture for compositional memory. This is concerned with the ability to recognise novel situations and objects in terms of the novel pattern of structural relationships between their familiar component parts. This work effectively treats analogy as a primitive capability of memory. Current standard machine-learning techniques have limited capacity to deal with patterns of relationships and consequently have difficulty recognising novel configurations of familiar components or recognising familiar patterns of relationship when the components have been changed. If successful, this work will have fundamental implications for cognitive science.

Ross Gayler’s Specialties:

Credit scoring, Application scoring, Behaviour scoring, Behavior scoring, Customer scoring, Collection scoring, Profit scoring, Profit optimisation, Response modeling, Customer retention, Fraud scoring, Identity fraud, Basel II, Basel 2

Analogical reasoning, High-dimensional vectors, Compositional connectionism, Vector Symbolic Architecture, Holographic Reduced Representation, Binary Spatter Code, Cleanup memory, Clean-up memory, Map Seeking Circuit, Graph isomorphism, Graph matching

Ross Gayler’s Experience

Honorary Associate

La Trobe University

(Educational Institution; Higher Education industry)

March 2009 — Present (5 months)

Academic affiliation with the School of Communication, Arts and Critical Enquiry in the Faculty of Humanities and Social Sciences in support of my cognitive science research.
Senior R&D Consultant

Veda Advantage

(Privately Held; Financial Services industry)

August 1998 — Present (11 years)

Act as a repository of technical expertise on statistical predictive modelling of customer behaviour in consumer finance and marketing. Directly apply that expertise to non-standard client projects and internal product development. Also provide technical advice to external and internal projects.

(Allowing for renamings, mergers, acquisitions, parent and subsidiary companies you might also know my employer as: Baycorp Advantage, Data Advantage, Equigen, Equigen Consulting, CRAA, Credit Reference Association of Australia)
Independent Researcher, cognitive science

self employed

(Research industry)

January 1987 — Present (22 years 7 months)

Cognitive science research on Vector Symbolic Architectures (very high-dimensional connectionist systems that are well suited to tasks usually construed as symbolic). Attempting to develop a practical, implementable connectionist architecture for compositional memory. Places analogy as a primitive property of memory.

I co-organised the AAAI 2004 Fall Symposium on Compositional Connectionism in Cognitive Science (http://www.cs.wlu.edu/~levy/aaai04/).
Honorary Associate

La Trobe University

(Educational Institution; 1001-5000 employees; Higher Education industry)

October 2003 — December 2006 (3 years 3 months)

Academic affiliation with the School of Communication, Arts and Critical Enquiry in the Faculty of Humanities and Social Sciences in support of my cognitive science research.
Honorary Senior Fellow

University of Melbourne

(Educational Institution; 1001-5000 employees; Higher Education industry)

January 1996 — December 2002 (7 years)

Academic affiliation with the School of Behavioural Science in support of my cognitive science research.
Manager, Advanced Modelling

GE Capital

(Public Company; Financial Services industry)

May 1997 — August 1998 (1 year 4 months)

Ran a 2 person unit to provide advanced modelling support to front-line modellers and work on non-standard projects. Examples of non-standard projects: Developed a model to predict fraudulent applications for credit that was able to be implemented in existing systems with minimal effort; Developed a method for improving the performance of old predictive models that eliminated the need for system changes; Developed models of the evolution of portfolio losses; Developed models of risk, revenue, and responsiveness for targeted marketing to launch a new credit card product; Developed predictive models and strategies for increasing the profitability of debt collection.
Decision Methodologies Manager

ANZ

(Public Company; Banking industry)

October 1995 — May 1997 (1 year 8 months)

Established the credit scoring unit at ANZ. Recruited and trained the first cohort of modellers.
Business Consultant

Experian

(Public Company; 201-500 employees; Information Services industry)

January 1992 — October 1995 (3 years 10 months)

Experian was CCN Asia Pacific at the time.
Development of statistical predictive models of default risk in consumer finance. Specification of systems for automated customer management. Methodological R&D for the CCN Group.
Senior Knowledge Engineer

HiSoft Solutions

(Privately Held; 1-10 employees; Computer Software industry)

July 1989 — November 1991 (2 years 5 months)

Development of expert systems (rule-based) for consumer lending.
Research Scientist

Queensland Department of Health

(Government Agency; Public Policy industry)

June 1982 — July 1989 (7 years 2 months)

Statistical and computing consulting to a health policy research group.
Systems Programmer

University of Queensland

(Educational Institution; Higher Education industry)

January 1981 — December 1981 (1 year)

Systems programmer and manager at Psychology Department. Installed 2nd Unix system in Queensland (to the best of my knowledge). Designed departmental computing strategy covering real-time experimental control and data collection, data analysis, document preparation and general computing.

Ross Gayler’s Education

The University of Queensland

PhD , Psychology , 1978 — 1987

I was enrolled as a part-time student.
Thesis: Development of a methodology and theoretical framework for melodic discrimination (1988)
This included an extension of statistical modelling based on the Theory of Signal Detection, which turned out to be remarkably useful to my credit scoring career.

Activities and Societies:

Queensland University Musical Society
The University of Queensland

BSc (Hons) I , Psychology , 1977 — 1977

My Honours thesis research was into neural feature detector models of speech perception.

Activities and Societies:

Queensland University Musical Society
The University of Queensland

BSc , Psychology, Computer Science , 1974 — 1976

When possible I chose a mixture of subjects that approximated toward an Artificial Intelligence course. In psychology I studied subjects such as perception, psychophysics, cognition, psycholinguistics, mathematical psychology, and statistical methodology.

Activities and Societies:

Queensland University Musical Society
MLSS 08

Additional Information

Ross Gayler’s Websites:

Ross Gayler’s Interests:

consumer finance, customer behaviour, customer behavior, credit scoring, profitability modelling, profitability modeling, customer management, CRM, strategy optimisation, application scoring, behaviour scoring, behavior scoring, customer scoring, collection scoring, profit scoring, profit optimisation, response modelling, response modeling, customer retention, customer segmentation, fraud scoring, identity fraud, identity verification, Basel II, Basel 2, cognitive science, connectionism, connectionist, neural network, neural networks, compositional memory, structural learning, relational learning, analogy, compositionality, systematicity, analogical reasoning, high-dimensional vectors, compositional connectionism, distributed representation, distributed representations, Vector Symbolic Architecture, VSA, Holographic Reduced Representation, HRR, Binary Spatter Code, cleanup memory, clean-up memory, Map Seeking Circuit,graph isomorphism,graph matching,Predictive State Representation