![](../images/_general/nice.jpg) |
The Fifth International Conference on Business
Intelligence and Technology
BUSTECH 2015
March 22 - 27, 2015 - Nice, France |
T1.
Applications of Machine Learning to Software Testing
Marcelo De Barros, Microsoft Corporation, USA
T2.
Modern Engineering Principles for Large Scale Teams and Services
Marcelo De Barros, Microsoft Corporation, USA
T3.
Artificial Intelligence of Humor: On Computational Humor
Victor Raskin, Purdue University - W. Lafayette, USA
Julia M. Taylor, Purdue University - W. Lafayette, USA
Detailed Description
T1.
Applications of Machine Learning to Software Testing
Marcelo De Barros, Microsoft Corporation, USA
Software Testing is evolving. Long gone are the days when to test a software all that was needed was an input and an expected output. Software Testing is shifting towards anomalies detection, quantification and prevention at large-scale and highly dynamic environments. In this tutorial Marcelo De Barros will talk about how to identify software testing scenarios in which Machine Learning can be successfully applied. Marcelo will describe real stories of how well-established and novel Machine Learning algorithms such as Clustering, Neural Networks and Markov Chains were applied towards detection, quantification and prevention of software anomalies, saving lots of time (and money) to the company.
T2.
Modern Engineering Principles for Large Scale Teams and Services
Marcelo De Barros, Microsoft Corporation, USA
In the online world, velocity becomes imperative for success. At Bing, we ship our code to production multiple times a day. Since we're always trying different experiments, we need to ensure that every build meets the desired quality level. Hence we have these two variables that are inversely proportional: time to ship an experiment (minimized) versus time to validate an experiment across multiple platforms (maximized). To achieve the goal of shipping the entire code base daily maintaining high quality several concepts had to be introduced and/or redefined such as high reliability of automation, distribution of test automation, scalability of test automation to multiple platforms, shift to test in production, utilization of production traffic (forking) for validation purposes, and so on. Cultural changes were also at the center of this transformation. In this talk, well go into the details of this transformation that allowed our team to become a super-agile-high-quality organization.
T3.
Artificial Intelligence of Humor: On Computational Humor
Victor Raskin, Purdue University - W. Lafayette, USA
Julia M. Taylor, Purdue University - W. Lafayette, USA
Goal
As the role of humor in social computing and the need for the computer to detect and generate humor is increasing in robotic and agent intelligence becomes evident, it is important to understand that verbal humor intelligence is part of AI, and as such, is formalizable and computable. The formal linguistic theory of humor, while offering an insight into humor structure, provides a basis for formalizing and computing jokes in meaning-based AI applications. The target audience at IARIA 2015 is those AI, CS, and related areas researchers who are interested in the seemingly unformalizable phenomena and willing to overcome their “fear of semantics.”
History
This tutorial has been preceded by several (2005-2012) tutorials at the International Conferences in Humor Research for a multidisciplinary but largely non-technical audience and at WorldComp 2012-2014 for a largely non-humor-oriented advanced technical audiences. The proposed tutorial has had no exact precedent in this form, but we did organize a 2012 AAAI Fall Symposium under the same title.
Content
The tutorial will follow the outline below but will be quite open to the audience’s questions which may lead to various extensions:
- Introduction into a formal study of humor
- Theories of humor
- Linguistic theories of humor
- Formal theories of humor
- Computational humor
- Semantic approach
- Statistical approach
- Hybrid approach
- Humor intelligence as part of AI
- With-Humor Applications
- Companion dialog applications
- Social media analysis (including real “sentiment analysis”)
- Humor-based stylometry
Description
The tutorial will review the theoretical and computational humor research to date and introduce a number of simple to advanced AI applications. It is intended for 4 hours, including 1+ hours for questions and discussion.
Prerequisite
An interest in and some preparation in AI and/or its contributing disciplines is desirable but the tutorial is planned as largely self-explanatory.
Credentials
Victor Raskin
Linguistics/CERIAS/Computer Science
Purdue University
500 Oval Drive
W. Lafayette, IN 47907-2038
765-409-0675
Julia M. Taylor
Computer and Information Technology/CERIAS
Purdue University
401 N. Grant Street
W. Lafayette, IN 47907-2011
765-494-9525