1st AAAI Workshop on Integration of Learning and Diagnosis (WILD 2016)

About

There is promising opportunity in overlapping model-based reasoning and data-driven learning approaches. In model-based reasoning, a model describes the behavior of an underlying system. In data-driven machine learning, a model describes interdependence of properties of some entity that may or may not come from an underlying system.

The 2016 AAAI Workshop on Integration of Learning and Diagnosis (WILD) focuses on an interdisciplinary research that integrates machine learning and model-based diagnosis. This is a subarea of the larger overlap of model-based reasoning and data-driven machine learning, particularly in the context of diagnosis. Research in diagnosis focuses on identifying the root causes for encountered issues.

The workshop aims to encourage interaction and collaboration across researchers and practitioners with diverse backgrounds including artificial intelligence (AI), machine learning, model-based diagnosis, control theory, cyber security and software engineering. The goal is to leverage a diverse set of expertise to combine model-based diagnosis and machine learning to not only increase accuracy but also extend capabilities by solving problems formerly not conducive to analytic approaches.

Important dates

  • Paper submission
    October 30, 2015
  • Notification of acceptance
    November 23, 2015
  • Camera-Ready copy due to AAAI
    December 7, 2015
  • Workshop
    February 12-13, 2016

Location

Phoenix Convention Center
100 N 3rd St
Phoenix, Arizona, USA

For detailed information about WILD location visit AAAI 2016 Website.

Organization

Co-Chairs

Program

TBD

Topics

We are looking forward to submissions on topics that combine and/or contrast machine learning and diagnosis problems and challenges, including papers focusing on the following issues:

  • Automatic learning of behavioral models by characterizing knowledge from data.
  • Applying machine learning to construct behavioral models.
  • Leveraging machine learning results to tune the parameters of the already known (first principles) physics model.
  • Devising diagnosis techniques to improve and explain machine learning results.
  • Identifying challenges of applying existing machine learning techniques for diagnosis problems.
  • Exploring directions for a hybrid model-based and data-driven approach for other application domains such as cyber security and condition-based maintenance.

Submission

One author of each paper submitted must be committed to attend the workshop and present the paper.

Papers should be no longer than 7 pages, describe a novel contribution to the integration learning and diagnosis, and should carefully support claims of novelty with citations to the relevant literature.

All paper categories must conform to the two-column AAAI conference publication format and must be submitted in pdf format.

Submissions will be handled via EasyChair.