MotivationTraditional supervised learning methods typically require the training data are fully labeled. Nowadays, the data size increases with an unprecedented speed. Fully labeled data becomes infeasible in many real situations, and consequently incomplete labeled data (or data with weak supervision) is ubiquitously existed. For years various approaches have been developed to learn with weak supervision, and learning from big data with weak supervision is showing its superiority to learning with fully labeled yet small data.
However, there are still many open problems and in recent years many interesting challenges have been realized. For example, safe semi-supervised learning that prevents unlabeled data hurting the performance is desired; developing data-adaptive active learning strategies have not fully touched; effective partial label learning in the presence of class imbalance data; deriving high quality labels from noisy crowds; borrowing supervision from auxiliary sources, etc.
ScopeThe main goal of this session is to provide a forum for researchers in this field to share the latest advantages in theories, algorithms, and applications on learning with incompletely labeled data. Authors are invited to submit their original work on learning with incompletely labeled data. The topics of interest include, but are not limited to:
- Semi-supervised learning
- Active learning
- Partial label learning
- Multi-instance learning
- Multi-label learning
- Multi-instance multi-label learning
- Learning with noisy labels
- Weak label learning
- Transfer learning
- Zero-shot learning
- Scalable or efficient learning algorithms
Special Session ChairsYu-Feng Li, (Nanjing University) firstname.lastname@example.org
Sheng-Jun Huang, (Nanjing University of Aeronautics and Astronautics) email@example.com
Min-Ling Zhang, (Southeast University) firstname.lastname@example.org
Paper SubmissionPlease read the following paper submission guidelines before submitting your papers:
All special session papers must be submitted through the IEEE WCCI 2016 online submission system and please select respective special session title under the list of research topics in the submission system. Our special session is Advanced Supervised Learning Techniques and Its Application. Such decision will be made by the Special Session Organizers in consultation with the Special Session Chair.
To help ensure correct formatting, please use the style files for U.S. Letter (http://www.ieee.org/conferences_events/conferences/publishing/templates.html) as template for your submission. These include LaTeX and Word.
Only papers prepared in PDF format will be accepted.
Paper Length: Up to 8 pages, including figures, tables and references. At maximum, two additional pages are permitted with overlength page charge of US$125/page, to be paid during author registration.
Paper Formatting: double column, single spaced, #10 point Times Roman font.Margins: Left, Right, and Bottom: 0.75" (19mm). The top margin must be 0.75" (19 mm), except for the title page where it must be 1" (25 mm). No page numbers please. We will insert the page numbers for you.
Important DatesPaper submission: 15 Jan 2016 (Please specify our special session ‘Machine Learning with incompletely labeled data’ when you contribute your submission)
Acceptance notification: 15 Mar 2016
Final paper submission: 15 Apr 2016
For more information, please refer to our website