ACM RecSys Workshop on Recommender Systems Evaluation: Dimensions and Design (REDD 2014)
Location: Silicon Valley, California, USA
Date: October, 2014
Evaluation is a cardinal issue in recommender systems; as in almost any other technical discipline, it highlights to a large extent the problems that need to be solved by the field and, hence, leads the way for algorithmic research and development in the community. Yet, in the field of recommender systems, there still exists considerable disparity in evaluation methods, metrics, and experimental designs as well as a significant mismatch between evaluation methods in the lab and what constitutes an effective recommendation for real users and businesses. This workshop aims at providing an informal forum to tackle such issues and move towards better understood and commonly agreed evaluation methodologies, allowing one to leverage the efforts and the workforce of the academic community on meaningful and relevant directions to real-world developments.
REDD 2014 places a specific focus, on the one hand, on the identification and measurement of different recommendation quality dimensions that go beyond the monolithic concept of simply matching user preferences. Novelty and diversity, for instance, have been recognized as key components of the utility of recommendations for users in real-world scenarios, with a direct positive effect on business performance. Considering the business perspective, performance metrics related to sales, revenues, and user engagement along the recommendation funnel should also be used. Additionally, from an engineering point of view, aspects such as efficiency, scalability, robustness, and user interface design are typically major concerns; often prioritized over the effectiveness of the internal algorithms at the core of the system. On the other hand, once a relevant target quality has been defined, a clear evaluation protocol should be specified in detail and agreed upon, allowing for the comparison, replicability and reproducibility of the results and experiments by different authors and enabling incremental contributions.
REDD 2014 aims at gathering researchers and practitioners interested in better understanding the unmet needs of the field in terms of evaluation methodologies and experimental practices. The main goal of this workshop is to provide an informal setting for discussing and exchanging ideas, experiences, and viewpoints. REDD seeks to identify and better understand the current gaps in recommender system evaluation methodologies, help lay directions for progress in addressing them, and foster the consolidation and convergence of experimental methods and practices.
Topics of Interest
We invite the submission of papers reporting original research, studies, advances, or experiences that focus on recommender system utility evaluation. The topics the workshop seeks to address include–though need not be limited to–the following:
- Recommendation quality dimensions
- Effective accuracy, ranking quality
- Novelty, diversity, unexpectedness, serendipity
- Utility, gain, cost, risk, benefit
- Robustness, confidence, coverage, ease of use, persuasiveness, etc.
- Matching metrics to tasks, needs, and goals
- User satisfaction, user perception, human factors
- Business-oriented evaluation
- Multiple objective optimization, user engagement
- Quality of service, quality of experience
- Evaluation methodology and experimental design
- Definition and evaluation of new metrics, studies of existing ones
- Adaptation of methodologies from related fields: IR, Machine Learning, HCI, etc.
- Evaluation theory
- Practical aspects of evaluation
- Offline and online experimental approaches
- Simulation-based evaluation
- Datasets and benchmarks
- Validation of metrics
- Efficiency and scalability
- Open evaluation platforms and infrastructures
Two submission types are accepted: technical papers up to 6 pages, and position papers up to 3 pages. Each paper will be evaluated by at least two reviewers from the Programme Committee.
The papers will be evaluated for their originality, contribution significance, soundness, clarity, and overall quality. Within a required quality standard, position papers will be evaluated for presenting new perspectives and insights, and their potential for provoking thoughts and stimulating discussion. All submissions shall adhere to the standard ACM SIG proceedings format. The accepted papers will be published in the CEUR Proceedings series.
Submission instructions can be found on the Submissions Page.
- Paper submission:
July 21, 2014July 28, 2014
- Notification: August 21, 2014
- Camera ready: September 5, 2014
- Workshop: October 10, 2014