Abstract

No abstract

This paper introduces the large scale visual search algorithm and system infrastructure at Alibaba. The following challenges are discussed under the E-commercial circumstance at Alibaba (a) how to handle heterogeneous image data and bridge the gap between real-shot images from user query and the online images. (b) how to deal with large scale indexing for massive updating data. (c) how to train deep models for effective feature representation without huge human annotations. (d) how to improve the user engagement by considering the quality of the content. We take advantage of large image collection of Alibaba and state-of-the-art deep learning techniques to perform visual search at scale. We present solutions and implementation details to overcome those problems and also share our learnings from building such a large scale commercial visual search engine. Specifically, model and search-based fusion approach is introduced to effectively predict categories. Also, we propose a deep CNN model for joint detection and feature learning by mining user click behavior. The binary index engine is designed to scale up indexing without compromising recall and precision. Finally, we apply all the stages into an end-to-end system architecture, which can simultaneously achieve highly efficient and scalable performance adapting to real-shot images. Extensive experiments demonstrate the advancement of each module in our system. We hope visual search at Alibaba becomes more widely incorporated into today's commercial applications. CCS CONCEPTS • Information systems → Image search; • Computing methodologies → Visual content-based indexing and retrieval;

This research was motivated by a desire to help office workers change their sedentary behavior because a prolonged sedentary posture increases the risks of developing musculoskeletal injuries and chronic diseases, thus threatening their physical and psychological well-being. Regular breaks involving low-effort physical activities are effective in reducing the adverse impacts of inactive behaviors. In this article, we present the design of a posture-based interactive system called HealthSit, which was developed to promote a short lowerback stretching exercise during work breaks. Through a within-subject study involving 30 office workers, the effectiveness of HealthSit in facilitating the stretching exercise was examined by making comparisons between an interaction-aided, a guided, and a self-directed exercise mode. We also used HealthSit as a research probe to investigate the interactivity of the system in enhancing user experience and the psychological benefits of the fitness breaks. Compared with the other two modes, the interaction-aided exercise mode significantly improved the quality of the stretching exercise and enhanced motivation and emotional state. These results confirm the effectiveness of HealthSit in supporting fitness breaks as a new workplace technology. Based on our study, a set of design implications have been derived for technology-assisted fitness work breaks. KEYWORDSWork breaks; lower-back stretch; posture-based interaction; workplace fitness-promoting technology

This paper presents a field study on using peer-based cooperative fitness tracking (PCFT) to promote workplace fitness. The social bonding achieved through a collective fitness goal and the sharing of fitness data between two co-workers has been explored as a motivational factor that can encourage physical activity. The study involved 10 dyads of co-workers in two groups (a distributed vs. a co-located group) based on their proximity at work. The effectiveness of the proposed PCFT was examined by comparing fitness data over a period of three weeks: the baseline week, the PCFT intervention week, and the post-intervention week. The proximity effects on PCFT were investigated by comparing the fitness data, goal commitment, and interview results between the two groups. The quantitative results showed that the physical activity of participants in the co-located group improved significantly after the PCFT intervention. The qualitative results suggested that PCFT may improve the awareness of being physically active, stimulate exchange of knowledge to support active lifestyles and facilitate including fitness breaks in the daily work routine. Based on these findings, we discuss design implications for the future development of the PCFT-based applications and their potential contribution to increased office vitality.