On-line learning and stochastic approximations
On-line learning in neural networks
Communications of the ACM
Interpreting TF-IDF term weights as making relevance decisions
ACM Transactions on Information Systems (TOIS)
Anatomy of a collective intelligence blood supply chain
Proceedings of the Fifth International Conference on Management of Emergent Digital EcoSystems
Hi-index | 0.00 |
With the proliferation of the emerging Online Social Networks and other conventional communication services, there is an increasing need for a tool which can facilitate individual users to effectively socialize across multiple, heterogeneous platforms. While the diverse nature of the heterogeneous services already makes the design of a cross-platform socialization tool challenging, an even more daunting task is to tackle the "noisy" nature of the Social Networking Services (SNS). Existing solutions all lack flexibility and extensibility, especially in supporting advanced users to better manage their cross-platform socialization via customized information processing. In this paper, we propose PIXS (Programmable Intelligence for Cross-platform Socialization) -- an open-source, extensible middleware which provides efficient information acquisition and dissemination across heterogeneous SNSs. A distinguishing feature of PIXS is its support of script-based operations. As a proof-of-concept to demonstrate the flexibility and effectiveness of PIXS, we have developed for it a Python-based semi-supervised learning application which can prioritize incoming messages from different platforms via a Rank Preserving Regression (RPR) framework. This framework can readily incorporate the domain knowledge of the end user. Our SGD-based approach also enables adaptive and incremental training of the ranking system according to the gradual evolution of the user preference. Performance evaluation based on real message traces shows that the proposed system can boost the user's efficiency in identifying and forwarding important messages across heterogeneous SNS platforms. Additional use-cases of PIXS are also discussed.