While many public services are getting web-enabled, there are many hurdles in getting those services to a common man. A key problem is the low Internet penetration among this population. Despite the growth in mobile phone usage, mobile internet has not taken off. More than 50% of the handsets sold just support voice and text messaging. We therefore, propose a platform to deliver Internet-based services over these low-featured phones. Our approach enables rapid creation of SMS and voice interfaces for any IT service without any programming. In addition the platform enables rapid creation of mobile widgets for Android, Palm and Windows phones. Customized location aware applications can also be enabled for Android devices through the platform. A live pilot is currently ongoing in India for health and career related services.

Most enterprise data is distributed in multiple relational databases with expert-designed schema. Using traditional single-table machine learning techniques over such data not only incur a computational penalty for converting to a ”flat” form (mega-join), even the human-specified semantic information present in the relations is lost. In this paper, we present a two-phase hierarchical meta-classification algorithm for relational databases with a semantic divide and conquer approach. We propose a recursive, prediction aggregation technique over heterogeneous classifiers applied on individual database tables. A preliminary evaluation on TPCH and UCI benchmarks shows reduced training time without any loss of prediction accuracy.

As the web grows in size, interfaces & interactions across websites diverge - for differentiation and arguably for a better user experience. However, this size & diversity is also a cognitive load for the user who has to learn a new user interface for every new website she visits. Several studies have confirmed the importance of well designed websites. In this paper, we propose a method for quantitative evaluation of the navigational complexity of user interactions on the web. Our approach of quantifying interaction complexity exploits the modeling of the web as a graph and uses the information theoretic definition of complexity. It enables us to measure the navigational complexity of web interaction in bits. Our approach is structural in nature and can be applied to both traditional paradigm of web interaction (browsing) and to emerging paradigms of web interaction like web widgets.

Supporting fast access to large RDF stores has been one of key challenges for enabling use of the Semantic Web in real-life applications, more so in sensor-based systems where large amounts of historic data needs to be stored. We propose a semantics-based temporal view mechanism that enables faster access to time-varying data by caching into memory only the required subset of RDF triples. We describe our experience of implementing such a framework in the context of a wide area network monitoring system. Our preliminary results show that our solution significantly improves client access time and scales well for moderate data sets.