The need to find or construct tables arises routinely to accomplish many tasks in everyday life, as a table is a common format for organizing data. However, when relevant data is found on the web, it is often scattered across multiple tables on different web pages, requiring tedious manual searching and copy-pasting to collect data. We propose KTabulator, an interactive system to effectively extract, build, or extend ad hoc tables from large corpora, by leveraging their computerized structures in the form of knowledge graphs. We developed and evaluated KTabulator using Wikipedia and its knowledge graph DBpedia as our testbed. Starting from an entity or an existing table, KTabulator allows users to extend their tables by finding relevant entities, their properties, and other relevant tables, while providing meaningful suggestions and guidance. The results of a user study indicate the usefulness and efficiency of KTabulator in ad hoc table creation.

Existing data stores of enterprises are full of connected data and users are increasingly finding value in performing graph querying, analytics and visualization on this data. This process involves a labor-intensive ETL pipeline, where users write scripts to extract graphs from data stored in legacy stores, often an RDBMS, and import these graphs into a graph-specific software. We demonstrate GraphWrangler, a system that allows users to connect to an RDBMS and within a few clicks extract graphs out of their tabular data, visualize and explore these graphs, and automatically generate scripts for their ETL pipelines. GraphWrangler adopts the predictive interaction framework and internally uses a data transformation language that is a limited subset of SQL. Our demonstration video can be found here: https://youtu.be/k92Qk6vuIsU

Precise indoor localization is of great importance to automatically track people or objects indoors and plays a vital role in modern life. Despite a number of innovative research present in the literature indoor localization still remains an open problem. To trace the main reason we identify that in the present literature the tendency is to pinpoint the exact coordinates of a target device although most of the location based services (LBSs) do not require exact coordinates. To support LBS, one can simply divide the area of interest into several zones and perform ``zone-fencing'', i.e., find under which zone the user is currently located at. In this paper, we propose a zone-based indoor localization scheme using neural networks. With the results from real world indoor settings, we show that a number of empty clusters is generated when the traditional counter propagation network (CPN) is applied as is. But a slight modification to the CPN reduces the number of empty clusters significantly and provides promising accuracy. The proposed scheme outperforms ``k-Nearest Neighbor algorithm" (k-NN) and its promising accuracy makes it suitable for real-world deployment.

The ability of smart devices to recognize their owners or valid users gains attention with the advent of widespread highly sensitive usage of these devices such as storing secret and personal information. Unlike the existing techniques, in this paper, we propose a very lightweight single-time user identification technique that can ensure a unique authentication by presenting a system near-to-impossible to breach for intruders. Here, we have conducted a thorough study over single-time usage data collected from 33 users. The study reveals several new findings, which in turn, leads us to a novel solution exploiting a new machine learning technique. Our evaluation confirms that the proposed solution operates with only 5% False Acceptance Rate (FAR) and only 6% False Rejection Rate (FRR) over the data collected from 33 users. We further evaluate the performance through comparing its performance with some existing machine learning techniques. Finally, we perform a real implementation of our proposed solution as a mobile application to conduct a rigorous user evaluation over 27 participants using three different devices in order to show how the solution works in practical situations. Outcomes of the user evaluation demonstrate as low as 0% FAR after letting intruders to mimic the actual user, which ensures extremely low probability of being breached. Moreover, we let 2 users to continuously use our application over 25 days in different states during their operation. Outcomes of this evaluation demonstrate as low as 1% FRR confirming the usability of our technique in long-term usage.