Clemens Drews

Bio: Clemens Drews is an academic researcher from IBM. The author has contributed to research in topics: Set (abstract data type) & Social media. The author has an hindex of 19, co-authored 52 publications receiving 1675 citations.

Summarizing sporting events using twitter

Abstract: The status updates posted to social networks, such as Twitter and Facebook, contain a myriad of information about what people are doing and watching. During events, such as sports games, many updates are sent describing and expressing opinions about the event. In this paper, we describe an algorithm that generates a journalistic summary of an event using only status updates from Twitter as a source. Temporal cues, such as spikes in the volume of status updates, are used to identify the important moments within an event, and a sentence ranking method is used to extract relevant sentences from the corpus of status updates describing each important moment within an event. We evaluate our algorithm compared to human-generated summaries and the previous best summarization algorithm, and find that the results of our method are superior to the previous algorithm and approach the readability and grammaticality of the human-generated summaries.

Where Is This Tweet From? Inferring Home Locations of Twitter Users.

Abstract: We present a new algorithm for inferring the home locations of Twitter users at different granularities, such as city, state, or time zone, using the content of their tweets and their tweeting behavior. Unlike existing approaches, our algorithm uses an ensemble of statistical and heuristic classifiers to predict locations. We find that a hierarchical classification approach can improve prediction accuracy. Experimental evidence suggests that our algorithm works well in practice and outperforms the best existing algorithms for predicting the location of Twitter users.

Home Location Identification of Twitter Users

Abstract: We present a new algorithm for inferring the home location of Twitter users at different granularities, including city, state, time zone, or geographic region, using the content of users’ tweets and their tweeting behavior. Unlike existing approaches, our algorithm uses an ensemble of statistical and heuristic classifiers to predict locations and makes use of a geographic gazetteer dictionary to identify place-name entities. We find that a hierarchical classification approach, where time zone, state, or geographic region is predicted first and city is predicted next, can improve prediction accuracy. We have also analyzed movement variations of Twitter users, built a classifier to predict whether a user was travelling in a certain period of time, and use that to further improve the location detection accuracy. Experimental evidence suggests that our algorithm works well in practice and outperforms the best existing algorithms for predicting the home location of Twitter users.

Home Location Identification of Twitter Users

Abstract: We present a new algorithm for inferring the home location of Twitter users at different granularities, including city, state, time zone or geographic region, using the content of users tweets and their tweeting behavior. Unlike existing approaches, our algorithm uses an ensemble of statistical and heuristic classifiers to predict locations and makes use of a geographic gazetteer dictionary to identify place-name entities. We find that a hierarchical classification approach, where time zone, state or geographic region is predicted first and city is predicted next, can improve prediction accuracy. We have also analyzed movement variations of Twitter users, built a classifier to predict whether a user was travelling in a certain period of time and use that to further improve the location detection accuracy. Experimental evidence suggests that our algorithm works well in practice and outperforms the best existing algorithms for predicting the home location of Twitter users.

Method, computer readable media and apparatus for the selection and rendering of audio files in a networked environment

Abstract: A networked virtual jukebox renders audible music or other audio files to all within audio range of the virtual jukebox. The order of rendering is determined by requesting methods, which include networked peer-voting input, recent play history, random selection and voting. Voting is received from each networked device in communication with the networked virtual jukebox using all types of input methods such as keyboard, mouse, and voice input. The networked virtual jukebox can also operate unattended by playing music and/or audio files based on random selection of past voting.

System and method for synchronizing operations among a plurality of independently clocked digital data processing devices

Abstract: A system is described for maintaining synchrony of operations among a plurality of devices having independent clocking arrangements. A task distribution device is to distribute tasks to a synchrony group comprising a plurality of devices to perform tasks distributed by the task distribution device in synchrony. The task distribution device distributes each task to synchrony group members over a network. Each task is associated with a time stamp that indicates a time, relative to a clock maintained by the task distribution device, at which synchrony group members are to execute the task. Each synchrony group member periodically obtains from the task distribution device an indication of current time indicated by its clock, determines a time differential between the task distribution device's clock and its respective clock and determines therefrom a time at which, according to its respective clock, the time stamp indicates that it is to execute the task.

Using Social Media to Enhance Emergency Situation Awareness

Abstract: The described system uses natural language processing and data mining techniques to extract situation awareness information from Twitter messages generated during various disasters and crises.

Learning to interpret natural language navigation instructions from observations

Abstract: The ability to understand natural-language instructions is critical to building intelligent agents that interact with humans. We present a system that learns to transform natural-language navigation instructions into executable formal plans. Given no prior linguistic knowledge, the system learns by simply observing how humans follow navigation instructions. The system is evaluated in three complex virtual indoor environments with numerous objects and landmarks. A previously collected realistic corpus of complex English navigation instructions for these environments is used for training and testing data. By using a learned lexicon to refine inferred plans and a supervised learner to induce a semantic parser, the system is able to automatically learn to correctly interpret a reasonable fraction of the complex instructions in this corpus.

It's Mine, Don't Touch!: interactions at a large multi-touch display in a city centre

Abstract: We present data from detailed observations of CityWall, a large multi-touch display installed in a central location in Helsinki, Finland. During eight days of installation, 1199 persons interacted with the system in various social configurations. Videos of these encounters were examined qualitatively as well as quantitatively based on human coding of events. The data convey phenomena that arise uniquely in public use: crowding, massively parallel interaction, teamwork, games, negotiations of transitions and handovers, conflict management, gestures and overt remarks to co-present people, and "marking" the display for others. We analyze how public availability is achieved through social learning and negotiation, why interaction becomes performative and, finally, how the display restructures the public space. The multi-touch feature, gesture-based interaction, and the physical display size contributed differentially to these uses. Our findings on the social organization of the use of public displays can be useful for designing such systems for urban environments.

DiamondSpin: an extensible toolkit for around-the-table interaction

Abstract: DiamondSpin is a toolkit for the efficient prototyping of and experimentation with multi-person, concurrent interfaces for interactive shared displays. In this paper, we identify the fundamental functionality that tabletop user interfaces should embody, then present the toolkit's architecture and API. DiamondSpin provides a novel real-time polar to Cartesian transformation engine that has enabled new, around-the-table interaction metaphors to be implemented. DiamondSpin allows arbitrary document positioning and orientation on a tabletop surface. Polygonal tabletop layouts such as rectangular, octagonal, and circular tabletops can easily be constructed. DiamondSpin also supports multiple work areas within the same digital tabletop. Multi-user operations are offered through multi-threaded input event streams, multiple active objects, and multiple concurrent menus. We also discuss insights on tabletop interaction issues we have observed from a set of applications built with DiamondSpin.