_demos_COMBINED
Screencasts of our technologies
* xLiMe Schowcase
Screencast provided by Expert System Iberia, Madrid, Spain
The xLiMe Showcase web app lets you explore the various types of media that xLiMe is able to process and interlink.
Visit the 'live' version of the showcase here.
Developers can see the open-source projects here.
Follow the link to view the screencast.
* xLiMe Schowcase
Screencast provided by Zattoo, Berlin, Germany/Swiss
Zattoo’s web application allows the user to discover TV shows via similarity recommendations ("related shows"). Based on automatic speech recognition and subtitles analysis, xLiMe provides meaningful “related shows” recommendations even where electronic programme guide data are incomplete or unspecific. In this video, we show recommendations from xLiMe in action, in a prototypical integration into Zattoo's web application.
Follow the link to view the screencast.
* Entity summarization
Screencast provided by the Institute AIFB, KIT, Germany
ELES (pronounced "Alice") is a lightweight combination of entity linking and entity summarization. In the demo, we use an entity linking service to analyze text and link fragments to entities of the DBpedia knowledge base. The LinkSUM summarizer (interfaced via the SUMMA API definition) produces fact-based summaries of DBpedia entities. The two applications are combined on the client side through the "Internationalization Tag Set 2.0" W3C recommendation and lightweight jQuery-based interfaces.
Follow the link to view the screencast.
* x-media and x-lingual analytics
Screencast provided by VICO Research & Consulting GmbH, Germany
With the help of advanced analytical features from the xLiMe project, like event detection and visualization of clustered topics VICO Analytics is able to help researchers in understanding big amounts of user generated content in different languages without reading hundreds or thousands of documents every day.
Follow the link to view the screencast.
* Event Registry - correlate
Screencast provided by the Jožef Stefan Institute, Slovenia
In Event Registry developed during xLiMe project users can identify correlations between different data types, such as mentions of an entity in the media or its popularity on social media, stock prices of companies, number of sold items, etc.
Event Registry is a media monitoring platform that collects news published globally. One of the features it provides is also tracking of how frequently a particular entity is mentioned in the media. This allows one to obtain a time-series data of mentions over several years. Along with mentions of entities, Event Registry also stores time-series information about other data types, such as stock prices of companies, trading volumes, exchange rates, etc. With several millions of time-series, Event Registry can then identify for a given time-series what are other time-series that correlate the most with it. This can be used to identify simple relations between various data types as well as potential causation.
Follow the link to view the screencast.
* Semantic Visual Annotator
Screencast provided by the University of Trento, Italy
Semantic image and video annotation tools aim to automatically extract high level information from visual content.
The xLiMe semantic visual annotator can be applied to detect and localize within an image a specific object of interest or a brand logo, thus enriching image and/or video metadata.
The xLiMe semantic visual annotator allows you to follow your favourite brand or to monitor a topic of interest both on social network or TV mainstream.
Follow the link to view the screencast.
* XKnowSearch!
A cross-lingual semantic search system developed at the Institute AIFB, KIT, Germany.
XKnowSearch! - Exploiting Knowledge Bases for Entity-based Cross-lingual Information Retrieval.
(Chrome and Safari are recommended. Currently Firefox has a problem with the Demo. Please note that the performance of the server is low.)
Abstract: In recent years, the amount of entities in large knowledge bases available on the Web has been increasing rapidly, making it possible to propose new ways of intelligent information access. On the other hand, within the context of globalization, there is a clear need for technologies and systems that can enable multilingual and cross-lingual information access. In this paper, we demonstrate XKnowSearch!, a novel search system for multilingual and cross-lingual information retrieval, which supports traditional keyword search and also allows users to influence the search process according to their search intents. By exploiting multilingual knowledge bases on the Web, keyword queries and media data on the Web can be represented in their semantic forms, i.e., entities in knowledge bases, which facilitates query disambiguation and expansion, and also bridges the language barriers between queries and media data in different languages.
Follow the link to view the screencast.
* Dynamic content based on "Hot Products" from TV and Social Media
Screencast provided by econda GmbH, Karlsruhe, Germany
This screencast (without audio) introduces the econda xLiMe use case. The xLiMe technology allows to detect products – such as shoes – in TV streams and in Social Media messages. This information is used to generate recommendations in online shops. The use of external evidence results in dynamic content that is up-to-date, interesting and inspiring, and follows current trends in real time.
Follow the link to view the screencast.
* xLiMe Semantic Integrator
Screencast provided by the Institute AIFB, KIT, Germany
xLiMe Semantic integrator is a framework which leverages implicit or latent semantics in the textual data to provide search, recommendations and analytic's to various channels of information such social media, news and TV. It not only supports recommendations or search within same channels , but also across channels and languages.
Follow the link to view the screencast.
* xLiMe Frontend Library
Screencast provided by Expert System Iberia, Madrid, Spain
The xLiMe frontend library facilitates the prototyping of end-user applications on top of the xLiMe platform. In this video we introduce the library and demonstrate some of its components.
The link to the library itself can be found under the video on youtube.
Follow the link to view the screencast.
Project Year 3: Our current demos and resources
* Twitter Annotator (JSI)
Twitter Annotator is a scalable service written in Python that can perform several NLP tasks including tokenization, sentiment classification, and named entity recognition using internal or external tools such as Stanford CoreNLP. Link
* Twitter Corpus (JSI)
A manually annotated corpus of social media messages in German, Spanish and Italian. All messages include Part of Speech Tags, Named Entities, and Sentiment (Polarity, Message Level). Link
* Twitter Sentiment Corpus Generator (JSI)
Automatically generate a message polarity sentiment corpus from collected tweets by using smileys/emoticons and optionally news source tweets for Neutral/Objective class. Link
* xLiD- Lexica: Cross-lingual Linked Data Lexica (KIT)
We constructed the cross-lingual linked data lexica by exploiting the multilingual Wikipedia and linked data sources, especially DBPedia. We provide the reference and co-occurrence associations between entities and labels/words. To derive such associations scross languages, cross-language links that connect Wikipedia describing equivalent entities have been employed. Link
* X-LiSA: Cross-lingual Semantic Annotation (KIT)
We demonstrate a cross-lingual annotation framework that can link words or phrases in unstructured text in one language to resources in the structured KBs, such as DBpedia, grounded in the same or any other language with the goal of overcoming language barriers by facilitating multilingual and cross-lingual access to information. Link
* XKnowSearch! Exploiting Knowledge Bases for Entity-based Cross-lingual Information Retrieval (KIT)
We demonstrate XKnowSearch!, a novel search system for multilingual and cross-lingual information retrieval, which supports keyword search and allows users to influence the search process according to their search intents. The goal is to facilitate query disambiguation and expansion, and also bridge the language barriers between queries and media data in different languages. Link
* xLiMe Semantic Integrator (KIT)
It is a framework helpful to collect, integrate, search and compare different media types such as news, TV metadata, TV subtitles, TV speech to text data and social media in English, German, Spanish and Italian languages. Link
return to top
Project Year 2: Demos from JSI
1. What is correlating in news with European Central Bank as mentioned in news
2. Event Registry - system for global media monitoring that is integrating information from news, TV and social media
return to top
Project Year 2: Demo from ECONDA
Product recommendations based on external evidence - e.g. tweets, Facebook messages, Youtube comments, blog posts
return to top
Project Year 2: Demo from ZATTOO
TV recommender that semantically relates TV programmes to online news articles
(password-protected; may request access from joerg.schindler[at]zattoo.com)
return to top
Project Year 2: Demo from VICO
Event-based, automatic categorization of cross-lingual, cross-modal data streams for monitoring of semantically modelled research objects
(credentials can be requested from eduardo.torres-schumann [at] vico-research.com)
return to top
Project Year 2: Demo from KIT
The xLiMe Live Filtering and Search interface
return to top
Project Year 1: Use case SEARCH (ZATTOO)
With around 12 Million registered users, Zattoo is the biggest Internet (OTT) TV provider in Europe. Headquartered in Zurich, Switzerland, the company is present in six European countries. Zattoo's TV offering is available on PCs, smartphones and tablets (Android, iOS, Windows), on smart TVs, set-top boxes and game consoles. Depending on the country, Zattoo offers up to 150 TV channels, including the primary national and regional channels plus a selection of international channels. In Switzerland, all channels are available not only for watching live but also in catch-up mode ("Zattoo Recall"). Zattoo users can choose between a free, ad-financed model and pay subscriptions.
The following screenshots provide insights in the first year's prototype: xLiMe provides different types of contents that are related to the actual live TV content the user is watching, based on speech-to-text from the audio of 15 news-focused TV channels (English, German and Spanish), in close-to-realtime.
Related Tweets enhance the TV viewing experience by letting the user follow discussions about topics that are related to the content the user currently watches.
Related online news articles allow the user to learn more about topics or keywords that were automatically detected in the TV program the user is currently watching.
The search delivers results from the 7 days catch-up archive of currently 15 channels. When clicking on a search result, the chosen TV show starts playing right at the referenced cue position where the respective search term or keyword has been detected via speech-to-text (“in-scene video search”).
Project Year 1: Use case MONITOR (VICO)
VICO Research & Consulting harvests, analyses, and monitors large amounts of data from social media. Monitoring results are presented in a highly flexible, intuitive Dashboard which allows the user to select information relevant to the task at hand. Within the xLiMe project, along with providing social media data, VICO implements a Use Case for tools and technologies developed. During the first year of the project the main task was to extend VICO's monitoring system to deal with TV-data.
The screenshots below show examples of the current status of the xLiMe Use Case dashboard.
VICO's Social Media Monitoring Dashboard has been extended to handle TV-Data along with diverse social media data and news.
VICO's Monitoring Dashboard can be used to compare monitoring results for different brands or topics.
VICO's Monitoring Dashboard can show the original documents from the monitoring results, with highlighted keywords. As an example this screenshot shows excerpts from speech-to-text transcripts from TV-shows.
Project Year 1: Semantic Search Showcase
The xLiMe semantic search showcase demonstrates the applicability of language-specific ranking through the xLiMe autocomplete and xLiMe entity summarization functionalities. It retrieves the most recently published annotated media items for a specific entity. The current functionality is an early prototype that will be extended by further languages and more intelligent autocomplete and summarization technologies. Please check the early prototype.