ScienceWISE: A Web-based Interactive Semantic Platform for Scientiﬁc Collaboration

0 likes345 views

Sciencewise is an interactive, web-based platform designed for scientific collaboration. It adds a semantic layer to arxiv.org and features tools for scientific bookmarking, online storage and querying of publication metadata, and expert finding. The platform aims to enhance the accessibility and usability of scientific publications.

ScienceWISE
��
A
��Web-‐based
��Interac2ve
��Seman2c
��
Pla7orm
��for
��Scien2ﬁc
��Collabora2on
��
Karl
��Aberer,
��Alexey
��Boyarsky,
��
Philippe
��Cudré-‐Maurox,
��Gianluca
��
Demar-ni,
��and
��Oleg
��Ruchayskiy
��

ScienceWISE
��
•  An
��on-‐line
��pla7orm
��for
��scien2ﬁc
��publica2ons
��
•  A
��seman2c
��layer
��on
��top
��of
��ArXiv.org
��

ScienceWISE: A Web-based Interactive Semantic Platform for Scientiﬁc Collaboration

•  Scien2ﬁc
��bookmarking
��
•  Store
��and
��query
��publica2on
��meta-‐data
��online
��
•  Expert
��Finding
��

The document discusses an integrated socio-technical crowdsourcing platform designed to enhance e-science by leveraging collective intelligence from various fields of expertise. It emphasizes the need for highly-expressive machine-readable formats and ontologies to facilitate knowledge extraction and interoperability among scientific artifacts. This platform aims to automate scientific processes, improve analysis methods, and foster collaboration among scientists, addressing the limitations of individual researchers in a complex scientific landscape.

Entities, Graphs, and Crowdsourcing for better Web SearcheXascale Infolab

��

Gianluca Demartini presented on using entities, graphs, and crowdsourcing for better web search. He discussed using crowdsourcing to perform entity linking and disambiguation on web pages through a system called ZenCrowd. ZenCrowd combines algorithmic and manual linking by automating manual linking via crowdsourcing tasks and using a probabilistic reasoning framework to assess workers. He also discussed using entity factor graphs for scientific literature disambiguation by modeling workers, links, clicks, and constraints within a probabilistic framework. The system was experimentally evaluated on news articles from several sources linked to entities from knowledge bases like Freebase and DBPedia.

Dagstuhl2014eXascale Infolab

��

The document discusses the Zencrowd architecture designed for micro-matching tasks, utilizing crowdsourcing to enhance entity extraction from HTML and RDFa pages. It highlights a significant average precision improvement of 0.14 with minimal worker involvement and explores the adaptation of transactive memory concepts through social network analysis and crowdsourcing methods, achieving a 46% F1 improvement over competitors. Key components include a probabilistic network decision engine and a micro-task manager to optimize worker efficiency.

Entity-Centric Data ManagementeXascale Infolab

��

1) Entity-centric data management stores information at the entity level and integrates information by interlinking entities. This provides advantages over keyword-based and relational database approaches. 2) The XI Pipeline extracts mentions from text and performs named entity recognition, entity linking, and entity typing to associate entities with text. 3) Approaches like ZenCrowd and TRank leverage both algorithms and human computation through crowdsourcing to improve entity linking and fine-grained entity typing.

Ontology-Based Word Sense Disambiguation for Scientific LiteratureeXascale Infolab

��

This document presents an approach for ontology-based word sense disambiguation for scientific literature. It leverages the structure of community-based ontologies to improve sense identification. The approach represents concepts as context vectors based on their relations in documents and ontologies. It evaluates techniques based on minimum distance between concepts in ontologies, shortest path between concepts in ontologies, and neighboring concepts in ontologies. Combining these graph-based models with context vectors achieves the best precision for word sense disambiguation on two scientific datasets.

SANAPHOR: Ontology-based Coreference ResolutioneXascale Infolab

��

This document presents SANAPHOR, an ontology-based coreference resolution system that improves upon existing approaches by leveraging semantic information. It first links entities in document clusters to semantic types and ontologies. It then splits or merges clusters based on these semantic relationships. The system was evaluated on the CoNLL-2012 dataset, where it improved coreference resolution performance over the baseline Stanford system, particularly for noun clusters. By utilizing semantic knowledge, SANAPHOR demonstrates the benefits of enhancing syntactic coreference resolution with an additional semantic layer.

Braintalk cuso nmeXascale Infolab

��

The document discusses the challenges of analyzing and querying large scientific datasets, particularly in projects like the Human Brain Project and the Large Hadron Collider, which generate massive amounts of data. It highlights methods for scalable spatial access and data exploration, utilizing connectivity for efficient query execution and prefetching techniques to enhance performance. The conclusions emphasize the necessity for advanced data management strategies to enable scientific breakthroughs in an era of exponentially growing data.

Fixing the Domain and Range of Properties in Linked Data by Context Disambigu...eXascale Infolab

��

This document proposes three methods - LEXT, REXT, and LERIXT - for disambiguating the domain and range of properties in linked data by using context information. LEXT uses the type of subject resources, REXT uses the type of object resources, and LERIXT uses both. The methods were evaluated against expert judgments and achieved up to 96.5% precision for LEXT and 91.4% for REXT. LERIXT generated too many new sub-properties.

Beyond Triplets: Hyper-Relational Knowledge Graph Embedding for Link PredictioneXascale Infolab

��

1) The document presents HINGE, a new method for embedding hyper-relational knowledge graphs that aims to better capture information from facts containing multiple relations and entities. 2) HINGE uses a CNN to learn representations from base triplets and their associated key-value pairs to characterize the plausibility of facts. 3) An evaluation on link prediction tasks shows HINGE outperforms baselines and demonstrates that the triplet structure encodes essential information, while other representations discard important information.

It Takes Two: Instrumenting the Interaction between In-Memory Databases and S...eXascale Infolab

��

The document discusses the interaction between in-memory databases and solid-state drives (SSDs), focusing on performance challenges and the need for effective instrumentation. It introduces an instrumented SSD prototype called Cosmos+ that allows for detailed analysis of workload interference and promotes adaptive scheduling capabilities. The conclusions emphasize that SSDs can be utilized more transparently to improve database interaction and enhance near-data computing possibilities.

Representation Learning on Complex GraphseXascale Infolab

��

This document discusses advances in representation learning on graphs, focusing on techniques for embedding heterogeneous graphs and complex structures without relying on meta-paths. It highlights various methods, including 'just' for embedding heterogeneous graphs and 'nodesketch' for efficient graph embeddings via recursive sketching, showcasing their performance advantages in node classification and clustering tasks. The document concludes by mentioning future directions for graph representation learning, such as attributed and dynamic graphs.

A force directed approach for offline gps trajectory mapeXascale Infolab

��

This document discusses a force-directed approach for offline GPS trajectory map matching, highlighting the concept of placing GPS trajectories onto real maps and the difference between online and offline methods. The proposed method uses network visualization techniques to adjust GPS points based on their proximity to roads, demonstrating improved accuracy through computational experiments with taxi data in Rome. Initial results show enhancements in trajectory length and average lateral error, indicating the algorithm's effectiveness across various map matching methods.

Cikm 2018eXascale Infolab

��

The document presents a novel heterogeneous graph embedding technique called 'just' that employs random walks with jump and stay strategies, eliminating the dependency on meta-paths. Experimental evaluations demonstrate that 'just' achieves state-of-the-art performance in node classification and clustering tasks while maintaining efficient runtime compared to existing methods. The authors highlight the advantages of their approach in balancing heterogeneous and homogeneous edges during embedding learning.

HistoSketch: Fast Similarity-Preserving Sketching of Streaming Histograms wit...eXascale Infolab

��

This document proposes HistoSketch, a method for sketching streaming histograms that preserves similarity and adapts to concept drift. It works by: 1) Generating weighted samples from histograms such that the probability two sketches match equals histogram similarity. 2) Incrementally updating sketches using a weight decay factor to forget older data and adapt to drift over time. 3) Evaluating HistoSketch on classification tasks involving synthetic and real-world streaming data, finding it approximates histogram similarity well using small, fixed-size sketches while adapting rapidly to drift.

SwissLink: High-Precision, Context-Free Entity Linking Exploiting Unambiguous...eXascale Infolab

��

This document presents SwissLink, a high-precision context-free entity linking system. It extracts unambiguous surface forms (labels) from knowledge bases like DBpedia and Wikipedia to link entity mentions without context. It catalogs the surface forms, removes ambiguous ones using ratio and percentile methods, and performs fast string matching to link mentions. Evaluation on 30 Wikipedia articles shows the percentile-ratio method achieves over 95% precision and 45% recall, balancing precision and recall.

Dependency-Driven Analytics: A Compass for Uncharted Data OceanseXascale Infolab

��

The document discusses the challenges of analyzing massive, loosely structured, and evolving data, which incurs high cognitive and computational costs. It introduces Dependency Driven Analytics (DDA) as a solution that employs a dependency graph (DG) to serve as a conceptual map and index for raw data, facilitating improved query and processing efficiency. While current implementation shows significant performance improvements, it also highlights open challenges, including real-time processing and scalable graph technologies.

Crowd scheduling www2016eXascale Infolab

��

The document proposes a novel crowdsourcing system architecture and scheduling algorithm to address job starvation in multi-tenant crowd-powered systems. The architecture introduces HIT-Bundles to group heterogeneous tasks and control task serving. The Worker Conscious Fair Scheduling algorithm balances fairness and priority while minimizing worker context switching between tasks. Experiments on Amazon Mechanical Turk show the approach increases throughput over baseline schedulers and adapts to varying workforce levels and job priorities.

Efficient, Scalable, and Provenance-Aware Management of Linked DataeXascale Infolab

��

This document discusses the management of linked data, emphasizing the need for efficient, scalable, and provenance-aware systems. It introduces new techniques for data co-location and partitioning to improve query processing in the cloud and highlights the importance of data provenance for ensuring quality and trustworthiness. Additionally, it outlines a distributed linked data management system and various query execution strategies that utilize provenance information.

SSSW 2015 Sense MakingeXascale Infolab

��

This document summarizes a presentation given at SSSW 2015 on making sense of semantic data. It discusses challenges in understanding semantic web data, including a "language gap" between semantic web languages like SPARQL and natural language. It presents an approach to bridging this gap through automatically verbalizing SPARQL queries in English. Evaluation results show this helps non-experts understand queries better and faster than the SPARQL format. It also discusses the "semantic gap" caused by mismatches between a question's semantics and a knowledge graph, and presents an approach using templates to generate SPARQL queries from natural language questions.

LDOW2015 - Uduvudu: a Graph-Aware and Adaptive UI Engine for Linked DataeXascale Infolab

��

Uduvudu is a graph-aware UI engine for linked data that generates tailored representations for users using a catalog of matchers and templates. Unlike typical visualization toolkits, Uduvudu focuses on storytelling through customizable templates and structured data processing founded on semantic principles. The system leverages RDF graphs to adapt displays based on user context such as device and language, enhancing reusability and efficient data handling.

Executing Provenance-Enabled Queries over Web DataeXascale Infolab

��

The document discusses the execution of provenance-enabled queries over RDF databases, emphasizing the importance of data provenance in assessing data quality and trustworthiness. It outlines various query execution strategies, including post-filtering, query rewriting, full materialization, pre-filtering, and partial materialization, with experimental results demonstrating performance improvements for these methods. The findings indicate that tailored queries with provenance data can significantly enhance execution speed, with partial materialization being the optimal strategy despite its implementation challenges.

The Dynamics of Micro-Task CrowdsourcingeXascale Infolab

��

The document discusses the dynamics of micro-task crowdsourcing on Amazon Mechanical Turk (MTurk), presenting findings from a large-scale analysis of 2.5 million batches and 130 million hits collected over five years. Key insights include classification of hit types, factors affecting batch throughput, and the elasticity of supply and demand in the marketplace, highlighting a periodicity in task availability. The research emphasizes significant features for predicting batch performance, suggesting that crowdsourcing can be efficient if optimized appropriately.

CIKM14: Fixing grammatical errors by preposition rankingeXascale Infolab

��

The document discusses a method for grammatical correction through ranking prepositions based on their likelihood of correctness, emphasizing its relevance in machine translation and speech recognition. It outlines the approach's processing pipeline, using n-grams and statistical measures like pointwise mutual information (PMI) to improve preposition selection. The proposed two-class classification algorithm outperforms existing methods, particularly when incorporating skip n-grams.

OLTP-BencheXascale Infolab

��

The document introduces OLTP-Bench, an open-source benchmarking framework designed for testing relational databases with a focus on OLTP and web workloads. It highlights the framework's extensible architecture, rich metrics, and the variety of workloads that have been tested, as well as future enhancements such as improved tooling and support for NoSQL. The aim is to reduce redundant efforts in benchmarking infrastructure within the research community and ensure more effective resource management during experiments.

An Introduction to Big DataeXascale Infolab

��

The document discusses a seminar on big data led by Prof. Dr. Philippe Cudré-Mauroux, outlining key topics like the 3 Vs of big data (volume, velocity, variety), its economic impact, and the burgeoning job market anticipated in the field. It highlights big data as a new economic asset that is transforming various industries, emphasizing the need for new processing methods for effective decision-making and insight discovery. The presentation also covers big data infrastructures, the role of Hadoop, and the current state of big data initiatives particularly in Switzerland.

Internet Infrastructures for Big Data (Verisign's Distinguished Speaker Series)eXascale Infolab

��

The document discusses the emergence and impact of big data, highlighting the increasing demand for new infrastructures to handle high-volume, high-velocity, and high-variety data. It projects significant job creation related to big data and emphasizes its role as a new economic asset in various sectors. Additionally, it outlines specific big data projects and technologies such as Hadoop and anomaly detection in smart cities.

Hasler2014eXascale Infolab

��

This document discusses a project called MEM0R1ES that aims to automatically organize a person's digital information from various devices and online services to generate useful digital memories. The project develops techniques for entity search, typing, clustering, and elicitation to extract, integrate and expose personal information from heterogeneous graphs. It has produced several open-source software components and published results in top conferences. The document outlines current research directions and concludes that the project addresses important societal issues through stimulating collaboration between institutions.

Crowdsourcing is for the taileXascale Infolab

��

Crowdsourcing is useful for curating information about tail entities, which are less popular entities like local restaurants, niche sports, or emerging music bands. Targeted crowdsourcing platforms like Pick-A-Crowd aim to match tasks to workers who can provide higher quality answers by considering a worker's social profile and task context. Transactive search uses the knowledge of crowds to reconstruct memories and answer questions by targeting the right people to search sources like Twitter photos or event attendees.