Dr. Joeran Beel

Summary

Recommender systems like on Amazon generate up to 30% of enterprises' consumption. Consequently, many large enterprises employ hundreds of software engineers to develop recommender systems and make them as effective as possible. Small and medium-sized enterprises (SME) often do not have the skills and resources to develop recommender systems from scratch. Instead, they typically use recommendations-as-a-service (RaaS). RaaS providers operate recommender systems on their servers, and SMEs pay a monthly subscription fee to request recommendations from the RaaS' API. Current RaaS providers offer suboptimal recommendation effectiveness because their systems do not adapt optimally to the characteristics of their SME partners. This is due to three reasons. 1. RaaS providers use only a single recommendation-framework. This limits both the number of potential algorithms to use and the type of data they can process. 2. RaaS providers follow a one-size-fits-all approach and optimise their systems globally, i.e. they use one algorithm for all their SMEs. However, research shows that recommendation algorithms perform significantly different with different SMEs; hence one algorithm for all SMEs will deliver suboptimal effectiveness for at least some of them. 3. RaaS providers do not optimise the presentation of recommendations. However, the presentation of recommendations is important as, for instance, the number of recommendations and the displayed information has a strong impact on how users perceive the quality of recommendations. We create the spin-out company Darwin & Goliath, which will offer a unique and highly adaptable meta-recommendation framework as a recommendations-as-a-service. The proposed solution features three USPs. 1. Darwin & Goliath uses multiple state-of-the-art recommendation frameworks (Apache Mahout, TensorFlow, and others). Each of these frameworks provides dozens of pre-implemented algorithms. Consequently, Darwin & Goliath chooses from a wide range of algorithms when it comes to identifying the best algorithm for each SME. 2. Darwin & Goliath applies a novel technology for algorithm selection, which we call "micro optimization". Not only identifies Darwin & Goliath the best algorithm for each SME, but for each recommendation request. This means that even for the same SME, Darwin & Goliath may use a different algorithm for different users, items, and contexts. 3. Darwin & Goliath automatically finds the most favourable way of presenting recommendations for each SME.

Funding Agency

Enterprise Ireland

Programme

Commercialisation Fund

Project Type

Research Development

Summary

Docear is a unique solution to academic literature management, i.e. it helps you organizing, creating, and discovering academic literature. Among others, Docear offers: 1. A single-section user-interface that allows the most comprehensive organization of your literature. With Docear, you can sort documents into categories; you can sort annotations (comments, bookmarks, and highlighted text from PDFs) into categories; you can sort annotations within PDFs; and you can view multiple annotations of multiple documents, in multiple categories - at once. 2. A "literature suite concept" that combines several tools in a single application (pdf management, reference management, mind mapping, ...). This allows you to draft your own papers, assignments, thesis, etc. directly in Docear and copy annotations and references from your collection directly into your draft. 3. A recommender system that helps to discover new literature: Docear recommends papers which are free, in full-text, instantly to download, and tailored to your information needs.

Funding Agency

German Federal Ministry of Economics

Programme

EXIST

Summary

Docear is a unique solution to academic literature management, i.e. it helps you organizing, creating, and discovering academic literature. Among others, Docear offers: 1. A single-section user-interface that allows the most comprehensive organization of your literature. With Docear, you can sort documents into categories; you can sort annotations (comments, bookmarks, and highlighted text from PDFs) into categories; you can sort annotations within PDFs; and you can view multiple annotations of multiple documents, in multiple categories - at once. 2. A "literature suite concept" that combines several tools in a single application (pdf management, reference management, mind mapping, ...). This allows you to draft your own papers, assignments, thesis, etc. directly in Docear and copy annotations and references from your collection directly into your draft. 3. A recommender system that helps to discover new literature: Docear recommends papers which are free, in full-text, instantly to download, and tailored to your information needs.

Funding Agency

ego.START

Programme

Seed Funding

Summary

See http://sciplore.org

Funding Agency

German Federal Ministry of Economics

Programme

EXIST

Summary

Domain-Independent Semantic Annotation of the Text (DISCANT): Surrounded by huge and exponentially growing volume of information of various nature, every day we face challenges with keeping track of the latest news in a domain of interest or finding good-quality answers to specific questions. The problem of information overload has been addressed by modern information systems and search engines, however, their capabilities are strongly limited by using unstructured textual formats for exchanging and storing information. Textual documents often contain a lot of concrete facts and quantifiable information, but the communication of these facts and information through human languages renders them inaccessible to machines. This "semantic bottleneck" problem substantially slows down communication and knowledge propagation in the world. In order to equip machines with the ability of effective processing of the text, DISCANT ("Domain-Independent SemantiC ANnotation of the Text") aims at creating a comprehensive framework for semantic annotation of textual documents of arbitrary domains, such as scientific papers, legal documents, customer reviews or clinical trial reports. We will develop approaches, methods and tools for two classes of solutions: an environment for discovering entity and relation types in a given domain and a system for automated semantic annotation of the text. The project proposes a novel approach based on a combination of unsupervised natural language processing and machine learning techniques. DISCANT will advance machine understanding of the text, contributing to the release of important knowledge buried in textual documents, the creation of machine-readable knowledge repositories and more effective solutions for semantic search and personal recommendations. As a consequence, DISCANT will equip the consumers of textual documents with better tools for overcoming data deluge and information overload, enabling them to make better-quality, data-driven decisions.

Funding Agency

EU/SFI

Programme

Marie Curie EDGE Fellowships

Project Type

Research Project