Project

Neural SPARQL Machines

Goal: Neural SPARQL Machines (NSpM) are deep-learning architectures based on Long Short-Term Memories. A module named Generator builds a training dataset requiring little human effort. Using a machine translation approach, a NSpM aims at translating natural language utterances into SPARQL queries.

https://github.com/AKSW/NSpM

Methods: Machine Translation, Neural Networks, SPARQL, RDF, Long Short-Term Memories, Recurrent Neural Networks

Date: 1 July 2017

Updates
0 new
6
Recommendations
0 new
2
Followers
0 new
16
Reads
1 new
572

Project log

Tommaso Soru
added a research item
A booming amount of information is continuously added to the Internet as structured and unstructured data, feeding knowledge bases such as DBpedia and Wikidata with billions of statements describing millions of entities. The aim of Question Answering systems is to allow lay users to access such data using natural language without needing to write formal queries. However, users often submit questions that are complex and require a certain level of abstraction and reasoning to decompose them into basic graph patterns. In this short paper, we explore the use of architectures based on Neural Machine Translation called Neural SPARQL Machines to learn pattern compositions. We show that sequence-to-sequence models are a viable and promising option to transform long utterances into complex SPARQL queries.
Edgard Marx
added a research item
The Semantic Web community has successfully deployed a significant number of RDF datasets on the Web. However, for ordinary users, to access information available on this datasets, it is still a cumbersome and time-demanding task that requires the knowledge of description frameworks the use of formal languages. We introduce the Neural SPARQL Machine (NSpM) Telegram Chatbot, a chatbot built on top the open-source Telegram platform that facilitates the natural language querying of RDF datasets through the NSpM architecture. We show the different APIs implemented by the proposed chatbot as well as the processing steps from a natural language query to its corresponding answer. Additionally, we demonstrate how to instantiate as well as how to query other RDF datasets.
Tommaso Soru
added an update
We are pleased that our work Neural Machine Translation for Query Construction and Composition was accepted as a poster at the 2nd Workshop on Neural Abstract Machines and Program Induction (NAMPI v2 @ ICML 2018), co-organized by DeepMind AI and other top institutions. See you in Stockholm on July 15th!
 
Tommaso Soru
added a research item
Research on question answering with knowledge base has recently seen an increasing use of deep architectures. In this extended abstract, we study the application of the neural machine translation paradigm for question parsing. We employ a sequence-to-sequence model to learn graph patterns in the SPARQL graph query language and their compositions. Instead of inducing the programs through question-answer pairs, we expect a semi-supervised approach, where alignments between questions and queries are built through templates. We argue that the coverage of language utterances can be expanded using late notable works in natural language generation.
Tommaso Soru
added a project reference
Tommaso Soru
added an update
I will give a presentation titled Translating Natural Language into SPARQL for Neural Question Answering at the 6th Leipzig Semantic Web Day in Leipzig, Germany.
Some talks will be in German and some in English. Event details and programme available here:
 
Tommaso Soru
added an update
Google has finally released the list of students that will participate in the 2018 Google Summer of Code (GSoC). We are happy to announce that two students will work this summer on two DBpedia projects about different challenges in building A Neural QA Model for DBpedia. The mentoring will be carried out by AKSW in partnership with DICE Research. More details following.
  • Aman Mehta will be mentored by Tommaso Soru [1,2] and Ricardo Usbeck [1,2].
  • Tarun Sethupat will be mentored by Edgard Marx [1] and Rricha Jalota [2].
Congratulations!
 
Edgard Marx
added a research item
The role of Question Answering is central to the fulfillment of the Semantic Web. Recently, several approaches relying on artificial neural networks have been proposed to tackle the problem of question answering over knowledge graphs. Such techniques are however known to be data-hungry and the creation of training sets requires a substantial manual effort. We thus introduce Dbnqa, a comprehensive dataset of 894,499 pairs of questions and SPARQL queries based on templates which are specifically designed on the DBpedia knowledge base. We show how the method used to generate our dataset can be easily reused for other purposes. We report the successful adoption of Dbnqa in an experimental phase and present how it compares with existing question-answering corpora.
Tommaso Soru
added an update
DBpedia has been selected for the 7th time in a row as a Google Summer of Code organization. We are currently looking for a student to mentor on the project "A Neural QA Model for DBpedia". The work will be supported by Google. The terms can be found at the address: http://wiki.dbpedia.org/gsoc
If you are interested in contributing to DBpedia and Neural SPARQL Machines, please introduce yourself in the project description page: https://github.com/dbpedia/GSoC/issues/1
 
Tommaso Soru
added 4 project references
Tommaso Soru
added a research item
In the last years, the Linked Data Cloud has achieved a size of more than 100 billion facts pertaining to a multitude of domains. However, accessing this information has been significantly challenging for lay users. Approaches to problems such as Question Answering on Linked Data and Link Discovery have notably played a role in increasing information access. These approaches are often based on handcrafted and/or statistical models derived from data observation. Recently, Deep Learning architectures based on Neural Networks called seq2seq have shown to achieve state-of-the-art results at translating sequences into sequences. In this direction, we propose Neural SPARQL Machines, end-to-end deep architectures to translate any natural language expression into sentences encoding SPARQL queries. Our preliminary results, restricted on selected DBpedia classes, show that Neural SPARQL Machines are a promising approach for Question Answering on Linked Data, as they can deal with known problems such as vocabulary mismatch and perform graph pattern composition.
Tommaso Soru
added a project goal
Neural SPARQL Machines (NSpM) are deep-learning architectures based on Long Short-Term Memories. A module named Generator builds a training dataset requiring little human effort. Using a machine translation approach, a NSpM aims at translating natural language utterances into SPARQL queries.