Giter Site home page Giter Site logo

pythia's Introduction

Pythia

A Semantic Web Oracle.
By Eugen Sawin [email protected].

Description

Pythia is an experimental platform for semantic web search with keyword queries.

The goal is to translate keyword queries, as typically used with full-text search engines, to semantic queries suitable as input for semantic search engines with high confidence.

The hybrid approach combines full-text search and natural language processing to provide the best search experience.

Required Data

The hybrid approach requires a full-text index and knowledge base.

Currently, the full-text search is conducted via the Google Custom Search API, avoiding the requirement for an index pipeline. Likewise the Freebase API is used for a lightweight access to entity data, in addition to an offline format of the YAGO database.

Due to its huge size, the ontology index is not included in this repository, but will be made available online to allow for reproduction of the system and results.

Requirements

  • POSIX.1b-compliant operating system (support for librt)
  • GNU Make
  • GNU GCC 4.6+ (sudo apt-get install build-essential)
  • OpenSSL (sudo apt-get install openssl libssl-dev)
  • Python 2.7+ (optional, for style checking)

Dependencies

Required

Optional

Building

Building dependencies

To build all dependencies automatically, use:

make depend

Alternatively, you can build each dependency separately by following the instructions here, otherwise you may skip the rest of this section.

Building gflags

The repository contains a slightly modified gflags version with less verbose help output, which is used by default. If you prefer using your installed version, modify the makefile accordingly (follow the comments).

make gflags

Building glog

To build to included glog version, use:

make glog

Building POCO C++ Libraries

To build the POCO submodule, use:

make poco

This will check out the required POCO version from GitHub and build it, which can take a while.

Building SENNA

To get the required SENNA version, use:

make senna

This will download SENNA from the web (~200MB) and modify the source files, so that it compiles with g++. Downloading the files may take a while.

Building Flow

To build the Flow submodule, use:

make flow

Building Pythia

To build Pythia use:

make

To build the version optimized for speed (-Ofast), use:

make opt

For debugging, use the debug version:

make debug

Usage

To start Pythia use:

./pythia

To show the full usage and flags help use:

./pythia -help

Testing (depends on gtest)

To build and run the unit tests use:

make check

Profiling (depends on gperftools)

To build Pythia with profiling turned on use:

make profile

Getting cpplint

Code style checking depends on a modified version of Google's cpplint. Get it via

make cpplint

Checking style (depends on cpplint)

To test code style conformance with the Google C++ Style Guide use:

make checkstyle

License

Pythia - A Semantic Web Oracle.
Copyright (C) 2012, 2013 Eugen Sawin

Pythia is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.

You should have received a copy of the GNU General Public License along with this program. If not, see http://www.gnu.org/licenses.

The full copyright notice is contained in the file COPYING.

pythia's People

Contributors

eamsen avatar

Stargazers

 avatar  avatar

Watchers

 avatar  avatar  avatar

pythia's Issues

Establish ground truth

Provide a ground truth for query to result entities and from query to result semantic query.

Evaluate semantic results

Phase 2 results as returned by Broccoli need to be evaluated and results displayed jointly with the phase 1 results.

Customizable scoring functions

Provide an interface for custom scoring functions.
Currently scoring functions are hard coded in script.js and the backend.

Include important verbs in semantic query

Verbs which describe an important relation between the entities should be included in the semantic query, while common verbs like is should be ignored.

Consider simple verb form conversion.

Improve control over scoring weights

Provide a better way to balance the effects of the score components.
Currently, the static and dynamic weights do not provide optimal results in all cases.

Entity identification

Add entity identification by any means to improve result quality and reduce the number of duplicates.

Improve top candidates cut off

Current scheme to cut off by a set percentage of the exponential moving average does not provide optimal results in some cases.

Automate and visualize evaluation

Add automated evaluation:

  • select appropriate quality measures
  • visualize evaluation on the frontend
  • update evaluation async in the background on scoring function adjustments

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    ๐Ÿ–– Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. ๐Ÿ“Š๐Ÿ“ˆ๐ŸŽ‰

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google โค๏ธ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.