What is Indra?

Indra is an efficient library and service to deliver word-embeddings and semantic relatedness to real-world applications in the domains of machine learning and natural language processing. It offers 60+ pre-build models in 15 languages and several model algorithms and corpora.

Indra is powered by spotify-annoy delivering an efficient approximate nearest neighbors function.

Features

Efficient approximate nearest neighbors (powered by spotify-annoy);
60+ pre-build models in 15 languages;
Permissive license for commercial use (MIT License);
Support to translated distributional relatedness;
Easy deploy: Deploy the infrastructure in 3 steps;
Access to the semantic models as a service;
Supports multiple distributional semantic models and distance measures.

Pre-build Models

Indra delivers ready-to-use pre-build models using different algorithms, data set corpora and languages. For a full list of pre-build models, please check the Wiki.

Model Algorithms

Supported Languages

EN - English
DE - German
ES - Spanish
FR - French
PT - Portuguese
IT - Italian
SV - Swedish
ZH - Chinese
NL - Dutch
RU - Russian
KO - Korean
JA - Japanese
AR - Arabic
FA - Persian
EL - Greek

Install

To install, please use the 3-step tool IndraComposed.

Getting Started

This guide provides the basic instructions to get you started using Indra. For further details, including the response format, additional parameters and the list of available models and language, please check the Wiki.

Requesting Word Embeddings `(POST /vectors)`

{
	"corpus": "googlenews",
	"model": "W2V",
	"language": "EN",
	"terms": ["love", "mother", "santa claus"]
}

For further details, check the Word Embeddings documentation.

Requesting Nearest Neighbors Vectors `(POST /neighbors/vectors)`

{
	"corpus": "googlenews",
	"model": "W2V",
	"language": "EN",
	"topk": 10,
	"terms": ["love", "mother", "santa"]
}

For further details, check the Nearest Neighbors documentation.

Requesting Nearest Neighbors Relatedness `(POST /neighbors/relatedness)`

{
	"corpus": "googlenews",
	"model": "W2V",
	"language": "EN",
	"topk": 10,
	"scoreFunction": "COSINE",
	"terms": ["love", "mother", "santa"]
}

For further details, check the Nearest Neighbors documentation.

Requesting Semantic Similarity (Pair of Terms) `(POST /relatedness)`

{
	"corpus": "wiki-2018",
	"model": "W2V",
	"language": "EN",
	"scoreFunction": "COSINE",
	"pairs": [{
		"t2": "love",
		"t1": "mother"
	},
	{
		"t2": "love",
		"t1": "santa claus"
	}]
}

For further details, check the Semantic Similarity documentation.

Requesting Semantic Similarity (One-to-Many) `(POST /relatedness/otm)`

{
	"corpus": "wiki-2018",
	"model": "W2V",
	"language": "EN",
	"scoreFunction": "COSINE",
	"one" : "love",
	"many" : ["mother", "father", "child"]
}

For further details, check the Semantic Similarity documentation.

Translated Word Embeddings and Semantic Similarity

For translated word embeddings and translated semantic similarity just append "mt" : true in the JSON payload.

Public Endpoint

We have a public endpoint for demonstration only hence you can try right now with cURL on the command line.

For word embeddings:

curl -X POST -H "Content-Type: application/json" -d '{
	"corpus": "wiki-2018",
	"model": "W2V",
	"language": "EN",
	"terms": ["love", "mother", "santa claus"]
}' "http://indra.lambda3.org/vectors"

For semantic similarity:

curl -X POST -H "Content-Type: application/json" -d '{
	"corpus": "wiki-2018",
	"model": "W2V",
	"language": "EN",
	"scoreFunction": "COSINE",
	"pairs": [{
		"t2": "love",
		"t1": "mother"
	},
	{
		"t2": "love",
		"t1": "santa claus"
	}]
}' "http://indra.lambda3.org/relatedness"

Citing Indra

Please cite Indra, if you use it in your experiments or project.

@InProceedings{indra,
author="Sales, Juliano Efson and Souza, Leonardo and Barzegar, Siamak and Davis, Brian and Freitas, Andr{\'e} and Handschuh, Siegfried",
title="Indra: A Word Embedding and Semantic Relatedness Server",
booktitle = {Proceedings of the Eleventh International Conference on Language Resources and Evaluation (LREC 2018)},
month     = {May},
year      = {2018},
address   = {Miyazaki, Japan},
publisher = {European Language Resources Association (ELRA)},
}

Contributors (alphabetical order)

Andre Freitas
Brian Davis
Juliano Sales
Leonardo Souza
Siamak Barzegar
Siegfried Handschuh

julsal/indra