Aakash Thatte

(GSoC) Summary of Neural Extraction Framework

Wed, 27 Sep 2023 00:00:00 +0000

Hello. This is the last blog related to GSoC as my project nears to an end and the final evaluation approaches. In this blog, I will try to summarize the project, write about what could be added further and future scope etc.

Summary of Neural Extraction Framework

GSoC project page - Link
Github Repo - Link

Neural Extraction Framework has been a wonderful experience for me. This project was in its third iteration this year in the Google Summer of Code and I am delighted to share that we have achieved to have an end-2-end framework to automatically extract relational triples from wikipedia articles and link them to entities and predicates in DBpedia.

(GSoC) Week-14 Recent Updates

Thu, 07 Sep 2023 00:00:00 +0000

Hello. This blog will be about a few updates I made in recent weeks. Also, I am excited to share that I will be giving a lightning talk in the Google Summer of Code Virtual Contributor Lightning Talks 2023 on my project. Its going to be a short talk(3 minutes max), I will be summarizing my project and I am really excited for it!

In think blog, I will give some info on the recent updates, like adding coreference resolution, some small speedups we got in the code, etc.

(GSoC) Week-11, 12 and 13 Recent Updates

Sat, 26 Aug 2023 00:00:00 +0000

Hello. Its been a while since I have put an update on the project(travelling and network issues). I had been working on some polishing, enhancing and re-experimenting, with the models that we are using, the time it takes to process the inputs etc. In this blog, I will give a review of these.

Restating the Goal

Just to make sure we are clear about the end goal, given a wikipedia page(an entity), we need to disambiguate the predicates/relationships between the page and all other pages(other entities) that are linked from that page. More precisely, we want to find the relationships between those pages which are only connected by the dbo:wikiPageWikiLink predicate and no other predicate, or mine the text to see even a relationship exists or not. There are cases where some page is linked in references, but is not related to the wikipedia page. The goal is to find all those hidden relationships by processing the wiki page text.

(GSoC) Week-9 and 10 Mid Evaluations and remaining TODOs

Sun, 06 Aug 2023 00:00:00 +0000

Hello. I recently passed my mid term evaluation for GSoC 2023. It has been a great experience so far. DBpedia finally has an end-2-end neural extraction framework. Got insightful feedback from the mentors. In this blog, I will discuss about some tasks that still need to be done in order to make this end-2-end relation extraction reliable and correct.

Background

To give a quick overview - our end-2-end framework takes in raw text as input and gives a list of (<subject>, <predicate>, <object>) triples as output. We are using REBEL relation extraction model for joint entity-relation extraction, GENRE for entity linking and text based vector similarity for mapping relations to predicates. This pipeline is explained in the [previous blog]({% post_url 2023-07-24-Week-8-End-2-End-RE %}) and works great.

(GSoC) Week-8 End-2-End Relation extraction

Mon, 24 Jul 2023 00:00:00 +0000

Hello. It feels amazing to start this blog as we have been able to create an end-2-end system for relation extraction from text, the ultimate goal of this Neural Extraction Framework project.

Before we begin, check the sample below of how the extracted and mapped triples look like:

Expectations

The expectation from an end-2-end relations extraction is to take as input a wikipedia article text and return a set of relational triples. These relational triples are in the form of <subject_entity> , <predicate> , <object_entity>/<literal>. Also, initially we want these triples to have the wikipedia page itself as the subject entity. For example, if the page is about Berlin_Wall, we would like to pick the triples where the subject entity is Berlin_Wall. This way, we have same objective as extracting from infobox, just that we also make use of the text to extract relations and thus enhance the process to mine more information about that entity/page.

(GSoC) Week-7 Predicate mapping

Wed, 19 Jul 2023 00:00:00 +0000

Hello. In this blog, we will be looking into mapping natural language relations to predicates in DBpedia. This task is extremely important in order to achieve our ultimate goal of a neural extraction framework.

Background

In the previous post, we had discussed about REBEL model for relation extraction. This model extracts the head entity, relation and tail entity from texts. But they are in natural language form - for example, “lives in” - we want them to be in the form of URIs - for example “https://dbpedia.org/ontology/residence" - so that we can link them to dbpedia predicates. For mapping entities to dbpedia resource URIs, we already have a few methods for entity linking.

(GSoC) Week-6 Relation Extraction - REBEL

Thu, 13 Jul 2023 00:00:00 +0000

Hello. The previous blog was the last in the entity-linking part of the project. Now we will move onto relation extraction. In this blog, we will discuss about REBEL - a relation extraction model. Given a text, this model can find relational triples(head-relation-tail) in the form of natural language text.

Background

REBEL is a text2text model trained by BabelScape by fine-tuning BART for translating a raw input sentence containing entities and implicit relations into a set of triplets that explicitly refer to those relations. It has been trained on more than 200 different relation types. REBEL is a joint model, meaning that it extracts entities and relations simultaneously.

(GSoC) Week-5 Entity Linking - Benchmarking

Tue, 04 Jul 2023 00:00:00 +0000

Hello. In this blog, I will discuss about the benchmarking(or evaluation 😅) of different entity-linking approaches that I have discussed in previous few blogs. I have used the WNED dataset for benchmarking/evaluation.

Dataset

I used the WNED(WNED-WIKI from the KILT Benchmarking) dataset available here. The dataset contains 345 wikipedia articles and 6,821 entity mentions from them, along with the entities they refer to(as wikipedia article titles). In the dataset, there are raw-text wikipedia article content which contain entity mentions, and then there is an wikipedia.xml file that contains information like the article title, each mention and its associated entity. I collected all entity mentions and their entities in a file:

(GSoC) Week-4 GENRE for entity linking

Tue, 27 Jun 2023 00:00:00 +0000

Hello. In this blog, I will be writing about using a text generation model to disambiguate entities, GENRE. It stands for Generative Entity Retrieval, presented in the Autoregressive Entity Retrieval paper by facebook research.

In a nutshell, GENRE uses a sequence-to-sequence approach to entity retrieval (e.g., linking), based on fine-tuned BART architecture. GENRE performs retrieval generating the unique entity name conditioned on the input text using constrained beam search to only generate valid identifiers.

(GSoC) Week-3 Entity Linking - Techniques used

Mon, 19 Jun 2023 00:00:00 +0000

Hello. In this blog, we will continue with entity linking(read [previous blog]({% post_url 2023-06-13-Week-2-Entity-Linking %}) for intro and simple demos), methods that I am using for entity linking in my project and some recent techniques that can be used to perform EL.

Background

In the previous blog, I had mentioned of using DBpedia lookup, DBpedia spotlight etc. In this blog, I will try to discuss how these work, discuss some approaches that I am using in my GSoC project to perform EL and also discuss some ideas like zero-shot classification that can be used to perform and enhance EL.

Indexing in information retrieval

Sun, 18 Jun 2023 00:00:00 +0000

Hello. This blog is about indexing techniques used in information retrieval. I will try to cover some basic indexing techniques and also provide code for some of them.

What is indexing?

An index is a data structure that is created to efficiently store and retrieve information from a collection of documents or data. The index allows for quick lookup and retrieval of documents based on user queries or search terms. In search engines, web pages are indexed, when user enters a query, the engine finds the relevant documents quickly using the index. Below are some indexing techniques:

(GSoC) Week-2 Entity Linking - Intro

Tue, 13 Jun 2023 00:00:00 +0000

Hello! In this blog, we will discuss about entity linking, a process used to map recognized entities to resources in a knowledge base, like DBpedia.

Background

Entity linking is the process of linking or mapping entity mentions(the identified entities in a text) to actual resources in a knowledge base. This is an important step towards our ultimate goal of relation extraction. Entity mentions in text need not be exactly same as the entities in the knowledge base, so it is very essential that we map them to the correct entity(in the KB) so as to extract correct relations. It is an important step which acts as a bridge to connect unstructured text to knowledge bases.

(GSoC) Week-1 Named Entity Recognition

Thu, 01 Jun 2023 00:00:00 +0000

Hi! In this post, I will be talking about NER, a very popular and important task in NLP!

Background

Named Entity Recognition, which is a subtask of natural language processing (NLP), involves identifying and classifying named entities within a given text. The goal is to identify and categorize the entities present in text. This also forms a step in our goal of relation extraction.

Entities present in text can be calssified(broadly) into a variaty of types like PERSON, ORGANIZATION, LOCATION, TIME, etc. It is easy for humans to identify and categorize entities from text. When we read or hear a sentence like "Messi is a football player of Argentina", we can easily identify that Messi is a person and Argentina is a location/country. Identifying these further helps us understand the relation between them, in this case, the relation is - player of. We can think of named entities as answers to wh questions like Who, What, When ,Where etc. in a sentence. But it is not as easy for a machine to understand and process text as we do. For this, we need either rule based methods or machine learning/deep learning approaches.

Accepted in GSoC 2023!

Sun, 21 May 2023 00:00:00 +0000

Hello and welcome!

I am delighted to write my first blog post about my acceptance in GSoC 2023 with DBpedia. Throughout the summer, I will work on the Neural extraction framework project with DBpedia.

Background:

It all started last year when I came to know about GSoC. It was that time when I learned about open-source, and I liked the idea of anyone being able to contribute and change the codes of so many softwares we use every day. I found an organization and worked hard contributing and writing a proposal. Though I was not selected for GSoC, I added two new distance metrics to a vector database later that summer. Since then, I have kept looking for opportunities in open-source projects, especially those I use.

About

Mon, 01 Jan 0001 00:00:00 +0000

Hi, I’m Aakash! I hold a B.Tech in Data Science and AI and work as an ML Engineer at Needl, where we are working to solve information overload problems in the finance domain. As part of my role, I work on developing and deploying NLP solutions, working on document AI, and more recently, around agents for automated insights generation.

During my bachelors, I have had a good time exploring Linear Algebra, Probability & Statistics, classical ML, Adversarial ML, GPU architecture (using CUDA) and Psychology.