Jakarta Artificial Intelligence Research Projects
Jakarta Research is a research community based in Jakarta that focuses on Artificial Intelligence. Our aim is to improve people’s life through artificial intelligence. We see a lot of AI applications that solved real-world problem, and we want to be a part of it. We started with natural language processing applications which all of our researchers were familiar with. We also make such a rule for the project, it has to be do-able and can be finished within two months. Below are the projects and events that we have been successfully finished and on-going project. Moreover, we also add incoming project that remains discussed to be the next our to-do list.
Quora Question Pairs
In this project, we were trying to identify a pair question is a paraphrase or not. It is similar with text similarity text but it takes a different approach. The task is largely known as sentence-pair classification. This task usually use in forum websites such as Quora which the user may create a question that has been already posted and answered. We experimented with several machine learning algorithms and built neural network architecture based on long short-term memory. But, it turned out BERT perform better from the rests. You can read the details in paper pre-print.
TorchSenti
TorchSenti is a natural language library that focuses on sentiment analysis tasks which aims to provide sentiment analysis dataset and pre-trained models. The library build on top of PyTorch, we want to support research community to expand the knowledge and contributors to solve current problems. Those features and resources helps NLP researchers to benchmark and evaluate their proposed method. However, this library may be a starting point for everyone that want to learn sentiment analysis in depth. Find the details on the repository.
Maleo
Data Scientists spend more than a half of their time for data cleansing including for text. With that problem, in Jakarta Research, we are building a tools that make data scientists job easier to clean the text data such as removing hyperlinks, punctuations, mistyping, etc. You can find more on Github.
Twitter User Behavior
Social media is one of the main resource to analyze text data. In this project, we are trying to find a pattern of people that has the same behavior as in a scripted account. Twitter data is largely open with some limitations, and in this project we are focusing on politic and social issues such as sexual harassments, social fraud, general election, etc.
Deep Learning Discussion
Deep learning discussion is a weekly meeting which each researcher should present that they think interesting or people should know about. The discussion started to discuss information extraction from a book, but as time goes by, it covers a board topics from general to specific one.
Research Program for Students
This program aims to help students in a senior year that want to release research paper or being involve in research project. We create this program because we found some students are struggle for computing resources, datasets, and guidance in order to start their research. Therefore, if you are a student and interested with this problem, visit this link.
In coming events and projects
More workshops and research projects. Absolutely! these are often discussed on on several meetings. We have a lot of things to do but we have no much time to do that, we proposed a lot on incoming projects such as language model, text summarization, dialogue system, semantic segmentation on image, object detection, speech recognition, speech enhancement, network analysis. But, we are open if anyone wants to do research in Jakarta Research, maybe there is an available resources that you can use.
Besides that, we also plan to build homepage on github.io, utilizing Discord API, setup CI/CD for our Github, and many more. If you are interested to contribute to our research community, just comment below your email.
Cheers!
If you think this article is helpful, please share and give this article 50 claps :D
