Building Natural Language Understanding system

We are building smart Digital Assistant using Deep Learning with a focus on Semantic Parsing and Natural Language Understanding. We actively keep track of State of Art Technology in various fields of NLP which include Neural Information Retrieval, QnA, Semantic Parsing, Joint Intent Classification and Slot Tagging, Joint Name Entity Recognition and Sequence to Sequence. We have paper accepted at top journal such as ACL.

Building a conversational A.I on embedded devices.

Summary: Building a RocketChat-based platform to enable and facilitate communication between learners, students and instructors, lecturers.Product Traction:_ The product is accepted into RMIT Activator, an Startup Incubator host by RMIT University, https://rmitactivator.edu.au/_ Lecturers use the product at RMIT University and British University of Vietnam_ Working with RMIT School of Science Working in a team to:_ Develop and customise and for livechat functions_… Show more Summary: Building a RocketChat-based platform to enable and facilitate communication between learners, students and instructors, lecturers.Product Traction:_ The product is accepted into RMIT Activator, an Startup Incubator host by RMIT University, https://rmitactivator.edu.au/_ Lecturers use the product at RMIT University and British University of Vietnam_ Working with RMIT School of Science Working in a team to:_ Develop and customise and for livechat functions_ Develop and deploy automated tests for livechat functions_ Improve and customise mobile application interface_ Maintain, develop and improve RocketChat core & Rest API _ Develop integrated applications: Attendance Checker & Classroom Response System _ Deploy and maintain the scalable system on Amazon Web Service, Azure Cloud app service Show less

Melbourne Network Society Institute_ Used a combination of a variety of techniques to solve the problem of 2D facial recognition that commonly arises in security or surveillance system. They include Signal Processing & Machine learning algorithms_ The algorithms were developed, tested and prototyped in MATLAB. They can also be scaled to work in large capacity using parallel and distributed computing._ Achieved above 90% accuracy for many famous dataset such as Extended Yale Database… Show more Melbourne Network Society Institute_ Used a combination of a variety of techniques to solve the problem of 2D facial recognition that commonly arises in security or surveillance system. They include Signal Processing & Machine learning algorithms_ The algorithms were developed, tested and prototyped in MATLAB. They can also be scaled to work in large capacity using parallel and distributed computing._ Achieved above 90% accuracy for many famous dataset such as Extended Yale Database B, Ferret, etc. Show less