All-in-one system for education, skills and careers
Social Learning Platform to connect, share and learn
College / University Management Software(ERP)
Digital Learning Materials – eBooks, Videos, Notes, etc.
Industry-required latest skill courses with certifications
Internships to gain real-time industry experiences
Jobs & placements directly from the industry
Latest educational news & events from India & abroad
Latest events exclusively for Higher Education
Scholarships exclusively for Higher Education
September 10, 2019
Researchers at IIT-H developed a new model which helps machines to explain their output using artifical intelligence (AI), this can make machines to make decision like human.
Currently there are devices such as phones, tablets, television sets and other electronic gadgets which make use of AI for machine translation, speech recognition etc,.
Nowadays during an appointment with doctor,we may be asked by chat-box to enter the report details of various tests and based on that,chat-box itself can conclude whether we need to meet the doctor or not. One of the major drawback in this case is that as its just a machine it won't be able to explain why it has arrived at the conclusion.
This drawback has been addressed in the model developed by IIT-H researchers in which the machine can explain the reason behind the conclusion it has arrived upon using AI techniques.
According to Vineeth N. Balasubramanian, associate professor, department of computer science and engineering, IIT-H ,"Our model can be used in any machine learning based device that involves prediction. From predicting which movie to watch next to occurrence of earthquake using seismic data, our model will help machines explain its conclusions,".
He also added that ,"Most machine learning based devices give an output based on a limited data stored in it. Human error, correlations in data can result in errors that go uncorrected. Once a machine learns to explain, it will learn concepts by itself,".
The main idea behind this research is to tackle the transparency of deep learning and to increase the end user's trust over machines. The aim of this research has been published in the 36th International Conference on Machine Learning,one of the highest rated conferences in the area of AI and ML,
This research has been carried out in collaboration with Honeywell, an American conglomerate which is in the process of patenting the model developed by IIT-H.
"Currently, the model works on a limited data. Large data may make the model turn slow. We are working on making it more efficient using large data," said Balasubramanian.
Download uLektz Apps for Latest Educational News / Events / Scholarships
July 25, 2019
May 06, 2019
September 27, 2019
February 26, 2019
January 03, 2020