DTU शोधकर्ताओं ने फेक न्यूज़, डीप फेक और हेट स्पीच का पता लगाने के लिए विकसित किया AI मॉडल; जानें कैसे करेगा काम?

ETVBHARAT

डीटीयू प्रो. दिनेश विश्वकर्मा अपनी टीम के साथ फेक न्यूज़, डीप फेक और हेट स्पीच को डिटेक्टर करने वाला एआई मॉडल बनाए.

Transcript

00:00In this time, we are here with the DTU Department of the IT Department, Professor Dinesh Kumar Vizsukarma.

00:07We know from Professor, which work is happening to us, and the AI model is in which way we can fix it.

00:13Thank you very much.

00:16Thank you very much.

00:46Thank you very much.

01:16Thank you very much.

01:18Thank you very much.

01:20Thank you very much.

01:22Thank you very much.

01:24Thank you very much.

01:26Thank you very much.

01:28Thank you very much.

01:30Thank you very much.

01:32Thank you very much.

01:34Thank you very much.

01:36news we can check that this is fake or this is real news.

01:42For that, our algorithms work.

01:46For that, the main purpose is where we can check that news is authentic and where is a portal

01:57that we can say that it is possible that it is wrong because we cannot say it wrong

02:03until we do not prove it.

02:06We did research on this research.

02:09We know that whenever we say something wrong,

02:13the way that we speak is smart,

02:18but sometimes it is such a thing that we know.

02:21So cognitive science says that when we say something wrong

02:25and we say something wrong,

02:27then the level of confidence,

02:30there is something wrong.

02:33So where we feel like we say hate speech,

02:37that no one is wrong,

02:39then it is seen that the lip movement is the same.

02:43These changes are the same,

02:45that the lip syncing is the same.

02:48So we see these things,

02:50which we try to know,

02:53that this may be a hate content.

02:57Now deep fake,

02:59this is a very threatening technology

03:02which has been very popular for some days.

03:07We have done a lot of research.

03:09We have done a lot of research.

03:11And now this research is now popular.

03:12We have done a lot of research.

03:13We have done a lot of research in the last year,

03:15two years.

03:16But we have done research in the last five years.

03:19Almost more than five years.

03:21Even in this field in this field.

03:24So when we saw this,

03:26it was the most important thing

03:28that this technology is very much misuse.

03:30This side of the positive and negative side

03:33is not much explored.

03:35But all the negative side is going to explore.

03:38So the first time came that

03:40this technology is the most important thing

03:42in pornography.

03:44And the popular celebrities

03:47can threaten from this type of technology.

03:51The popular leaders can threaten from this type of technology.

03:53And now we are seeing these things

03:56in the current days.

03:57There are many celebrities

03:59who have been targeted.

04:01Some politicians have been targeted.

04:03So this society has been very negative.

04:08And this is AI,

04:10the bad part of it is

04:13that people have been in AI.

04:15We have built up this way.

04:17This is the way that

04:18the fake and real

04:20is very difficult.

04:22It is a normal users.

04:24So for this,

04:26we have built up this algorithm.

04:29We have built up this.

04:30In the language of the algorithm,

04:31we have mentioned in the language.

04:32There is a computer program

04:34at the point of it.

04:36We can check any content from it.

04:37We have built up this system.

04:39We have built up this algorithm.

04:40And in the context,

04:41we have built up this algorithm.

04:42We have built up this algorithm.

04:45We have built up this algorithm.

04:47we have made the data we collect in the AI models

04:51we have created a face detection

04:55we have made the algorithm automatic identify

04:59so this happens to be in which way we are using

05:03the first approach that was the feature based

05:07which was the machine learning based

05:11now the concept of deep learning is more growing

05:15in that way we don't have the role of the face detection

05:19we can easily do it first

05:21we used the feature based approach

05:23so what is it that if we look at

05:27different geographical regions

05:29then some people have different skin tone

05:33and some people have different skin tone

05:35so that all the entire universe

05:39should cover us in the data

05:41so we don't have the data

05:45where people have different

05:47biographies

05:49and in that way

05:51we have the data

05:53so we try to cover the data

05:55worldwide

05:57then only we use it

05:59we use it

06:01and if it happens to people

06:03in research

06:05that people are doing it

06:07then people create data

06:09they are a little reluctant

06:11but it is that

06:13we are doing research in India

06:15so the role of India

06:17which is the role of the youth

06:19which is the role of the youth

06:21they are a lot of things

06:23that they are making

06:25but they are saying

06:27that before building

06:29our resources

06:31such as data sets

06:33to generate it

06:35and create it

06:37so that in our Indian scenario

06:39how the algorithms are performing

06:41and how the programs are working

06:43and how the programs are working

06:45for those things

06:47and now there is a lot of work

06:49and we hope

06:51in the time

06:53that it will be a good job

06:55the AI algorithm

06:57which you have developed

06:59is now in which form

07:01and it will be used for public

07:03in the form of app

07:05or in the form of website

07:07what is it

07:09you have asked a very good question

07:11if there is

07:13that we are researching

07:15so our objective

07:17in India

07:19that people are researching

07:21and publish it

07:23and after publishing it

07:25after publishing it

07:27they say that

07:29it is okay

07:30and it is still

07:31in India

07:32the researches are doing more

07:33in India

07:34but the scenario

07:35is now changing

07:36so

07:38when we research

07:39we see

07:40what is the work

07:42and what is the work

07:44and what is the work

07:46we see

07:48in which form

07:50it will be delivered

07:51in the form

07:52so

07:53for example

07:54there is

07:55fake news

07:56there is

07:58hate speech detection

08:00or deep faith detection

08:01so

08:02for this

08:03we see

08:04we have

08:05we have to make

08:06a application

08:07like

08:08web application

08:09or

08:10we have to make

08:11an app

08:12which is

08:13what is the work

08:15we can check that in this situation

08:18we can do more work

08:20because we focus more on product waste research

08:24so that our country will become our products

08:27and the country will be more useful

08:30and we can use it here

Category

Transcript

Recommended