Subscribe to ForumIAS

[Results] Prelims 2020 Over - Gearing for 2021

Because of the corona virus issue? If it does not happen then worst affected will be people whose interview is stuck, around 600-700 such people I think are yet to appear for their interview.


@neyawn sir and others, what is your take?

This is the official Thread for discussion of Prelims Postponement, Please do not create new threads for the same.

jack_Sparrow,curious_kidand137 otherslike this
5.9m views

13.7k comments

****PLEASE HELP****
guys can someone clarify regarding this- I marked a slight dot (~10% of bubble) on the wrong option before marking the right one. will this be considered as wrong? really appreciate any insight into how omr machine works or if someone has had a similar experience before.
4.6k views

Is there any possibility of tukkabaz pushing the cut off to 110. It did happen in 2016.

This time around a lot of people have their marks hovering in 20s and 30s.

Going by that cut off could cross 105 may be even reach 110 and some website has just predicted 105+-5 as cutoff.

I don't think Random guesswork can catapult the cutoff to beyond 105. That is just statistically improbable.


Yes it seems I just saw the cutoff on one website and panicked for no reason. I just saw 2015 question paper. Piece of cake, very conventional and cutoff was 107.33. Which means cut off here should not cross 100 even though intake was higher back then. This years questions were just too unconventional to push cutoff beyond 100.

4k views
Are carbon nanotubes used for artificial blood vessels? Any link, reference which states this?
4.4k views
Do you want a short stories App.
I working on AI since 3 years as software programmer.
I can give you that app to you   within  1 month. It is not only possible there are so many tools to do that. 
Bigdata+ AI= Shorr stories. 

Also ,  We are making Proctored exams to be human less... The project has been launched by Pearson USA now . You dont need human invigilation ... Just fee cameras and sensors and .. Live video frame perception and voice interception (noise sampling technique) ... and it is done. 

Because .. You do not use AI.. that does not mean.. It is not in market. 

Google .. Short stories app by AI .. or music apps by AI. There are so many startups in US and banagalore doing it and some of that already did that at good level. 

Problem with UPSC aspirants .. They are disconnected from how far the world is moving and how quickly. We are more entrenched in tracking SC cases, Govt policies or Economic stats ... Most of aspirants never track latest dovt in sciences or business. Reason is - Reading compilation and ... Even if people read news papers ... They filter such news ki ye nahi aayega. Reliance facebook agreement are iska upsc se lena dena nahi ... But FDi question isi pe hai. Reliance ka vo arm jisme Facebook ne imvest kiya vo non listed hai... FDi ..wham ?
and .. Lot of such questions this year which are not from so called core academic news. ... but from general spice news and they derived concepts from their and asked in exam. 
What I mean to say ... Devision of newspaper ki aap das min me pahle tick kar lo kaun sa news padna hai .. and then aur kuch dekhna bhi nahi 
.
 This is the prevailing you. 
But, UPSC attacked this view as well. The questions are from events instead of some academically aligned news in newspapers. 

Hope you have got me. 

Short stories is very simple to do in AI. Python have a very strong framework to create Deep fake programs and  Everybody is doing deep fakes. Machine learning amd robotic process automation  is thing of past now. Startups are implementing Deep learning now. 


Ok. And what about electricity transmission?

5.6k views
@Tyler Durden Bina tukkebazike ipaper me 45-50 questions ke upar surety ke saathsolve karna mushkil tha, for a well prepared student. Uske baad calulated risk aur tukka hi tha. Aur agar cut off 95 ke upar jaata hai matlab tukka public jeet rahi hai.


Tyler Durden,
4.5k views

Kntsaid

@Tyler Durden Bina tukkebazike ipaper me 45-50 questions ke upar surety ke saathsolve karna mushkil tha, for a well prepared student. Uske baad calulated risk aur tukka hi tha. Aur agar cut off 95 ke upar jaata hai matlab tukka public jeet rahi hai.


Go and check out discussion of people on this forum after Prelims on this forum...Almost every time people predicted it will go beyond 100. But it was in the range of 95 +/- ... Almost Always.

4k views

Is there any possibility of tukkabaz pushing the cut off to 110. It did happen in 2016.

This time around a lot of people have their marks hovering in 20s and 30s.

Going by that cut off could cross 105 may be even reach 110 and some website has just predicted 105+-5 as cutoff.

No way.

5.4k views

@gauravbrly409 , WTH dude!! Do you actually know what A.I is ? How is deep fake and deep learning related to creating "meaningful" stories ?. Don't be like newspaper reporters and try to pass your argument by bringing some technical words like Big data , which is just a statistical tools for prediction based on past data. Bro, A.I is only capable of learning from "sample" which is passed on to it. "Meaningful" stories requires cognisance , can you point out any cognizant A.I app out there ?  

vivekdeshmukh1504,
3.7k views

I have already answered that. 

Not possible at present level of AI... Even in future level of AI. Because ... The problem of wireless transfer of electrons at transmission level need some electric hardware devices ... and elctrical things. yaar itna to common sens laga lo.. Why stretching toooooo far. 

AI has nothing to do with it directly. 


13.6k views

Yes .. Learning 

You need to provide metadata .  which has to be created through multiple sources and algorithms. Big data is required at back end .. to create such data framing and scooping ... and AI engine will work over the data created over that.
AI engine will have neural networking algorithm to bring cognizance. 

Hope you got it. anything more ??


GreenArrow,
13.6k views
@rafael How AI can make it more efficient .. Can you explain ... 

Indeed wireless transmission is real. There is just one organization which made it to demo stage. And AI is particularly unrelated to it. Secondly, AI just do automation by applying human intelligence. The basic principle of AI ..that what AI provides is in contradiction with this. 

Also - It has been written Big Data ... Then .. Yes .. Big data can make it happen.. Because it will helo in scooping large amount of research data and finding patterns. etc.... 
Let us discuss science here..

Bongiyaan chodna band karo. 


Give me 2 reasons that how can make it efficient 



and what efficient means here. How can you make a thing efficient for which you have no proper tech. 


Aise to mai bolunga ... Tokamak ban gaya hai .. Nuclear fusion is possible to be commercialzed. AI can make nuclear fusion more efficient .. we will make nuclear fusion plant soon. Had hai yaar. 

You win... add 2.67 

Aurora,Ayushi7and6 otherslike this
5.8k views
@Biradar I did not get you completely. Is there a NBFC A with its name A on a credit card? I am seeing news reports of NBFCs wanting to issue credit cards but none about them having credit cards in their own name or even them being permitted to issue credit cards to individuals.


No not issuing credit cards. I am saying they might have business credit cards to their name. Otherwise what can be the reason for 3 only? 

Ok. I understand you now. Two coaching centre discussions have just said all 3 right like it was the most simple thing in existence without giving any explanation. Would like to know Shankar's reason behind only C.

3.2k views
@gauravbrly409 AI can make it efficient like detecting if there is some problem within transmission or more power loss occurring then it can shut off the process or something like that? Basically what I am saying is that maybe AI has no role in direct transmission but we can monitor the process through that.
3.3k views

Wireless Transmission electricity answer was wrong as always said. .. and I supported earlier theories of tesla and all that it is only in research phase and  in labs of power sector companies this time. 

The question asked -At present level of

So at present level .. Wireless transmission is in very nascent stage. Recently , In August 2020 (may be upsc preparing question paper at that time) 

New Zealand startup announced a demonstration to show that they have achieved it in lab and want to demo the concept. 

I searched and it is all over the Internet. The below link I am posting from IEEE publishing window itself which published this development on 20 aug 2020. 

Emrod’s laboratory prototype currently operates indoors at a distance of just 2 meters. Work is under way to build a 40-meter demonstration system, but it, too, will be indoors where conditions can be easily managed. Sometime next year though Emrod plans a field test at a still-to-be-determined grid-connected facility operated by Powerco, New Zealand’s second largest utility with around 1.1 million customers.


https://spectrum.ieee.org/energywise/energy/the-smarter-grid/emrod-chases-the-dream-of-utilityscale-wireless-power-transmission



So - There should not be iota of doubt -At present level of tech , Wireless electricity transmission is not feasible... It is actually still in demonstration stage by various companies. 


Also - Wireless charging is different concept - It uses near field magnetic induction as a transfer mechanism to supply very tiny induced emf. It is not what wireless transmission is. 


Transmission -This term specifically used in power sector to supply bulk of electricity power (high potential) to reasonably long or medium distances. 


So , I do not know .. why people making it right. Its basics of physics yaar that we have all studied in class 12. 





 because it was an SNT question and so every option must be correct. Coaching also seems to have gone by the same logic. If u apply knowledge we all know wireless electricity transmission by AI is not feasible at this stage. 

Also going by the wording of question short stories and wireless transmission is not possible. They asked at the present state Nd also effectively. These are key words. While marking B I knew all coaching key will give D. 

This reason + if you know song and short story writing is not AI's cup of tea yet, it will straight away eliminate 3 options out of 4. This is a format employed by UPSC to make deceptively difficult questions simpler.

3.1k views

Deep fake and Deep learning are actually related .. If you want true user experience ..  A ssume a  person real looking ...on some screen ...and narrating the story.  The image of person is generated by AI ... For ex somebody's mom. You have provided metadata for that to backend. The AI algorithm will create a deep fake on it .. and even you will see that person's all activities on screen like it has recorded.. but that is deep fake. 

Big data  is not only a statistics tool. Statistics is just very small component of big data tech. Big data is way way larger than in capabilities than stats. The actual thing abo big data is working on a heap  of data grams(think of as a garbage pile)  and analysis of data .. to join other kind of profiled data and .. and some user inputs .. and brings you a required data. 
Big data is not just one tech. There are minimum 5 to 10 layers of different techs involved in big data. 

Dont ask me to go more technical now.. Mai yahi kar rahi hu itne saal se.. Meri roji roti hai ye. UPSC us side business. 

 


GreenArrow,
13.4k views
@gauravbrly409 That's prove it you actually don't know how A.I works. 
"AI engine will have neural networking algorithm to bring cognizance." Do you even understand this ?

I have created a project based on neural network for detection of intrusion in network based on Snort IDS. So, I do know what I am talking about.

You provide sample data to neural network and mine was backtracking neural network , and based on its final result it adjust its weightage of parameters.
A.I work as good as you provide "right" "sample" "data" to it. 

There is no cognisance. Period. And we don't want A.I to have same, don't we ? 

3.5k views
Yaar jab electron beam chodne ka 11000 voltage ka device hi nahi hai. Abhi Wireless electricity ka hazard to public ka koi study nahi hui hai ... To matlab kyu lage ho bina wajah stretch kar kar ke justify karne me sir. 

Mains ka paper nahi hai .. ki kuch common words likha ke general general pe 2 no de hi dega. 

Evidence do naa. 


shaanu4595684,
13.4k views

Kntsaid

@Tyler Durden Bina tukkebazike ipaper me 45-50 questions ke upar surety ke saathsolve karna mushkil tha, for a well prepared student. Uske baad calulated risk aur tukka hi tha. Aur agar cut off 95 ke upar jaata hai matlab tukka public jeet rahi hai.


It's true. People attempt more questions these days thinking that net gain will be positive. And every option has 25% chance of being correct. If you eliminate 1 or 2 options, that probability even increases. And if you mark 1 answer correct by fluke also, it will make up for 3 incorrect answers marks-wise. People like this trade off. It's like gambling. From what I have seen, people who loose in this process are mostly the ones who attempt less.

Aurora,Ayushi7and1 otherslike this
4.9k views

Dude .. Ai have two part .. 
Neural networking thing .. is AI core. 

But, If you think of asAI Tech, Things change.  


14.6k views
@gauravbrly409 Read again , of what I have said. Deep fake , deep learning is different from Cognizant system. 

As far as Big data and sampling is concerned , its just prediction based on past data. However, you wanna fancy it , that is the crux of Big data. 


4.7k views
Write your comment…