Sunday 19 March 2017

Examinee, the Examiner.




A few years ago, while working with my six year old school-going son I had stumbled over an idea and had indeed registered a domain name ('Intellitute') to implement it. Nonetheless, I subsequently decided to call it off for the fear of loosing focus on a mammoth task, I had already set myself for. The idea, then relevant, is today apparent. And I wish someone takes up this opportunity to execute such an idea. I do not intend to pursue any financial gains from its implementation and would only require that whosoever executes the idea partially or fully, credits this blog/author for it on his product.

It is common for parents to help their children in understanding lessons in science or literature. We were no different. Once playfully, I proposed my son to set a test for me in order to compensate for him having to take tests constantly. Bang came the discovery - the lessons for which he tested me, were better and more in-depth studied and understood by him than those in which I tested him. When I carefully applied my mind it seemed logical - in a bid to trip me off, he ensured to set a test which was according to him exceptionally difficult. To set an exceptionally difficult test, rather than to answer it, one needs to know the lesson exceptionally well & in-depth. I repeated the experiment a few times and discovered a lucid pattern - the knowledge of my son in a topic of his liking would be further strengthened by getting him to invest in creation of a test, rather than taking one.

AI (Artificial Intelligence), BIG DATA Analysis, NLS (Natural Language processing) have all progressed fast enough to ensure that examinations can be personal rather than being set commonly for all the children of a nation (as it is now).

The idea is - ideal assessment should entail setting up of an examination (question paper) by an examinee first. An AI Engine can be built to scrutinize & evaluate the questions set by the students. Indeed, one can use various students to do the rating at the first instance even if the complexity of the question is not adjudged by an algorithm.

After substantial participation there will be an unimaginably large data bank of questions that can be ranked by the Engine. The Engine will then sort, shuffle and present back the questions set by students themselves ensuring that none of the questions that are set by the student appear back to him. The student would be examined for both his exam setting capabilities vis-a-vis others as also his answering capabilities. A question bank could always be made available to the AI Engine in case of mass collaboration in exam setting.

In learning, I also contemplated that every lesson provided audio-visually should have the capability of clearing doubts, inserted through chats, the AI Engine could very quickly learn all possible doubts about every lesson and have various approaches created and ranked on how best to clear a specific doubt of a specific type of student (depending on historical/prior success of clearing doubts). Every lesson would be followed by a testing & test-taking exercise. 

An AI Engine will not only be able to provide absolute and relative ranking/grades but would also be able to assess much better - special abilities in students. In talent assessment in most academic fields an AI Engine can trip off any human teacher.

AI Engine could provide the flexibility of letting students to dwell as deep as they desire, setting only the minimum limit of knowledge that a specific age student needs to know.

All lessons & the learning-output of students should be evaluated on the basis of DNSEA a methodology for description and quality-ranking, that I propounded in my work/book titled "Invention of Description" [ASIN: B072S1CT47 on Amazon Kindle]

This I believe will set a new paradigm in the field of education, which will allow more ability & knowledge in skill-building & competency-enhancing rather than being too general in nature, as is currently prevalent. The approach is all the more important as humankind shifts from degrees and diplomas to skills. And as self learning is becoming common (owing to an exponential rise in average human capabilities and information awareness) high quality assessment will be the key indicator, not tutoring. And it is a proven fact. IIT-JEE or other Entrance Exams are completely neutral to the school one has studied from. A student might have never gone to a formal school and may have accomplished education from open school. Yet if he/she surpass others they are admitted into university. Similar shall be the parameters for getting-good-jobs or excelling in future. 

The approach may be preferred by some and abhorred by others, but as I see it, it is an inevitable consequence for human progress to continue with a gusto that has been associated with last few decades.

... by Deepak Loomba

6 comments:

  1. Novel idea, but would the education czars be willing to give up their blinkered logic to experiment with QNSEA?

    ReplyDelete
    Replies
    1. The czars of today - No. But czars of tomorrow - yes. Putting the idea up for grabs is motivated by my assumption that a czar of tomorrow will execute it not the ones who have educational fiefdoms today.

      Delete
  2. Very doable only if concerned organizations look at it seriously for the future intelligent humans.

    ReplyDelete
  3. pretty innovative concept but I had a few concerns.
    1. This system assumes that the collective knowledge of the group is larger than the single knowledge of the trainer which might not be case in some cases.
    2. Given how the system is today, everyone is more concerned about passing the tests rather than deriving actual learnings and this suggestion makes "gaming the system" quite feasible i think. For instance, the examinees just submitting 3 questions each from a total of 10 questions (which everyone shares among themselves).

    I think the intentions behind the system are pretty sound but then again to make the system a fool-proof system is a huge task (as you rightly pointed out). And I am not sure that given the mindset of the students/teachers right now (to just pass it), this system will bring any positive change.

    ReplyDelete
    Replies
    1. Your reaponse is highly appreciated. Collective knowledge of a group is always more than that of one, assuming that like all other exams this exam is mandatory for all.

      On the possibility of fooling the system - its marginal to none because 1) those providing poor/no questions will fail in the test setting exercise which could mean (if need be) fail altogether; 2) the system has its own bank, so if questions are below the rank of those in question bank, then the system will push the bank questions.
      Lastly, people want to just pass and not learn not because they want to 'just pass', but the education system is such that encourages, incentivizes and hence habituates majority to it. Change the system and you will change the attitude towards education from 'degree acquisition' to 'skill acquisition'.

      Delete