WikiConference India 2016/Submissions/Sum total of Human Knowledge, Next step in Human Evolution and Wikipedia's role in it
| Main page | Hackathon | Programs | Edit-a-thon | Press coverage | FAQ | Sitemap |
- Title of the submission
Sum total of human knowledge, Next Step in Human Evolution and Wikipedia's role in it.
- Your Username (For the submission author)
Thewinster (Link)
- Type of presentation
Talk
A request to the admins and volunteers: Although the concept and implementation of the idea is straightforward here, the full presentation is very long and meandering. I am continually writing drafts and iterating over them. Please find the work in progress on this public google doc link:
https://docs.google.com/document/d/1EGojRl46Hrk7XQd8BUNz9MYjdUtf8yDKvbCQBF1SCxQ/edit?usp=sharing
- Abstract (in about 300 words)
There was once a time when we had encyclopedias written on paper. Wikipedia came along, and it changed how we consumed and maintained information. We increased our reference to an encyclopedia from once in a blue moon to several times in an hour.
Can we do better? We have descriptive knowledge represented in our wikipedias. Can we make a shift to a different form in which "the sum total of humanity's knowledge" exists?
The presentation operates on the premise that we can make this shift. I will explore how simple browser plugins ala "Pocket" can kickstart such a processing. I will lay down a case where such innovation is the platform from which India will understand itself, and innovate. Note my emphasis on understanding itself, and innovation in contrast to merely saying "India will have access to knowledge".
Different themes explored
The presentation will have touch on several themes, show cases and case studies (all boiled down to their essence in view of time constraints):
- Neuromorphic Social Network
- Increasing the number of maximum possible number of editors on Wikipedia, by drastically reducing the bar of first contribution.
- Quick snippets of which crowdsourcing platforms worked, which didn't and their levels of cognitive load. (Khan Academy went away from crowd sourcing, Wikiversity never took off), quick perspectives from the disciplines of growth hacking and UX. How does this help for wikipedias which are slow, do not have critical mass and also, for other sister projects!?
- The synthesis: How does Wikipedia pick up pace, develop itself, sister projects, Indic Wikipedias and develop ahead of closed source social networks into a global brain? Before we construct the global brain, will we be able to have a Hindi Wikibooks used by Indians?
Neuromorphic Social Network
Artificial Neural Networks are in vogue right now. Taking inspiration from them, I describe here neuromorphic social networks. These are social networks that process information, mimicking how our brain functions. This function is the form that I advocate the change to.
The basic functional similarity is this:
- In a brain, one neuron takes input from several other neurons. The neuron Processes this input and outputs it to several other neurons. This process continues almost ad-infinitum. All our emotions, hopes, dreams, memories, conversations, sights, sounds come from this simple network.
- In the proposed social network, people consume, process and disseminate information. The processing involves finding relevance, asking questions, highlighting important lines etc. This type of processing starts with one article and builds up several threads. In effect, we are consuming and processing the knowledge. Contrast this to descriptive knowledge. I contest that such a network will have emergent properties.
What do we mean by processing? In Neural Networks, a neuron changes incoming electrical impulses and modifies them to different impulses. In our context, processing can range from as basic an activity as annotating and highlighting important sentences, to performing critical analysis of claims, adding historical context to a particular newspiece, identifying conceptual similarity between atoms and solar system to pretty much any other cognitive function that is possible.
After laying this foundation, we will move on to exploring the network in the real world. If the 2012 Delhi Rape case happened in the presence of such a network, will India be better off? I argue that we will be much better off. Will we be able to publish textbooks in Hindi, Bengali and Nepali? I showcase the different stages of the production.
Taking the first step
To end, I will describe how and where do we go from where we are. We have an online Wikipedia, whose Indic language versions are not yet strong enough. From here, to go to constant processing of all relevant information of India, where do we start from? The answer, according to me, is in a few browser plugins. I will supplement the design with case studies of reCaptcha and DuoLingo.
http://vignette2.wikia.nocookie.net/animorphs/images/a/a2/The_Attack_Illustration.jpg In the above book, the Animorphs meet an alien species which remembers everything that their species has ever seen. Every member has access to a collective brain. Compare that to humanity, which lost the 1500 years because it forgot the things that it had learnt. Even the things that we know right now as a species, are not really accessible to us all. We still behave in ignorance. In view of this, the global brain gains prominence as an idea.
- Result
Review comments[edit]
The review committee is unable to make space for this proposal in the main program schedule. This does not mean that your proposal lacks merit. We would encourage you to register for the conference and make use of the opportunities to present and discuss your ideas in several other formats available. We will have places to write and display ideas, places to discuss, tables for demos and space for break-out sessions with interested participants.
- Thank you. :-) I have broken down the talk into simple case studies and will look forward to impromptu conversations as well as Lightning Talks at the Conference. As of today, I have registered for the conference. On second thought, the subject is much better treated in chunks instead of a large presentation.