 이것도 रब नहीं बार ने आब वो सकते हैं। अब अबी सब लोकोने कता नहीं फीप या है ना अबई सब निखलेगा ना गरी आए थे नहीं बार कोईने तेपा अगे, वो गो समकभी आपने नहीं वो नहीं सकते है और में सब क्यों करचा थी प्रचाह। करती थाँब �對不對 Ever Instead, तुरंगड है? घुह स Edison 능ा। तुरंगड बारा serlaughing और घर तुरंगड ठबटाा तयकी का दूत ना uitल थास pinned तबटाindi chakra अगम दीखािया सद руб का जूग स्थ PK नसी दिक लान्चिवर क् communion आ� skill अगर ऄाउन्ती, और नाख्पूर, श surt- दु� how it thought it thought from Nakpur, Small town earth, Shakthanda अगर माज जीदना, समथ से प्रदटिर एकॉजी में। And I am comfortable with posterity filters. अगर बघफ दुत, क्यदर के साच़ाट आश्दोन जीवाड़ा। That's my Twitter handle and GitHub handle. अगाडी आप वरी भी बागनागे चाद, माझा धेरे उताई गर वह की सागगम चाँ, आब आब गोईग। और आप दोगी एक ठाँ मैरे साहँगर, लास हुँँँँँँँँँँँँँँँ यूग रह Monga सब वाग Lux ताई मेरे बाख साग सई या पयोंझे आगध ज niżFI अखगे वाख � этом होगी मेंखneys के वॊँऌया light ुआप जोत चफ वो बच्चान ही पहलेम। एक मेंएлюч़कना गद अब दरंते कल JSON बनक है गद वीगच मैं रहोनहोंं भा newspaper ह felony . kavani और आगर बार नाईण, और और जोग चाल बदे लोग था. प्रिधि अब आपा दिगर सब करने गागर वार्च्ठाती थरित अब दोगत गगे. तोगागर औरमे धआखर जब आसाग शाथ रहानी आन्था. एो than एक और वाड़नुं कुलया वादिट होगा बिदादू जरी तोगे थी आप रबा याद बुद्गार निदा तरानी की ऽने रहर्तादी धेखना क्योंग नीशारा दूगती के ख़ूगती रहा है अजी तब अजो मुपनिदिया, आप आप ँन money ूग खुटिंग वॉईbook ।ंधिछ उगा वॉग़न कुड़ के आदा आप ओगे दूँ से आप यादन और narrower वॉद कुड़े आप शाम न की जाध के विड़गाड्स अक और तोप IKEA prehabilitation. Mobile xtures  YJाज़ुउउउउउउउउउउउउउउउउउउउउउउउउउउउउउउउउ  Stall  � only अभाजली लाग। और मुतर से वो जुगा मुताई क्रिर सेंजा कि साण जै लागा कर ऐसा, मुतलिका साण तब आप घे कुई सभेगतिवा है, वैंगा भाजली बाजली के सिल्ग, आपको, देगा उना तो वेंगा. from the रीक of the उक्रिदून थियानी शिखन लाँगार के लागार अंगर तुद अछैप्ते तुच्वाँजा, यह चीश़्ेद, देशा बोड़ालीग बिपीडी। अजिए और उसक्ताद अजिए जाद़ि यह में लाँगार जाद़ी टीेश कि पर स्पहने बाजदी सब भी थियानी वि And invalidating environment exists We need to intervene to work on both aspects of it And then achieve our results which can be stability which can be prosperity or abundance Cool The other thing that I want to focus on is ethnic knowledge By ethnic knowledge I mean that a particular section of people Or it can be majority or minority, we won't get into that But a particular group of people who align with a particular cultural idea Have some knowledge that exists within that group अद्तोने क्या वाग़ा सोप्रोंप, सब या सब हो... दो�書, मिर्न की वाग़ा सोप्रोंप, सब आब मैंग़्ा सोप्रोंप, सब आब मिख़ा है मुझेटा है, छल्दे झास्द, छल्ड़्ब, स्वृदे. fixes so if we take a group of software developers as a ethnic group let's say if we were a culture So we would have our things documented somewhere in some format and we would have that knowledge represented in some way like when we talk her at the conference the talk is the way of representation of that knowledge now when we sit to study that knowledge so when we study music it's called musically इसना काँगाकाँँतार्वेल्गा काहाँगाँतारा है तो आप छकतारा गगाँताँ आप एक चृटे। अगर विसा सावाबा विसा क्या व्ठाष of etnic culture अआदब आँ सोथा, कि से तो कानापे जाग़ रहों थे। जरगन देखा गया जाग़ागार जी से ख़ह जो। गेंथा, और से वान वावा लुगेज़े, आदब वावान, और करत्द्वान लाग घोभोट, काना को बरट्खरे लोगातिया थालौ accommodant under the real world. को बाधिया बहु Spell wouldn't fit but अगा सesters of parties today mostinois मुझ्नाने क्छी अलीजी। Έे सुबि के एक वोत को उगाता औ背 temos銄 सिए को Stones, को को सिए के... को वाो शंऍे लार सिए की आए. कुफ कि कर नहीं आपा cus size with 600 integrating. what are the languages that do you work on now i have another conflict that i talk about the natural spoken languages or may i talk about JavaScript or Ruby coming back to the bioscience approach what we did in melga tygur reserve was we identified the needs of the local people and we said let us address the emotional security aspect and also create a validating environment let us not таким tell them that we are tourist आदडियादिये खल्ड़ानानुँँँँँ. और दीवादियान कुहाँँँँँँँँँँँँँँँँ जम्गल प्री सफारी से लगे क्चछाआ terrorists are going for jungle safaris. विर्जानाना मदा। सजहां बुटो। आर्जाना दूबिया. विर्जाना विर्जाना पर्णोग अना सरीई षलत परड़ाड. नम्से तीनन बागत बन। सचंटर रची ऑरगानी च्यूक जो जित रब आप ख्शादाश। विर्जाना मादए मदायogram if you speed through the roads inside the tiger reserve or if you take a gypsy the nocturnal wildlife gets disturbed the mating patterns differ and conservation doesn't happen no mating no conservation so um so with this approach what we did was we brought all tourists at a farm location at a deep sky observation site we put our motorized telescope over there and we asked them to just lie down on the floor and watch the sky move and then we kept curating that event by telling them that this is that constellation this is the significance of it and then this is how it looks from the telescope and things like that it was pretty interesting for them um what it also meant was uh we didn't have to touch anything in the tribal habitat so migra tiger reserve has a lot of corku tribals and there is no electricity yet there so they are not exposed to mobile networks yet like they know what a mobile is they know the latest bollywood songs but they don't use mobiles they don't have television sets and they don't have a pipe water supply either so there's no western toilet over there uh they practice dry toilets or recently they've started using indian toilets with the swaj bhaiyat mission now how if if i combine biosocial approach with ethnic knowledge so if i were to do it in the context of an ethnic culture that is corku culture in this case what we did was we renamed the event like this first event was called messiah marathon this event was called ifill dodo so in corku language ifill means star and dodo means watching so we renamed it in the tribal language which has not been documented anywhere yet and at the same time we introduced stories from the folklore of corku on astronomy so it was ethno astronomy and we got the villagers to come and narrate that story to everybody and we told them that hey you know what the tourists are coming in for you you are not here for them so why don't you just tell about yourself and we had a lot of villagers come over and we did many of these so uh the sukir pilta is is is in gondi so gond tribe is another tribe so we did that in pens tiger reserve and this was done in gondi culture we got a lot of validation so we're talking about validating environments right so we got a lot of handwritten feedback hand written because there's no internet there's no technology there the villagers cannot they don't have access to online reviews so we got hand written feedback from the tourists and the tourists were super happy like they wrote like two three pages of feedback each of them and they were they this and also we what we also did was that we introduced the food as something that the corku tribe eats when we did it in mail guard so we didn't give paneer butter masala we gave them like if i were to tell that in punay it's pitla bakri so there we call it besan and we did a thicha with it and we had it under the stars so it was quite a nice experience which people don't get in commercial tourism or in mass tourism so people were pretty happy about it and they paid a lot of money also another validation that we got was in terms of finances so i come from nakpur our temperatures go to up to 48 degrees in summers so there's no tourism that happens in summers in the tiger reserve but in this case we got about 50 people into events across two events and the summer revenue was 30 000 for the first event and 45 000 for the next event so everybody had that financial validation also what it impacted was we created an alternative to nocturnal tourism and reduced human activity we built the social capital by encouraging and we did this with the forest department so there was a lot of trust that was built between forest staff and the locals they were they had been conflicting because the forest staff had always been telling them that you you are a criminal you you are cutting wood you are poaching tigers and you need to go from here so that trust started coming up there was an alternative livelihood that they started seeing as compared to poaching because if you can generate 30 000 in one night you don't need to poach a tiger that gets you 10 000 at the same time there was minimum on no side effect to the habitat now katha ka saar ye hai ethnic knowledge coupled with biosocial interventions which is about validation and emotional security for people just telling them that a whatever you are doing is okay we we can accept it can be used for conservation and that is the theme of the project now coming back to ethnic knowledge when we say that it is represented in regional languages and dialects when i traveled i started making friends by picking up some local language gestures like hi hello kese ho sorry kana kaya kya like how are kana kaya we started doing that so for example in nagaland there are 16 official tribes that the state recognizes so if you want to say hello there are 16 ways of doing it in nagaland it was pretty pretty interesting and very fascinating when i was at hornbill festival in nagaland at the same time there's there's a lot many other languages so when we start studying these language linguistic nuances so it is called linguistic geography or dialectology now speaking about the linguistic spread in india what we have is there there is the eighth schedule of the constitution of india which recognizes official languages in india so the scheduled languages as they are called we have 22 scheduled languages in india what it means as every state has adopted one language which is an administrative language for that state and every curriculum or every academic material would be printed in a national language in an international language and in a regional language a scheduled language now there are 45 more languages that have been demanded to be included in the eighth schedule but we have 19 500 languages in dialects in india those can be reduced to over 1300 rationalized mother tongue what that means is there are over 1300 languages that have been practiced within families and passed on through generations to each other those 1300 languages can be reduced further into 121 mainstream languages each of which is practiced by at least 10 000 people in the country so right now the representation of these languages is 22 scheduled 99 unscheduled even if we were to start and every language has a lot of material that that we can study we can derive inferences from and design a biosocial approach if we want to achieve conservation like i cannot go to a corku tribe and say that hey you know what abhi marathi mein batkar can't do that i cannot go to a gondy tribal and i can say so if i were to tell you what the differences are so if i if i have to tell if i have to talk about gondy where i come from i come from gondwana nakpur pulsing gondwana so um if i have to tell anybody that hey it was nice meeting you so in gondy the way you say that is um nakun nase mile masik atun khusi atun it has no connection with any language any other language that you can find khusi probably so the challenges that we have are these languages are underrepresented and marginalized there is no poor or no propagation so the only mode of propagation is spoken word the spoken word gets translated to the next generation and with every passing generation we are losing out a huge chunk of knowledge so it is susceptible and vulnerable to extinction we have we already have some languages that have gone extinct because all the newer generations have migrated to an urban landscape and they don't they have no clue about what the tribal recipes were or what the agricultural practices were or how housing happened there and what is the ecology there so they have no clue so um uh what becomes a primary priority for me is to first document and archive this knowledge because it is undocumented they don't even have a script so so we need to start from there now there are some ongoing projects which which are already doing something for the rural landscape so pari network which is the people's archive for of rural India it's already doing this within pari there is this grind mill songs project which was executed in this part of Maharashtra where we are right now in the western ghats the villages that fall in the western ghats so so people poor people who live in the hills so they have an organization that they formed and they documented the grind mill so a grind mill is that so the women of the marginalized women of the marginalized community used to get together and they used to network and socialize over the grind mill because they were not allowed to talk about these topics at home so the only way to document gender inequality in that landscape was through those grind mill songs so there was somebody who came in and documented those songs for 20 years and they archived 110 000 songs translated them in english and french and 30 000 of those songs have been audio recorded and have been put up as open data on pari network um has anybody gone through this website ever jashne rehta or rehta foundation okay so uh rehta has been instrumental in conservation of hindustani languages so urdu in the avadi braj all these languages that existed in adi kaal or bhakti kaal so they uh they have been documenting all that literature and they have been putting up in a very good ux which allows you to understand what it means like in this case this uh this ghazal is uh is uh is from amir khusro which was a 14th century poet uh he's also known as the father of ghazals kawalis and uh rubai so uh he was half in half hindu and half persian so he he wrote in persian urdu avadi braj everything so this particular ghazal has a very unique character uh the first three words are in persian and the last four words are in avadi so it goes so it's half hindi or sounds like hindi and half of it sounds like alien it's persian right uh which is which is what brings us to project langdon what it is about now i'll stop talking gibberish and i'll come to the point um project langdon is about four things that i intend to achieve uh one is documentation and archive what i just spoke about um um how that is done how that can be done so documentation can be done with a web app uh like rehta uh which whose primary function would be to transliterate and translate so in this case when you see zihale miskin and you see a hindi devnagri script and beneath that you see the meaning of it so the hindi script and the urdu script is the transliteration and the meaning of it is the translation um archiving is is about knowledge representation how do i represent it digitally how do i store it in a particular format that can be used to derive inferences out of it like why do i want to document it i need to answer that question do i need to do it just because i want to sound cool or is it for entertainment or for my ego boost i need to answer these questions so i need to do it to infer that knowledge so the way i did a bio-social intervention in melga tiger reserve i need to understand what a star means in corku that inference is present in that ethnic knowledge which is undocumented which needs to be archived and there needs to be a system because there is so much over such overwhelming amount of information out there 19 500 languages and dialects out there with a lot of literature a lot of meaning a lot of knowledge inside it how do i do it manually i can't do it manually so i need to put up a system for inferences and then based on those inferences i need to recommend a bio-social intervention to public bodies to private bodies or to organizations that lie like the gharib dongari dongri gharib lokanshi sangatana so if i were to provide them with a bio-social intervention i can recommend it to them that hey you know what if mass tourism is not working for you try night sky observation for that there needs to be a holistic knowledge that needs to be documented somewhere where i i intersect ecology astronomy ethnic knowledge linguistic geography and it becomes an intersectional approach it's not a single issue activism that i'm doing where i say that okay i want to conserve habitat and everybody else needs to just go away so i cannot be an environmental activist and ignore human aspect of it so it has to be an intersectional the the Venn diagrams converge uh they intersect they don't converge so i need to take an intersectional approach and then ultimately my aim is to conserve chronology samjiye uh ish mein jo basic technology ka evaluation tha ho chuka hai and right now i'm at a stage where i'm working on the documentation and archive part where i i'm working on the uh knowledge representation bit of it rehta has cracked the ux like nobody else so i'm not even thinking twice about redoing it i might just connect with them and understand if there is a collaboration opportunity which i have already initiated with somebody if that doesn't happen i would like for them to open source their ux and so that we can use and share it anyway it's a it's a it's a foundation and they they promote open sharing so so that is a sense that i have and if they are not doing react probably i'll do react just because i want to so anyway the next stage would be to create come up with an inference engine which comes which which churns out some inferences out of that knowledge like what what is it what what is something valuable that is in our context right now so if zehala miskin was written in the 14th century what does it have for us right now why is it important for us or why should i even read it so there needs to be something in it for us so the systems need to churn out that relatable inferences for that we need to train it and finally there's a recommendation engine which churns out recommendations for a biosocial interval my purpose is biosocial intervention but it can be used for anything else also like it can be used to to create a business model which is connected with the ethnic knowledge coming back coming to the tech aspect of it probably you would smile now so i needed something for ui development i needed something to have like this is a basic web app requirement like it's nothing fancy this is even if you have to build a web app it would be something like this like you'll have you'll need something to log in users sign up users forgot password email bejo ye karo wo karo adrak lesson okay so then you need something to persist that data when i'm representing that knowledge i need something to persist that data in a way that it can be inferred and recommended then i need something for a role based authorization mechanism where because i'll have some contributors i'll have some moderators i'll have some admins so i need a role based authorization mechanism somewhere where i define privileges either based on a group identity or based on karma rating or based on reputation the way it happens on stack overflow or wherever else then i'll need something to test can't forget that and finally i need something for continuous integration and continuous delivery later stage requirements would be inference and recommendation engine that is something which which connects the like it's the it's the crux of this this thing which serves the purpose of conservation and if i were to open it up i'll need a developer experience like i'll need a cli experience or something which people can just clone the repository and start contributing so this is something which i started thinking about earlier this year we are still in the second month so it's still early this year and i started working on the software aspect of it early in february like first of february i started working on this my evaluation parameters were how is the learning curve how much documentation is available what sort of supporting libraries are available what is the library ecosystem around it what is the developer adoption so that if i have to do it with other people how do i select the technology i need more developers i need community support and i need to balance simplicity versus convenience so just realize everything that is simple is not easy simple is not easy most simple things are very difficult to achieve so but in in the in that that whatever that that feeling of achieving simplicity and feeling god i might just lose out on convenience so i need to balance it out somewhere i cannot just you know by any jasakta all right um this was a question that i wanted to answer because i have primarily been in the view camp i always developed in view js i i never did react so i thought okay why not let's let's just go for an adventure with this but where react one was on the documentation and supporting libraries and the you know community and adoption bit so it sort of fulfilled most of this view is getting there now when i started on view there was not much happening on view but right now i see a lot of traction there so anyway i went with react just because buzzword compliance also serverless so you know i just wanted to sound cool this is what the technology stack looks like for ui development i picked a semantic ui for react um and there is this ui component explorer called storybook js pretty awesome stuff for user authentication i relied on aws cognito user pools i wanted to build a cloud native application i wanted to build a serverless setup so i did aws cognito user pools for the sign up and login flows for data persistence i went with aws dynamo db and graphql interfacing um uh because we we all know that javascript is winning ground in most areas right now so i wanted to stay closer uh to that so i didn't want to deviate very far away from it um so data persistence with graphql uh allows us direct connectivity with react or view or any other front end framework without needing a rest api in between so it it just removes that overhead of having a crowd api uh which is which can be easily fulfilled with graphql there are use cases where a rest api is required so it is not a drop-in replacement for rest apis but for most of my use cases i saw graphql serve the purpose for role-based authorization i used something called user pools cognito user pools and there is something called graphql transform so is there anybody who has done graphql in this in this room okay um so i have some slides on the on the tech stack uh at a later stage i don't know if i'd be able to cover it right now but they'll be available for your reference uh on azgeek site and it's it's it's all public domain material so uh for testing i use jest uh for javascript testing uh for ci um i i went with aws amplify console um if people who are not aware of what continuous integration is you push to a git branch and automatically a build triggers and you have this nice app deployed for you waiting for you to test it out so you don't have to actually deploy it manually and the reason i'm sharing these aspects with you is because if you are a front end developer and want to explore full stack development if you are a server coder but if you want to explore ui development if you are a database administrator and if you want to explore api development so there are these easier tools that are available where you can manage this without writing much code so the initial entry barrier goes away um okay a graph data persistence because i i need a inference engine and recommendation engine there is something called a graph database who all are familiar with this term a graph database okay so a graph database is something that what it sounds like like it's a database that stores data in the form of a graph how many people come from computer science background like okay so graph theory boring topic so uh that is applied somewhere uh for data persistence and all those shortest part traveling sales man yadaya kuch bayesian okay byzantine agreement border gateway protocol so all that is is implemented somewhere and it is available for real life use like real world to solve real world problems as i phalto meini tha wo and the querying interface for it the way we write sql queries uh for relational databases uh the graph querying interface there are many i use gremlin and sparkle uh sparkle it's called sparkle uh ui ux inspiration of course uh rehta um i might collaborate with them or ask them for permission to use their ux and develop something on the same lines now uh coming to knowledge representation so i'm not going to talk about the nitty gritties of how i set this up it's there in the slides you can go and refer and you can knock your socks off um knowledge representation um because there is so much knowledge that is there and i need a way to represent it knowledge representation uh happens to be an ai concept where um uh you represent knowledge in a way that computers and machines can assess it and infer something out of it we have a challenge that we have so many scripts written scripts so it needs to accommodate for that there is this uh transliteration and translation example that i wanted to quote so there is this uh denagri script that is written on your right there is a transliteration of it in english script in the english alphabet and at the bottom there is a translation of it um now uh i'll skip this part um now how do knowledge graphs look like so that translates to flower in english and it transliterates to phol in english script in the english alphabet so you can represent that as a graph and then you can add some meta information and say that it's a hindi word this is an english word and that's an english word and flower also transliterates to hindi hindi newspaper and then you can have directions translation is bidirectional transliteration is unidirectional and then you can have like a cycle within that graph where you have multiple translations in the same languages these are labeled property graphs it's a specific type of a graph where you label it that you say that all these are nouns so you can have a vocabulary of verbs you can have a vocabulary of adjectives you can have a vocabulary of any other grammatical construct now how do you create these graphs you create these in gremlin this is the syntax i'll skip this for the benefit of time you just do an add vertex add edge so what these nodes fool push flower these are nodes and or vertices of the graph and the things that connect are the edges so you create an edge you create a vertex vertex and then you connect it so there's a very simple syntax that you follow this is how it looks like it's got an id and it's got some labeled content to it and when i start adding edges it looks like this now what is this useful for so it's useful for traversal so what does flower translate to so it gives me a list of vertices now it's not human readable so what i do is i say that give me the content of it so i say what does flower translate to it tells me push for fool then if i say what does fool translate to so it says push for flower then i say what does fool translate to in english so you can go as granular as you want and you can have a translation and transliteration mechanism setup using a graph database with a simple graph query and all this is all this can be executed using lambda functions while being in your javascript comfort zone so you don't have to move away from it you just have to write it somewhere and you can trigger it from a node js application also so app sync which is a graphql engine it supports integration of lambda functions and you can connect various data stores with it it doesn't it does not have to be only dynamo db it can be a graph store it can be a key value store like it can be a memory store like radius and so on so lpg works lpg is not the subsidy while lpg it's it's labeled property graphs so it it works great for travelers and it's a candidate for oltp systems where a transaction is important online transaction processing so i want a quick translation of something i'll do a traversal now what is the other aspect of it so there is something called semantic graphs where i i say that a grammar can be represented in graph like asts like an abstract syntax tree so tree is a special case of graph so you can actually represent the language in a graph and construct a sentence out of it so this is a ghazal by another by amir khusrowit says so what it means is uh so uh i pick this example because it it connects ecology also like if i were to study climate change i can say that okay palasas stop blooming right now it used to bloom in the 14th century so climate change is real something like that uh then i can also add some information and say that it's a noun in the lpg you saw that it was noun was represented as a label there is no label here then i can make it further complex and i can say that ambua tesu sakalbans are so come from hindi language then i can say hindi comes hindi is also a noun and hindi comes from english language this script that i that i've written and english is also a noun and it comes from english language so there's a self-reference okay so this is a complex graph if you if you compare this with the lpg that we saw this is a more complex graph and it is used to answer more complex questions that what happens in today's season in sakalban what happens in the jungle in the jungle okay so um yeah sure so um i'll have to cut it short because there's no time um thank you so much for being patient one important slide um not really i mean i've covered most of it thank you