Leveraging DeSci and DeFi to Breach the Bastion of / Arnaud Castillo

Transcript

and how that is a potent mix for quantitative research in financial applications so look forward to hearing you talk Arnold Castillo take it away thank you okay um we hear you loud and clear yeah I'm hearing twice the song if it's possible or otherwise I will try

to we're checking right now oh thank you very much so in the meantime hello fries I'm really thrilled to be with you today and we are gonna speak about the centralized science and D5 so just a few words about crunch though crunch down we are a platform that connects from one side financial institutions and from the other side uh researcher from all around the world

um and what we do basically is that we predict the stock market you can go to the next slide Maybe so um basically it's very difficult to generate High I mean I mean generating high yield in finance it's possible especially through artificial intelligence but it is very very very challenging why because you need a specific expertise and few people have these skills it's I'm consuming you have a lot of people

that are working on the same thing and the problem is that the findings don't last because if all the researchers on the planets in the world are working on the same things they tend to converge to the same findings and basically they are doing the same discoveries and unfortunately after the the your findings that allow you to find this Alpha we call it on the market so this yield is uh not lasting so um if you can go to the next slide basically it's not like uh detecting a face you know uh or detecting a cat face a dog

face this is very very easy we call it uh I mean very easy it's difficult but it's another it's another difficulty but to detect for example face it's called phase detection and there is something that we call the signal to noise ratio so basically in a detecting a phase the signal is very high it's a very good signal so it's very easy to detect these face unfortunately if you go to the next slide you will understand that it's for stock it's way way more harder because it's very difficult to understand how one stock will perform versus another in the next week in 30

days 60 days and 90 days so to solve these issues because we are not smart enough we decided to build a community of researchers and so today if you can go to the next slide you will see that we are uh 2 000 I mean we are more than two thousand this was the previous number 21 and a half months ago because now we are 3500 and there is around 700 phds on the platform and because we are not smart enough to beat this I mean to to solve this problem it's forecasting problem we decided to build this solution uh with

the best financial institution and universities in the world so now what we proposed to our clients basically is to access thousands models on the Fly you can go to the next slide and these thousands model on the Fly can be accessed and can Empower any financial institution in the world so we turn any Financial um machine Financial problematic question into um pure machine learning exercise so no more specific knowledge is needed and

there is research phase and any financial institution can access this model easily so that's what is pretty good um we believe in a decentralized research model where communities will use web free tools like crunch down to borrow their common goals so if you go to the next slide you will see that crunch Dao has a framework that basically gives you provide a decentralized science capabilities I will come back to it in the next seconds the crunch dial platform that is able to aggregate financial institutions and crunch Alpha which means that any

researcher who has who finds yield in the market can share his signal with the community um so basically you can go to the next slide we are collecting thousands of moles and we Leverage The Design framework so coming to you I think it's important that you see how we do this can you go to the next slide again perfect so here what you can see is the way we are collecting the model so we have a variety in the of the community is very diverse we have people coming from more than 50

different countries they have difficult different backgrounds for example they have machine learning Computer Science Background astrodynamic statistics mathematics physics I mean they are coming from a lot of different background and on top of this they are using different methodologies so for example some will use clustering some will use regression some will use a neural network and in the end we have a lot of different combinations because the the education from the country you belong is one variable your background I mean

academic background a mathematician are you a computer scientist are you a physician and what methodology do you use so if we aggregate all this then we have a lot of deep diversity in our in the models we collect if you can go to the next slide yes so this is very interesting you know in the in the end of the last century there was a someone who whose name was Francis Galton maybe you know him and he he wanted to prove that the crowd was stupid

so he brought an Oaks in in the middle of a village and he asked it's true story he asked any um all the people to guess the weight of the Oaks and in the end of the day at the end of the day he was very happy because he was able to show that people did not find the solution you know no one guesses exactly uh the the weight of the Oaks yeah well then he came back at home and then he did his calculation and in the end what did he see he see he saw that doing the average of all the the the the the guests people found the collective intelligence

found that the this very accurate result I mean it was perfectly uh it was at two kilo it was I think two kilos different with the real weight of the oak so it was very very accurate and this is what we are doing this crunch uh leveraging the wisdom of the crowd we are asking them we are believing in time and we believe in diversity we believe in collecting diverse models from diverse people and and send something in them of course there is a mathematical retreatment it's not that easy of course it takes time but this is this is what we do

then if we go to the next slide there is um software that we built that is gold that is called that we called the design crunch down so it's a bit maybe difficult to see here but uh you can so basically everything because we are down it's very difficult to build knowledge incrementally and we need this because we are doing research but when we have you have 3 000 researchers on your platform it's very difficult to have a scientific uh uh clear and structured scientific discussions

so we came up with this solution where anyone can upload a paper basically a white paper is on the is on this uh platform where anyone can comment and you can just sign in with your uh metamask maybe there is a no unfortunately I don't I can't show you here but you can just sign in with your metamask upload your code and then you will sign your the the the your your paper with your metamask anyone can edit it uh change it suggest to change then any change is then put it and uh this makes it a real

decentralized Community doing research together so that we can build incrementally um we can go to the next slide because during this episode with this approach we are solving three problems we are solving the problem of incremental knowledge because our framework rules researcher and Alpha provider to discuss up through timeline signals by building easily on previous discoveries there is a also a reproducibility crisis we say because the problem is that if you have some finding it's very difficult to reproduce it and our technology enables

reproducibility by Design and of course anything you change is sent then to higher ipfs so it will be forever basically so we are able to there is no more problems with immutability and transparency if we go to the next slide one word about the tokenomics because I think it's important to understand how we did this tokenomics and what is the value in the token um basically we have on the top that you can see we have the Dao who is collecting and codes and models so

predictive codes these codes are having an output so predictions and uh are then sold to hedge funds or any asset manager and we basically we help them build their portfolio without prediction and because our the diversity and we call it also the orthogonality that we are providing we improve the performance of any financial institution or most of the financial institution I would say um then we take a performance fee on this on this performance and the financial

institution pay the doubt we swap the Fiat into uh crunch token and we pay back the the the the community with the performance of the research so the more the research perform of course the better results we get then the better than the financial financial institution that are working with us are performing better and so they are giving more food to the Dao Etc uh uh yeah yeah and I guess that seats for this short presentation I hope I was not too long thank you very much any questions

looks like no questions right now thank you very much though all right one more time thank you years [Music]