How to Brainstorm better than ChatGPT with Knowledge Graphs and GPT 4 | InfraNodus Tutorial

hello everyone in this video I'm going to  show you how you can use a combination of   data visualization and chat jpt to develop your  ideas in an interesting way so that instead of   having an interface like this where you have  to read through lengthy responses you can do   it more iteratively in short fragments so you can  really focus on developing your own thinking and   not taking the reades that are produced by GPT  so keep watching to learn how it works we'll   be using the app that's called InfraNodus, that  I developed and to do that you can go into the   apps import page and choose AI ideation live AI  ideation and once you choose this you start with   the topic that you're you're interested in you  can either write an abstract or just like the   key phrase that describes what you're interested  in for me it's heart rate variability um because   I'm currently working on an app that would allow  people to measure it in ways that goes beyond what   what the fitness trackers can do and I want to  learn a little bit more about this subject for   instance and not so much to learn about it because  I already know what it's about but to see what are   the specific nuances that I should take into  consideration when I'm developing this whole   discourse right so for example here I'm starting  from the basics I type in heart rate variability   it gives me an explanation of what that is so if  I see something I like that I want to add into   my stack of ideas I'm going to save it to the  graph here is just a definition of what heart   rate variability is as a measure of variation in  between every heartbeat used as an indicator of   physiological resilience and emotional stress  this is great because it reminds me directly   you know how I could also Market this app that  it can increase your resilience and adaptivity   and reduce stress here I can generate more  responses um but basically it's say more or   less the same thing but maybe in different words  I'm going to save another one because I like the   idea of phys phys The Logical flexibility that  seems very interesting too so I'm going to save   that as well and then what happens here is  when I want to develop it further um I have   some thoughts in mind so I can actually write  my own ideas for how it can go further or I   can also choose some stuff on the graph like  for instance emotional resilience heartbeat   and let's see maybe another word variability I  can select the terms I like on the graph that   I would like to be better connected and then I  go into the inside panel and just select this   one again so they marked and then it can generate  content related to these four ideas so it's just   going to come up with some thought that links  this idea of emotional resilience and heart   rate variability because that's the aspect that  I'm really interested in the connection of this   body marker to how we feel both on the level  of emotions and also physiologically right   so here here it says emotional resilience can  significantly influence heartbit variability   indicating an individual's ability to adaptively  respond in stressful situations this is great this   could almost read like an explanation of why  this app or this approach would be interesting   so I can save it into the graph generate  more responses emotional resilience can   be gsh through heartbeat variability a stable  heart rate indicates better Stress Management   and adaptability emotional so I want to show you  something here because this is not exactly true   I already know this the the subject so it made  a slight factual mistake it's not exactly that   the stable heart rate indicates better Stress  Management but that makes me think that maybe   I can go deeper into the subject and use the AI  to generate some ideas in relation to that right   so I will make a manual query and I'm going to  say okay what's the connection between HRV and   emotional and physiological resilience here I'm  using chat GPT GT4 mode so that usually generates   better answers and factual responses so here  it says how it's an indicator of physiological   flexibility a higher HRV signifies healthier  emotional and physiological resilience as it   reflects the body's ability to adapt efficiently  to stressor so this is great now I'm going to show   you something else I'm adding it into the graph  and I want to dive a little bit deeper into the   subject so I'm going to click elaborate here and  then by default it proposes me to make sort of   like an instruction or a prompt for chat jpt  or GPT 4 to develop this further right and the   default one is elaborate on the statement I  could also write explain why this is the case scientifically I'm improvising let's see  if it's going to work out if it's going   to give an interesting answer or not sometimes  it takes time to think so here it says that HRV   is grounded scientifically in a nervous systems  operations specifically the connections between   sympathetic and parasympathetic systems so this  is great because now we're getting deeper into   the science of things and look what happens when  I add this into the graph it will add a completely   new cluster so here it's shown here with with  the light green color right which shows me that   my discourse is developing in a New Direction so  I was in this region first like on this pink and   on this kind of dark green cluster and now I'm  getting into this parasympathetic sympathetic   one right so there I can actually select those  and maybe also HRV and then again use this   function to generate some content in relation to  parasympathetic and sympathetic so I can develop   this cluster a little bit further and this is  basically how you would use the graph live and   this is a big difference from how you would be  having a conversation if you did that using GPT   you know like you would basically like I think I  have here a conversation on on HRV so I have to   read through all of this I have to find the stuff  that I want to develop further by reading text   here directly visually I can see where I should  make connections I'm seeing this new cluster I   know that I want to connect it to the main idea  so I just select some nodes and ask it to do that   for me right and again this is not to make AI  think for you this is just a stimulator to come   up with some fragments of ideas which I can then  later rephrase myself make new connections between   them and then generate some original content  based on what I'm thinking about so here it says   it's kind of like a general statement interact to  influence okay that's great so here is explaining   me the biological sort of reason why it works like  that and I can again elaborate how it works uh biologically so you see here I'm giving it a  prompt of what it should do with the statement   that it itself generated so this is great because  it's a real human in the loop interaction with the   AI where I'm making it produce ideas but I'm not  making it write a whole textt for me I'm always   in control of every paragraph of every statement  it makes and I'm only choosing the stuff I like   and I'm using it to inspire myself to see in which  direction it could be interesting to develop this   discourse further right so here it's explaining  exactly what happens that the parasympathetic   system slows the heart rate promoting rest and  digest functions and the sympathetic system   accelerates it for flight of or fight responses  this balance create HRV and indicator of physical   and mental resist resilience this is great I  could even use the sentence in my own text or   in my promotion materials and so on right so then  I can generate some more stuff here is just kind   of like a rephrase of that but I'm going to keep  that also because that's interesting and finally   I can also generate actually more responses and  see if it comes up with something completely   different on another iteration sometimes you  might run it a few times to see if there's an   interesting thing to do here okay so then  one other thing that I like to do is to be   like okay I look at the graph again let's let's  take a zoom out View and this I always encourage   people to do it in foren notes because this is  what the graph is really good for now you see   things on a really detailed granular level you  see the main ideas like variability heartbeat   parasympathetic sympathetic nervous systems  and so on but this is a very granular level   so sometimes it's nice to zoom out and to see see  what are the main ideas you're working on so here   I click on High level ideas you also have the  button here in the analytics panel the analytics   panel is available here right and then it shows  me what this conversation is about all right so   I see that it's about Hardware at variability  stress adaptation resilience autonomous nervous   system so this is great because it gives me a  very clear idea of where the conversation is   going and also makes me think of whether I'm  missing anything I see I'm talking about the   actual thing I'm interested in and also stress  and resilience makes me think that I want to go   more into technical details and this is great  because this is exactly what this graph allows   me to do to always have this sort of feedback  reflection on my thinking process and if I use   it right if I ask myself what am I missing not not  only what's there but what am I missing then I'm   going to develop ideas in a very interesting  way so here I'm going to say okay maybe let's   focus on the technical aspect aspect and I will  write a question how is a heart rate variability   normally calculated by the way you can also take  the current context into account but I prefer not   to do that because I like to keep the answers more  open because then I can cover more territory at   the beginning and only later I start connecting  them together and kind of trying to make it more   coherent and this is also why the graph is really  good because it shows you which which part of the   process you're in somehow you know now we're just  throwing some ideas on the mental map um trying to   expand as much as possible at some point we will  need to start also connecting the ideas better   because we want the discourse to be coherent  and not all over the place but to be focused on   one thing so this is what we will do later so  here it says that HRV is typically calculated   using time domain methods measuring intervals  between successive heartbeats or frequency the   main methods considering distribution in various  frequency bands so this is good I'm getting to   technical details I'm going to save this research  further on those here it's explaining me a little   bit more what those time domain methods are I'm  going to save that as well for example standard   deviation of interbit intervals and I know that  for example Apple watch uses this approach but I'm   going to ask it to elaborate and please explain  if any other methods can be used in addition to   the ones listed below okay so I'm going to give  it its own response and then ask it to say to me   if we can develop it further and find some other  methods that are used in science to analyze hrp   and here we get to the more interesting stuff like  for instance nonlinear analysis like the trended   fluctuation analysis or Point curve plots that can  be used for assessment so I'm going going to save   this this I'm going to also edit it a little  bit so I will remove P care plots and let's see how we can elaborate on this why would DFA  be interesting in this context right so again   I'm using a prompt to specify how we can develop  an idea that was given out by Chad in interesting   way here it's explaining how DFA this nonlinear  analysis techniques adds a new dimension okay so   this is it just says it's unveiling potentially  enlightening Insight so a little bit too General   nonlinear lens unveiling concealed aspects in time  domain and frequency domain method and the third   answer intricate complexity so let's say like  which aspects are unveiled spe specifically and   let's go into more Precision here and by the way  you will also see that the deeper we go the more   topics the more specific topics appears here so  here we see that um it's adding hidden elements   so so it's revealing some hidden elements so let's  say we we save this and we decide to ask how does   DFA DET trended fluctuation analysis work and why  it can be interesting for measuring hrb okay save   this into the graph and I'm doing it all in real  time so bear with me if it takes a little moment   but I'm doing that so you can also use the same  workflow in your own Explorations and here we say   that it's a statistical approach for detecting  l long range correlations in Time series data   so there I get more specific and and let's see  if there's something else and again something   else so maybe I like this first one long range  correlations in sensor in in Time series data   so I just choose that part and ask it to elaborate  it further so as you can see it's generating some   ideas fragments of those ideas I'm taking them  in saving what I like and asking it to develop   it further and here you see it says that it  can unveil statistical nature of seemingly   random Trends um is an inovative technique that  unveils the presence of long range correlations   mathematical intricacy okay and then another  one unveils hidden long range correlations   and recurrent Trends so this is interesting this  is what it's doing actually it's it's un bailing   some hidden Trends and long range correlation  so I can save this as you can see I'm developing   this part on the hidden correlations and this  is the part that is maybe also something very   special for my content or for the project that  I'm working on because I really want to focus on   um sort of providing more data than other tools  or or approaches do so this is great because it   helps me find where I could specialize somehow  right so here I see this hidden correlations   and maybe I can even remove HRV and then hide it  from the graph slice it off and then I can also   hide time because we talked about it a lot heart  rate variability and what it enables me to do is   to find what are the topics that I could develop  for further so this I really recommend you to do   during the interaction that you have with this  tool that once you get or or you cover the main   topics which might even seem generic because you  already know what they're about you remove them   all from the graph to see what other the topics  that you should go into further so for example one   here I see is is this one is very interesting  is is the connection of HRV to resilience and   psychological and emotional well-being this  is great and then this technical aspect of um   sympathetic nervous system response balance so  maybe we can even sort of focus on this two and   to do that you go into blind spots and you can  highlight the gaps between different ideas in   the network so the Clusters which could be better  connected but are not and then ask InfraNodus to   generate some idea for you that would connect  them both so for example here we're lacking   some connection between autonomous nervous system  and heart rate variability and this is because we   removed all these terms but this is great because  we will force it now to generate some idea that   would connect those two clusters together  I could also do it myself so I could look   at those two topics and think of a connection  myself I can also use the eye and generate some   interesting question that can stimulate me to  think further in this direction right so here   it's asking what are the physiological responses  of the sympathetic nervous system that influence   emotional resilience and heartbeat interval as  indicators of typically fofl function and their   impact on overall body health so this is really  good because it's allowing me to go much deeper   into the specific aspect of how our nervous system  works and why HRV is an important metric so I'm   going to save this into my notes this is another  function you have here that you can save it into   notes if it's some idea that you want to develop  further um I don't want to be part of this main   discourse because this is just a kind of like a  notepad of ideas but when I get get to something   that I know I need to research further then I'm  going to save it into notes and go on with that   and then if I'm done with this Gap by the way I  can also turn on their names and then find another   Gap some stuff that are missing so for example  stress adaptation and data analytics so this   is great seemingly unrelated topics but exactly  because they are unrelated we will generate some   idea that would connect them in an interesting way  and Link them together so so let's see if there is   some interesting question that comes up what is  the correlation between traditional approaches   to creating healthy balance of stress and dynamic  pattern signifying efficient and good long-term   outcomes and statistical Series so this is a bit  too much like just using the keywords to generate   something but I always use it as inspiration to  make me think how I could fix this question to   make it interesting because actually the worse  is the phrasing the more interesting is going   to be the outcome because it's not so good the  phrasing because those topics are not related   but this is usually where the good ideas are  right so it's kind of like a balance you have   to strike between connecting things that are  hard to connect and also making them coherent   so in this case I'm going to see what are the  correlations between Dynamic pattern of stress   healthy good lifestyle in the long run so this can  be interesting to explore further I can save it to   notes or I can ask the system to elaborate on this  topic and generate interesting content in relation   to this question that it asked itself to come up  with some interesting idea so here it's talking   about um fluctuations and how more variability  means that you can handle stress better so this   I'm going to save in the graph because it's quite  interesting and also into my notes as well because   I find it useful for further discourse so in the  end once I go through a few iterations I can get   the nodes back into the graph that I removed and  then I see the full representation of all my ideas   and if I have to come back to this graph any time  later for further developing it into some text or   maybe some ideas or marketing material I can  always look at the graph see what it's about   see what the main topics are and then usually  I like to focus on some idea at the periphery   maybe here individual ability I will select those  notes to see in which context they were used so it   shows me here here exactly the sentence that uses  individual ability to adaptively respond and then   I'm going to develop this discourse further by  either thinking how it can go on in this direction   or by choosing to use gp4 to generate some ideas  um in relation to this topics and by the way I   could also do very simple thing by taking the  current context into account so once I select   those and take the current context into account  it's going to generate something that will also   relate to heart rate variability so the original  topic that I've been talking about you see and   then when I add this you will see that it links  everything on the periphery to the center which   usually makes the text more coherent so these  are some different approaches that you can use   to generate ideas you see we built like a very  sort of comprehensive and coherent at the same   time mind map of ideas by the way this measure  topical diversity shows how how diverse it is or   how focused it is at the moment it's medium you  have an explanation of what that means when you   click here but basically you want to aim for the  yellow or the green because if it's in the red   it means it's too biased if it's in the blue it  means it's too dispers so you want to be somewhere   focused but also to have enough diversity in  your ideas uhif you're writing let's say an   informative text right so this is how you would  approach this try it out on InfraNodus.Com and   let me know how it works also subscribe to this  channel so you can get informed about new videos   when they come out and if you have any questions  or comments please let me know and thank you

Channel: Nodus Labs

Views: 512,603

Keywords: chatgpt, gpt 4, ai, infranodus, thinking tools, productivity, ideation, creative thinking, creativity, research, hrv

Id: XMwa9q6wWkw

Length: 22min 8sec (1328 seconds)

Published: Fri Oct 06 2023