Whatever your question is, math already has a map to the answer
Bogdan Bocșe29:532020
Math can already answer almost anything: the hard part is asking the question
A virtual library holds every possible page of 3,200 characters, and any sentence you type can be found at an exact shelf and page. For anyone building AI, teaching math or fighting misinformation, the bottleneck is no longer the answer. Bogdan Bocșe, a solution architect and managing partner of AI startups Knosis.ai and Envisage.ai, asks what kind of shared language would let humans and machines frame the right questions.
Learn how math can teach kids isomorphisms, expose hidden privilege and help fight fake news
Bogdan Bocșe argues that mathematics already maps the space of possible answers, so the real challenge for humans and machines is learning to ask the right questions. He shows how simple math ideas can teach young children abstract concepts, expose hidden privilege and inspire a more inclusive shared language, then answers audience questions on starting with AI, fighting fake news and building digital twins.
“Because math can provide the answer, but the challenge is, can we provide the question?”
“It's not the machine's job to decide what's wrong and what's right for us.”
What you will learn
Reframe problem solving around asking precise questions, since mathematics already maps much of the space of answers.
Explain abstract concepts such as functions, bijections and isomorphisms through simple visual stories that young children can follow.
Use prime factor decomposition, as in Eugenia Cheng's thought experiment, to map hidden layers of bias and privilege.
Start experimenting with computer vision in the browser using TensorFlow.js, then define a concrete objective before choosing tools.
Evaluate tamper-proof publishing records as one approach to making fake news harder to deny or erase.
Details
Level
All levels
Format
Talk
Technical depth
Introductory technical
For
All career stages
Length
29:53
Published
July 2020
Questions this talk answers
Where should a beginner start with AI and machine learning?
Bocșe recommends starting with TensorFlow.js in the browser, using TensorFlow for server-side work, and then defining a concrete objective to zoom in on.
Can technology alone solve fake news?
He believes AI and blockchain can help, but a lasting solution needs a change in human mentality and discipline, for example a permanent record that stops publishers pretending they never published something.
Can math be used to create a person's digital twin?
Almost: he expects technology to create a shallow copy indistinguishable to others within five to ten years, but not to reach what is inside your head, such as a secret you never tell.
Is lambda calculus the inclusive language he is describing?
It is one tile of that larger language, but it does not handle cases where an association is not a complete function, such as a superposition or a probability distribution.
Why is the growth of programming languages a problem?
Human minds cannot sustain that growth in diversity: nobody can learn all of them, and they become harder to manage over time.
About this talk
Mathematics seems to be discovered, nonetheless, the same way the perfect form for a cutting tool was discovered by primitive humans. Just as the tool, mathematics still represents a concoction of essential secrets of the Universe and circumstantial limitation of the human body, mind and perception.
Speaker 1: Hello, everyone. I'm happy to be here and I'm honored to give this session. I'll be repeating this a couple of times until everybody joins. Please feel welcome to ask any questions. I know that the platform doesn't allow me to ask questions. For me to see the presentation and the questions at the same time. So after about 8 or 10 minutes, I'll try to take a break, a very short one, to look at the questions and to try to integrate them in the rest of the presentation.
As I have written in the chat panel for this session, I want this to be interactive and I want to encourage you to interact with the material I prepared. You also have the slide deck. That's the last link. So you can also make a copy or take it offline if you feel like it. And then I propose a small game that has to do with what do you want to build that is useful for the human citadel.
Speaker 1: So the theme is somehow in parallel to the presentation, which I will begin shortly. And the topic of the presentation is whatever your question is, math already has a map to the answer. Speaker 1: Now, I propose we wait for one more minute, then we get started. How do you feel about that? Are you ready to tackle the math monster?
Speaker 1: Okay, cool. Well then, let's get started then. I'll be sharing my screen. And let's go on. I'll just check with you briefly if you can see what you're doing. Speaker 1: Let's see my screen. I'll have to exit the presentation in order to ask you and then I'll come just to make sure. So were you able to see the presentation? Was it all good? Cool. Let's go back to it then.
Speaker 1: Okay. So I like to think of myself as the chief executive geek. I run two startups. As you can see from the names, Gnosis AI and Invisage AI both have to do with artificial and augmented intelligence. And I've Speaker 1: I've had a lot of experience in dealing with big data solution architectures, both in the startup and in the corporate environment for the last almost 14 years. Also, I have dedicated my part of my efforts to Deep VISS, which is an open source initiative that aims to make both machines and humans communicate more flexibly and more easily about their objectives.
Speaker 1: As I said in the introduction, I propose to you a little game of imagination while we go through the Speaker 1: Through the presentation and through the topics. Of course, you don't have to do it now. You can always do it after the Speaker 1: After the presentation if that's how you're more comfortable. And I'll make sure to follow up to any questions you may have or to come up with any suggestions on the ideas you may have. Now, I Speaker 1: Mathematics is often a boogeyman of education. It's a subject that's hard to reach and hard to follow. And most students are usually dissatisfied with their experience in mathematics, whether they're primary school or university level. But other than being a little literal pain, as reported by many, mathematics is a repository.
Speaker 1: Is a way that the best way we found to organize the many experiences that we've had in this universe in Speaker 1: In order to make sense of them. In many ways, mathematics is the ultimate essence of the language we were able to build. Now, the fact that mathematics is a language is not necessarily a done job. And I'll give you a small example. There is a thing called the library of Babel. You can look it up online. It's available.
Speaker 1: And you can play with it. It's a virtual library, which contains all the books that could be written in pages of 3,200 characters. And basically, this is an array of possibility. The thing is, you can actually look things up in this library. And you can get the exact chamber and the exact wall shelf book and page where the text is located.
Speaker 1: Of course, if you do the search several times, you'll come up with the exact same position for the book where the text is written. Now, the fact that all the possibilities have been generated and ordered into the structure of chambers and walls and shelves, leaves us with just one challenge. Speaker 1: What is the question we ask? Because math can provide the answer, but the challenge is, can we provide the question? Now, let's take a small break from math. Because we want to put together a little bit of the answer.
Speaker 1: We don't just want to have something theoretical. We don't just want to have something theoretical. So let's use our imagination. Speaker 2: And consider two questions for the imagination. Speaker 1: And consider two questions for the imagination. First of all, what is the question we ask? Speaker 1: What is the question we ask? Because math can provide the answer. But the challenge is, can we provide the question?
Speaker 1: Now, let's take a small break from math. Because we want to put this in a practical setting. We don't just want to have something theoretical. Speaker 1: So let's use our imaginations instead and consider two questions for the imagination. First of all, we keep hearing about the singularity, the moment when machines will become just as intelligent and more intelligent than us. And basically, the singularity means that we don't know what happens next.
Speaker 1: Now, the other thing I want to bring to your attention is making first contact. You've seen this in movies. There was a movie in the early 2000s called actually Contact with Jodie Foster, which explored the possibility of alien. Speaker 1: What would you say to an alien if you first encountered it? What would our first message be to an alien species?
Speaker 1: And I encourage you to take that question in for a second. What if you were the ambassador, if you were the astronaut who represents humanity for the first time in front of a different species? Speaker 1: What would you ask? What would you write? How would you say hello? Speaker 1: Now, I'm going to take a different approach to the same situation. Let's just say that I know a super company announces that they created AI. It's done. And you're a journalist who is tasked with interviewing this first form of artificial intelligence of artificial sentience. What would your three questions be?
Speaker 1: Assume that language is not a barrier. Assume that the machine speaks your language and we're assumed that the machine understands all the languages you understand. Still, what would you ask? Speaker 1: Now, a lot of these topics have to do with the conventions that our species, our civilization has made. Speaker 1: In order to express experience in both speaking and writing. And one of these conventions is the geometry or the topology of writing. How do we, what is the direction of our writing? What is the shape of our writing?
Speaker 1: And as you may know, these writing styles are pretty heterogeneous around the world. There is no one correct way to write. And this shows us from an intelligent design point of view that all the languages that we use, including programming languages, are particular cases of something more, more generic. Speaker 1: Now, programming languages have been around for a while and their number, as you can see in this diagram, which just shows a small number of them from the 1960s to the early 2000s. You'll see that there's this whole web of languages of programming languages. Now this, this looks more and more like the picture I showed you in the previous slide.
Speaker 1: Maybe you recognize here, the Tower of Babel. It is the story, the mythical story where wanting to be higher and higher and closer to the Supreme Being. People built a structure that was ultimately unsustainable. And to a greater extent, what we're doing with the technology in the notion of language is at this point, we're going to see what we're doing with the technology.
Speaker 1: In the notion of language is at this point unsustainable because human minds cannot sustain this growth of diversity. Nobody can learn all these languages. And they're becoming harder and harder to manage as we go on. What I showed you in the previous diagram was the situation until the present day. But the number of languages, programming languages and programming languages are so that we're not going to be better.
Speaker 1: And that is not a sustainable pattern. However, movies like Contact and Arrival challenge us to define language writing symbols in a way which is more inclusive of the potential semantics of the universe, which is more inclusive of the potential semantics of the universe, which is more inclusive of the potential semantics of the universe. Speaker 1: Which is more inclusive of the message or of the description or of the problem that we want to express clearly without doubt and without fake news.
Speaker 1: Now, I want to give you more specific examples of what I and my colleagues at DGSS think could be done. We have created this model that can introduce mathematical concepts. Speaker 1: To juniors and young children using a simple story and the simple story, which you can, which you can also see online on my profile and on my blog is a story of several families of animals living on the same island.
Speaker 1: And there's a very easy to use symbolic representation of these concepts. And ultimately the story that we want children to imagine in order to have a deeper understanding of abstract concepts is a story of all of these animals being forced to move on different islands. Speaker 1: Now, what this visual experiment goes to show is that very abstract concepts such as functions and bijections and isomorphisms and all these sort of things you learn in college level math can be introduced at an intuitive visual level to kindergarten children.
Speaker 1: Because what we want to make sure is that they have the mental flexibility to envelop the many languages that today has to offer. Now I'm going to take a short break and see if you have any questions or if you have any topics that you want to propose moving on in the presentation. Speaker 1: I do want to keep the schedule. I know you're all busy. So, uh, what are your questions regarding the use of mathematics to define a more inclusive and more efficient language that would allow us to at least be aware of the discrepancies.
Speaker 1: At least be aware of the discrepancies between one culture and another one context and another. Speaker 1: Okay. Well, in that case, I'll go on to the next part in case you don't, you don't have any questions. Speaker 1: Um, let's start again. Speaker 1: No, this example is not my own. It belongs to Dr. Eugenia Chang. I, uh, cited the source in the, in the bottom of the page and it's a simple thought experiment that has to do with decomposing numbers in their prime factors and in their factors and ultimately in their prime factors.
Speaker 1: Uh, you can see that you can write this, uh, pretty simple problem of what are the factors of 30. Speaker 1: You can write it in a visual way representing the origin of each one of these factors, basically, uh, to decompose this number into its components. Now. Speaker 1: You, you may know that, uh, prime numbers are extremely important in everything from cryptography, uh, blockchain probability prediction.
Speaker 1: Uh, but this next example. Speaker 1: Is not very intuitive, or at least it was very surprising to me when we analyze a very rich and very complex landscape. Speaker 1: Of bias of diversity of privilege. Speaker 1: It's easy to get, uh, hung on our own perspective. Speaker 1: And it's easy to, uh, forget that some of the ways we are biased and privileged.
Speaker 1: Are most of the times completely hidden to us. Speaker 1: However, if we take this more structured approach that is borrowed from a very simple example in mathematics. Speaker 1: We can get a wider, uh, understanding of. Speaker 1: Honestly, how biased or potentially biased, uh, we are because of our context of our constraints and of our privileges.
Speaker 1: Now, I think the greatest challenge that our generation faces is to define. Speaker 1: To define this new language. Speaker 1: To define the way it should work. Speaker 1: So that it includes. Speaker 1: What several diverse sentience can express. Speaker 1: It needs to be a language that allows for the flexibility of natural language.
Speaker 1: Where flexibility is, uh, desired. Speaker 1: But it should be a language that emphasizes, uh, lack of coherence or that, uh, shows any incompleteness or any controversial statement. Speaker 1: Without disallowing it, without denying it like a programming language, but rather by directing the attention of the reader. Speaker 1: Saying, hey, this piece right here is just a hypothesis.
Speaker 1: Or this piece right here has been contested by a great number of readers. Speaker 1: It's not the machine's job to decide what's wrong and what's right for us. Speaker 1: But it is our responsibility to task the machine with being able to bring to our attention what is most likely to be relevant to our decision.
Speaker 1: Now. Speaker 1: There are a lot of open questions. Speaker 1: Of. Speaker 1: How we use the definition of a new language for the purpose of inclusivity, eliminating bias, reducing the impact of, of fake news and, uh, closing the skill gap. Speaker 1: That is leading our societies, both from a gender perspective or race perspective.
Speaker 1: An income perspective from our point of view, having access to learning the language, a new language, whether it's a programming language. Speaker 1: Or a natural language, foreign language, let's say. Speaker 1: Shouldn't shouldn't be a luxury. Speaker 1: Should be a public digital service. Speaker 1: So in the end of our presentation, I think we have about two minutes left.
Speaker 1: Let's hear your questions and your challenges regarding what would you like to build that is useful for the human city. Speaker 1: I see a question here. Speaker 1: Um, with, uh, Lambda calculus, uh, Lambda calculus is, uh, is an alternative way of, uh, of looking at programming that. Speaker 1: Uh, keeps in mind the idea of the completeness or the closure of functions.
Speaker 1: Now, while Lambda calculus is, um, is definitely part of this greater language or, or, or of this more inclusive language paradigm. Speaker 1: It by itself does not tackle all the challenges. Speaker 1: And, uh, just to be, uh, just to give you an example. Speaker 1: Um, one of the, um, one of the potential challenges that, uh, Lambda calculus does not tangle is what happens when a closure isn't complete.
Speaker 1: When you can, you have an association between two types of elements, but it does not define a function. Speaker 1: Maybe defines a superposition, maybe defines a, uh, distribution of probability. Speaker 1: Uh, now. Speaker 1: It's definitely to answer your question. Speaker 1: Uh. Speaker 1: Lambda calculus seems cleaner. Speaker 1: Uh, indeed it is part of the greater language, but it is just one of the tiles that we need in order to reach our objective.
Speaker 1: Uh, another question I see is, uh, regarding, um, tutorials for AI for beginners. Speaker 1: And I actually had something prepared here. Speaker 1: I'll, uh, I'll, I'll show you a quick demo on open source demo as an example. Speaker 1: And if you want to dig in more, uh, sure Indira, it was my pleasure to answer your question.
Speaker 1: Uh, I'll paste in a few links here in the time we have left. Speaker 1: And if you want to know more, feel free to track me down on LinkedIn at me and ask whatever you need. Speaker 1: I'll do my best to answer in a timely fashion. Speaker 1: The example I, I pasted here just to make sure that everybody has a little bit of fun in the last, uh, in the last minute is, um, is an example.
Speaker 1: You can start it right now. Speaker 1: It doesn't break anything. Speaker 1: It will just ask for access to your webcam and you will be able. Speaker 1: It doesn't transmit anything in the internet. Speaker 1: Don't worry, but you will be able to see that, uh, artificial intelligence algorithms for computer vision now can run your browser directly.
Speaker 1: You don't have to go to a server to send your data for it to be processed. Speaker 1: So if I were starting today as a machine learning enthusiast, I would definitely start with building all of this in the browser, taking what's already built, uh, in TensorFlow.js. Speaker 1: I recommend TensorFlow.js. Speaker 1: See whatever other solutions.
Speaker 1: Uh, yeah. Speaker 1: Uh, one, one second. I'll take the fake news questions as well. Speaker 1: But I would, to answer, uh, what to begin with, start with TensorFlow if you want server side stuff, but also see what you can do with TensorFlow.js in the browser. Speaker 1: Because there is absolutely no more reason to put late the latency between the browser and the server on the critical path of user experience.
Speaker 1: Now, as you can see from the demo, uh, it moves pretty fast because, especially because it doesn't have to wait for the server to do the computation. Speaker 1: So start with TensorFlow.js and try to define an objective for what you want to build. Speaker 1: Let's talk about what you want to build because it is very hard to answer someone who's asking, Hey, what's the right tool for my job?
Speaker 1: If I cannot have a conversation about the nature of the job of the objective. Speaker 1: So start with, uh, something genetic like TensorFlow.js and then try to define an objective to zoom in on while you go on. Speaker 1: And, uh, I'm not sure if, uh, the session will be cut off or something, but, uh, to answer the next question, uh, on the idea of fake news, I'm, I'm going to state a pretty controversial, uh, opinion here.
Speaker 1: I think that machines can help in, uh, solving the fake news epidemic. Definitely AI can help. Blockchain can help both of them, but at the same time, I am confident that any complete, uh, permanent, uh, elegant solution to the fake news problem requires a significant change in human mentality, uh, human knowledge and human disciplines on behalf of the users.
Speaker 1: Specifically, uh, one of the simpler ways you can tackle fake news is to, uh, create the system blockchain based or PKI based. The technological solution is not that important, but the idea of the system is look, if you see a website, if you go on a website, you, you can underline part of what is written there.
You can, uh, a screenshot is taken both from your computer and from a different source. So you can't mess with the data and you write on that blockchain that, Hey, this item at this moment, uh, is controversial. Speaker 1: Or here is the source that here are the sources that say it isn't so. Now, if you find this way of making sure that creators of content can no longer delete the content or rather, uh, if you find a way to make sure they can no longer pretend they never published it.
Speaker 1: If you make what is published part of a permanent record, even if it's deleted. Sure. It's not visible, but it can be proven that it has been written. 80% of the, of the fake news job is done. Part of the problem, uh, of fake news is not necessarily lack of technology. Um, part of the biggest challenge is, uh, people's lack of willingness to engage that technology to pursue new publishing models and new business models.
Because to be entirely honest, the people who make money, make money out of this the way it is. So it is up to us to define a new game with slightly new rules that are, uh, easy enough for people to, uh, want to follow them, but, uh, well-defined enough for, uh, tighter, stricter definition of truth, because, you know, in the old days, when you wrote something in the newspaper and you sent it out, you couldn't take it back.
And it, and, and, and freedom is not a, is not about being allowed or not allowed to say something. Freedom means that everybody should be allowed, but freedom also means the responsibility, not to lie about what you said. And that level of trust has escaped media, has escaped, uh, social media, and to a greater extent has escaped digital society.
And it is up to our generation to define such tool and to build it. Because otherwise we can't expect the machine to clean up after our lack of discipline all the time. I hope that provides some answer to the question. Uh, sure. Alexandra, please connect, uh, on LinkedIn. And the invitation is open to anyone else who, uh, has other questions.
I do, I am in no rush. I'm not exactly sure how this session goes. But if you have any questions, I'll be happy to sit around and take them one by one. Uh, hope I didn't miss any of them. Uh, yeah. I, I have, I saw a very interesting question for, uh, Deepan Shusetya. Um, he, he asks whether, uh, I think that, um, math can be used to create one's digital twin or digital copy.
And the short answer is almost. So I think technology can be, uh, uh, current day technology can be used in the next five to 10 years to create a twin that is indistinguishable to others from the original. And so basically I think our technology today permits us to copy, uh, to make a shallow copy of a person and to, uh, have an integration of, let's say deep fakes with generated reality and with ray tracing that can go a very long way for simulating appearance and conversation and so on.
But I don't think we yet have the technology to go on the inside. So basically if you create a digital twin using this, uh, mathematical models, which I'd be more than happy to discuss offline, uh, then that twin will be recognizable to you because at the very last you will be able to set the password or to know a secret or a password that you never tell anyone and that you remember, and that will allow you to distinguish yourself from the twin.
You ask him, do you know the secret? And he'll say, no, I don't. And I don't think we can go deep enough yet with the technology we have to find the secret. You don't want to tell, but let's, uh, let me, uh, rephrase the answer this way. I think when, within the next five years, we will have the technological capacity to create a twin that will manifest all the features you choose this twin to manifest.
But not all your features because fortunately between the capability of the machine and what is in our heads, there is still a barrier, uh, of free will. And I think we should protect and enforce that barrier, not just try to break it all the time. Okay. Um, one moment. Double checking for any questions. Well, uh, if you have, uh, no further questions, I'll, I thank you for your time and your patience.
And if you are interested in embarking on this journey with us, uh, we'd be very happy to connect, to discuss any specific objectives that, uh, we may directly or indirectly help with. And make sure to, uh, double check the game because I think that if we help each other define what we can build for our common good, I think we will, uh, create a more coherent digital society.
Thank you very much.
Transcribed automatically; names checked against the speaker record. Select a time to jump there.
Article based on this talk
If math already holds the answers, can we learn to ask the right questions?
Bogdan Bocșe on math as a language, teaching abstraction to children, hidden privilege and the human discipline that technology cannot replace.
In short
Bogdan Bocșe argues that mathematics is the best language humanity has built for organizing experience, and that it already maps much of the space of possible answers. The harder task, for people and for machines, is framing the question. He uses that idea to propose a more inclusive shared language, to show how young children can grasp abstract concepts, and to argue that problems like fake news need human discipline as much as new technology.
Key takeaways
Reframe problem solving around asking precise questions, since mathematics already maps much of the space of answers.
Explain abstract concepts such as functions, bijections and isomorphisms through simple visual stories that young children can follow.
Use prime factor decomposition, as in Eugenia Cheng's thought experiment, to map hidden layers of bias and privilege.
Start experimenting with computer vision in the browser using TensorFlow.js, then define a concrete objective before choosing tools.
Evaluate tamper-proof publishing records as one approach to making fake news harder to deny or erase.
Math as a library of every answer
Bocșe opens by acknowledging that mathematics is often the boogeyman of education, a subject most students find hard to follow at every level. He offers a different view: math is a repository, the best way humans have found to organize what they experience so it makes sense. "In many ways, mathematics is the ultimate essence of the language we were able to build." (4:29)
To make the point concrete, he points to the Library of Babel, a virtual library online that contains every book that could be written in pages of 3,200 characters, as the speaker described it. Search for any text and you get the exact chamber, wall, shelf, book and page, and the same position every time. With every possibility already generated and ordered, only one challenge remains. "Because math can provide the answer, but the challenge is, can we provide the question?" (5:38)
He then invites the audience into an imagination exercise. If you were the first human to greet an alien species, as in the film Contact with Jodie Foster, what would you say? If a company announced it had created a sentient AI and you were the journalist interviewing it, with language no barrier, what would your three questions be?
A Babel of writing systems and programming languages
Writing conventions vary widely around the world in direction and shape, and Bocșe sees that diversity as evidence that every language we use, programming languages included, is a particular case of something more general. Showing a diagram of programming languages from the 1960s to the early 2000s, he compares the tangle to the Tower of Babel, a structure that grew until it was unsustainable. "Nobody can learn all these languages." (9:55)
Films such as Contact and Arrival, he says, challenge us to define symbols and writing in a way that is more inclusive of what people need to express clearly, without doubt and without fake news.
Teaching isomorphisms in kindergarten
Bocșe and his colleagues at DeepVISS built a model that introduces mathematical concepts to young children through a story about families of animals living on an island, with an easy symbolic representation. The children then imagine the animals being forced to move to different islands. "Now, what this visual experiment goes to show is that very abstract concepts such as functions and bijections and isomorphisms and all these sort of things you learn in college level math can be introduced at an intuitive visual level to kindergarten children." (11:35) The goal, he explains, is mental flexibility to embrace the many languages the world offers.
Prime factors and hidden privilege
Borrowing a thought experiment he credits to Dr. Eugenia Cheng, Bocșe shows how the factors of 30 can be drawn visually, tracing each factor back to its prime components. He found it surprising that the same structure helps analyze the complex landscape of bias, diversity and privilege. It is easy to get stuck in our own perspective, he notes, and many of the ways we are privileged are hidden from us. A structured decomposition gives a wider view of how our context and constraints shape what we see.
A language that flags doubt instead of deciding truth
Bocșe calls defining a new language the greatest challenge of his generation. It should keep the flexibility of natural language where that is useful, while highlighting incoherence, incompleteness and controversy: marking a passage as a hypothesis, or as contested by many readers, rather than rejecting it the way a compiler would. "It's not the machine's job to decide what's wrong and what's right for us." (16:38) He adds that access to learning a new language, whether programming or foreign, should not be a luxury but a public digital service.
Audience questions: AI, fake news and digital twins
Asked about lambda calculus, he calls it one tile of the larger language, but notes it does not handle associations that are not complete functions, such as superpositions or probability distributions. For AI beginners, he demos a computer vision example running entirely in the browser. "So start with TensorFlow.js and try to define an objective for what you want to build." (21:32)
On fake news, he states what he calls a controversial opinion: AI and blockchain can help, but a lasting fix requires a change in human mentality. He sketches a system where anyone can mark a passage on a website as controversial, with independent screenshots recorded so publishers cannot pretend they never published it. "Because otherwise we can't expect the machine to clean up after our lack of discipline all the time." (25:39)
On digital twins, he expects technology to produce a convincing shallow copy of a person within five to ten years, but not to reach inside our heads. "And I think we should protect and enforce that barrier, not just try to break it all the time." (28:25)
Watch the talk
"Whatever your question is, math already has a map to the answer" by Bogdan Bocșe, about 30 minutes: https://www.womentech.net/video/bogdan-bocse-whatever-your-question-math-already-has-map-answer
About the speaker
Bogdan Bocșe is Managing Partner at DeepVISS and has worked as an IT professional for 12 years in roles such as solution architect, project manager and technology consultant. He is managing partner of Knosis.ai and Envisage.ai, leads the open source, non-profit DeepVISS initiative bridging industry and academia in visual machine learning, and has filed patents in Romania and the US.
Questions this talk answers
Where should a beginner start with AI and machine learning?
Can technology alone solve fake news?
Can math be used to create a person's digital twin?
Is lambda calculus the inclusive language he is describing?
Why is the growth of programming languages a problem?
Drafted with AI from the talk transcript. Editor review pending.
Frameworks and ideas mentioned
Lambda calculus
Prime factor decomposition thought experiment (Eugenia Cheng)
Resources mentioned
Library of Babel (online virtual library)
TensorFlow.js
TensorFlow
Contact (film)
Arrival (film)
Cite this talk
Bogdan Bocșe. "Whatever your question is, math already has a map to the answer." WomenTech Network, 2020. https://www.womentech.net/video/bogdan-bocse-whatever-your-question-math-already-has-map-answer