New users are flooding into HTM Forum, so let’s get some discussion started! How about we get to know each other a little better? You might take some time to fill in some details about yourself in your profile, or just post below describing who you are and what you’re working on.
How did you hear about HTM?
Have you read any papers or watched any videos about HTM?
Did you read On Intelligence or see Jeff’s TED talk?
Do you follow my Twitch account to keep up with the latest research?
So I’ll start…
I read On Intelligence in 2006, but I never thought I’d actually be an engineer at Numenta. But here I am! I’m one of those rare software engineers who’s technical, but also an extrovert. You might say I’m a techtrovert.
I’ve been involved in open source projects since I started programming over 10 years ago, most notably on the Grails and Mojito projects. I was so excited to help Jeff Hawkins make NuPIC open source back in June 2013. This is a dream job for me, and I love helping people in our community build awesome things.
Recently, I’ve been live streaming my work on an interactive tutorial called Building HTM Systems. Before than, I spent a lot of time working on an educational series of YouTube videos called HTM School.
Binggo! Interesting here!
Hi, every body, my name is vaxin, I’m a Chinese, working at Beijing.
I encounter On Intelligence at 2015, after that my dream about A.I. fired again. So I continue to dig even everything about the secrets about brain and A.I., My goal is to seek an architecture that works like the human brain so that we can create various of assistants for us.
Now, I grew up and has own much more knowledge about this thing than one year ago. The nupic-theory mailing list helps me a lot. As the HTM Forum sets up, I hope in the future, we can discuss together and do more things for A.I.
Best wishes for all of you.
Vaxin
I’m Kevin and I live in St. Louis. I’ve mostly lurked on the mailing list but will occasionally offer unsolicited, unsupported, and sometimes irrelevant opinions. I’ve been thinking about brains since the late 80’s (blame Douglas Hofstadter), eventually getting a Ph.D in neuroscience working on nonlinear signal integration in cortical dendrites. I’m no longer doing science professionally (slinging code is more fun and pays better) but I keep an eye on the field. I’ve been following Numenta casually since more or less the beginning–I had several friends at Redwood so I kept hearing about this Hawkins guy. I’m honestly something of an HTM skeptic but I very much look forward to being proved wrong.
Ivy here from the 50th State in the USA, Hawaii. I too have been lurking for about a year now. In 2008 I found On Intelligence randomly in the library stacks and could not put it down. I have been talking about going back to school for my PHD but never found the right program…that has changed and am working on my application and proposal.
Thank you to all of the members on the mailing list, everyone here has inspired me to move into the next phase of my life, albeit broke but happy!
My name is James Chu, and I am probably one of the most “odd ball” in this group. I am a retired computer engineer now living in Shenzhen, China. I am interested in this group because through my years of researching into human behavior, I created what I called “The Simple Science of Being”. It is not psychology, philosophy or religion, but really a “science” in understanding and modifying human behavior. It originated from Werner Erhard’s work (David Ray knows about this), but I went much further and created an implementable form of science. I have always thought that the model I created can be applied to AI research, hence the interest. What is even more incredible is that back in 1996 when I first came across the writings of Mr. Hawkins, I already found that his idea actually matches with mine. Then years went by without any further news from him until I came across Numenta last year. Obviously my approach is completely different from brain anatomy, but rather from a “functional” point of view. In fact, my model can actually explain why SDR is the ONLY way to implement AI. Any way, at this early stage of AI development, my model is not very useful. However, I could imagine that at a later stage, my model would be invaluable in creating an AI that can actually behave like a human being, even complete with emotion. I will be more than happy to share my discovery with all of you in the future.
Fun fact: I am an artist!
I have very little experience in programming (I haven’t made any apps yet) and I learn on the fly through the Internet and staying active in forum based communities like this one!
I found out about numenta a few weeks ago while I was doing some research on streaming content via an RTMP server (although I found a better solution with SAAS!). I am very passionate about technology. I believe that through communities like this, humanity pulls itself infinitely closer to technological singularity. Every day, I dedicate some time to research and brainstorm ideas for developing an AI to manage my IoT. Although experimental in nature, my goal is to create a product that allows me to, first, manifest my ideas through conversations with my AI (a personal AI), and then develop more products with the AI that will allow the AI autonomous control of the products (all controlled through me of course). I believe that the Cortical Learning Algorithm will enable this possibility when programmed to study how I think, as I think, from a general level to a detailed level. As an artist, every moment in time is valuable because every moment is a place of inspiration for something new. By giving hardware the ability to see the world as I do through software, I’m essentially bridging the gap between possible and impossible.
I could write for days on this topic (I’m very passionate about technology) but I’ll leave it to whomever wants to hear to reach out to me .
I’m excited to be a part of this community! Thank you for putting it together @rhyolight.
Feel free to send a message my way if you want to talk!
I’m Matt! I’m originally from Germany but live and work in Shanghai, China. I read “On Intelligence” and the white paper beginning of last year. Ever since I’m fascinated by AI and the question of what intelligence actually is and how to replicate it in a machine/software.
As a techie, I wanted to understand the nitty gritty details of HTM and CLA, so I developed my own implementation (in C). Then, to compare with other machine learning approaches, I also started digging into deep learning and built convolutional neural nets.
I find HTM very compelling and probably closest to tackling AGI. I appreciate the inspiring discussion, especially in the theory forum, and look forward to continuing exploring together.
I’m Rosh from London…I currently work in hotels.com as a data scientist. I am currently reading the paper on HTM after reading the book on intelligence. I am very fascinated by the CLA algorithm and want to try it out on our data at some point!
Hello everyone! It’s good to see all of you here.
I’m Silver Wei, a PHD candidate from Harbin, China. The promising CLA offers me a effective tool of solving my problem. Although there’s lots of other deep learning models, it is HTM that expands my mind on artificial intelligence problem. May someday in the future, I could contribute to this community, albeit no extra experience in programming.
You can find an earlier version of my book “The Simple Science of Being” in Amazon, author: Jim Kay Ping Chu. I am planning to revise the book in the future, to make it even more logical. However, I must caution you ahead of time that my book can only be understood through practicing the method, and not by just reading it. This is necessary because of a built-in limitation of the human mind. If you want to know more, you can send me an email: jc20031212(at)qq.com.
Hi, everyone
My name is Jiang hao shen ,and I’m a graduate student major in electrical
engineering. I’m now doing a research project about anomaly detection , I
actually found HTM by chance when randomly typing things at Youtube, the
ability of dealing with temporal sequence and scalable time and spatial
hierarchy is what attract me ,because I think the definition of anomaly
depend on temporal and spatial context.I’m currently learning Python ,hope
to apply HTM in anomaly detection in image problems, I’m very interested in
the differences between applying HTM in 1-d data and images. If anyone has
tried it , please send me email and we can discuss ,thanks!
My name is David Ragazzi, Brazilian, son of Italian, Portuguese and Irish forefathers. Since I was a child, I’m passionate on Robotics (yeah, due to movies like Terminator and Matrix), however due to life circumstances I had to pursue a IT career. I’m graduated on Business focused on System Analysis and recently I gained a Master degree on Software Engineering. I always loved challenges and help people, because this I found in the computation a way of keep doing these 2 things at same time! After going through a personal crisis related to my current work situation, I came to think about what really matters for I being REALLY happy. It was when I realized that I deviated from my old passion and then I decided to pursue what really matters and makes sense for me. This said, I fell “accidentally” in the neuroscience field, because as many here I feel the most robotics and AI did not follow a biological approach to be design their products. During my personal research, I saw we neither have a theory about intelligence in order to we have robots really intelligent. Then finally, I read about Jeff Hawkins and his work… and here I am!
My contributions to the community:
Translation of “On Intelligence” to Portuguese.
Translation of CLA White Paper to Portuguese.
Full convertion of the NuPIC old build process to CMAKE which decreased complexity and the number of files in the repository.
Active participation in the efforts to make NuPIC Core independent of NuPIC.
Active participation in the restructuring of the NuPIC Core repository.
Active participation in the restructuring of the NuPIC repository.
Active participation in the efforts to make NuPIC installable with PIP.
I think that my next contribution will move HTM a (huge) step foward, but this is a subject that I will give more details only when I have a stable version of my work (soon).
I have worked with R&D in the last 10 years. I am currently R&D engineer for a TV manufacturing industry.
I very interested in machine learning applications and for this reason I ended up finding Numenta during my research.
I’m trying to use Nupic in some personal projects related to the financial market and so far it seems very appropriate to what I need.
As I’m just starting out, I still have much to learn about the HTM and their use for prediction problems. I hope to count on the help of experienced users in this forum.
Regards,
Ruan Belém
P.S. Hey @david-ragazzi, can I have your contact? Maybe we can have a talk in Portuguese.
Kent Langley. I read On Intelligence when it came out and it had a profound impact on my thinking about machine learning and AI. I’ve been learning, staying aware and researching on the topic ever since and looking for ways to apply the technology.
I run my own technology advisory and research firm ProductionScale, Inc. It’s been in operation since 2012. I have built and managed several companies in the last 21 years and helped lots of people do that say. Usually from a technology perspective.
As faculty, I teach in the Data Science and Exponential Organizations (driving organizational scalability with the judicious use of advanced technologies and processes) tracks at Singularity University. I have presented on Numenta, Nupic and derivatives of it many times in my lectures at SU.
Recently, I’ve had a new child (our 3rd), been working with a South American company to use data science to increase protein production for human nutritional quality improvement, worked with a Chinese marketing company to use Data Science / Software to help with Ad placements and scale operations in chinese e-commerce contexts, taught one of the largest chinese supply chain manufacturers in the world a new view on Data Science for their organization, worked with a mexican insurance company to help them understand how to apply technology and new business models to drive toward breakthrough innovation and some other things as well.
I’m very excited about the overall potential for HTM technologies and Matt, you in particular, have been phenomenal! So, thank you.
Interesting to read the introductions of very interesting people!
I am also an artist and a technologist. Studied engineering and later acoustics and music tech postgraduate. But my goal has been becoming more and more to use knowledge and the technique for the creation of pieces that serve as communication, exploration of ideas and dialogue.
Been interested in Numenta for almost a year now since I started researching it for an art project in which I used NuPIC to analyze audio to show the reactiveness of a system in the context of language. Notice, not language processing.
To be honest, ever since I started dwelling in the train of thought of this technology, or I must say, algorithm, I started having lots of understanding and ideas about the way we process information, how language works, how we create meaning, and movement, and more interestingly, in the way ideas evolve, language evolves and culture as a body of ethereal knowledge is carried and loaded into our brains by means of these abstraction layers. Fascinating!
I’m a CS PhD student in Australia, my work is focused on analysing road traffic data. Initially I used data from vehicle detectors beneath the road to infer incidents and other anomalous behaviour. My proof of concept used nupic as the main engine and won the 2015 HTM Hackathon. There’s still plenty of room for improvement there, but now I’ve moved onto a similar project using IoT devices to track people movement around a city. Hopefully I’ll be able to use HTM to detect anomalous people movement!
I fully expect that we can eventually replicate a human brain’s operation inside a computer given enough resources and accurate algorithms.
My name is Stan. I’m from Bulgaria, by way of NYC and Boston, and I live in San Francisco. I studied electrical engineering and computer science at MIT, where I got interested in brains, AI, and computational science.
I found out about Numenta in 2009 or 2010 from one of my friends who had interned there. I read On Intelligence and interned there myself in the summer of 2010. At the time, we were focusing on computer vision and learning transformation-invariant visual features. We started a reading group to go through Hubel’s “Eye, Brain, and Vision.” I helped prototype some of the spatial pooling algorithms. I bugged Jeff and Subutai with lots of questions. It was a great summer!
After I finished school, I worked at Twitter for a few years, where I built systems that detect trends and breaking news, and studied networks and user safety. More recently, I worked at a company that applies deep learning to medical image diagnosis. Now I am working on training machines to generate music, and other independent projects in AI.
I am particularly interested in mastering the recent advances in neural networks and understanding how the popular deep learning methods relate to more biologically plausible approaches like HTM.
Hello! I’m Andreas and currently live in vienna. I came across HTM-theory 1 year ago during research for a formal description of “intelligence”. The theory makes a lot of sense for me and gives a much more realistic explanation how the brain could work than I have seen so far. To get a better understanding I created the old Temporal Pooler in Matlab (at this time I didn’t understand the difference between the old and new Temporal Pooler). Also I have read the most papers on HTM-theory posted in this forum (when it starts with too much neuroscience I’m lost).
Recently I finished a master in applied mathematics and now I’m working on several projects with friends. One of the projects is to figure out how to define concepts like predictions, stability, hierarchy, information in relation to SDR’s and how to put them together for a perception model (for machines). Also I’m working on an implementation of this concepts but until now it’s more theoretical. In the future, I want to work in a startup based on HTM and other ML techniques to create problem-specific intelligent systems.
I’m always looking for people to talk about HTM and AI from a formal and algorithmic perspective. If anyone in Europe wants to meet some day to exchange ideas or work together on some projects, contact me:)
.