nuitblanche
Alfred-40B, an OSS RLHF version of Falcon40B
I am Igor Carron, the CEO of the company.
Our technology is different than the one used by Lightmatter and Lightelligence. We currently do not do integrated photonics.
We are starting from the ground up by first building a photonic co-processor that does one function (to be extended) of general interest to AI. the current function is a random projection. We believe that it will be complexified over time as we grow. But our particular focus is on those computations that are difficult to perform in electronics. In particular, our recent focus has been on Transformers/Large Language Models where we think our technology will be a game changer in the training and inference of these gigantic Machine learning models. For the training part, we published two papers at NeurIPS last year in the main conference. We also presented a more detailed one at one of the workshop: https://arxiv.org/abs/2012.06373 We are going to present a poster at HotChips this year about our foray in HPC: https://arxiv.org/abs/2104.14429 Other papers in AI are listed at the LightOn AI research page: https://lair.lighton.ai/
Physically, speaking, the Optical Processing Unit fits into a 2U rack (currently our focus has not been to miniaturize yet, we avoid permature optimization) that is connected to a CPU bus through a PCIe connection.
The technology has been running for the past three years anf half in datacenters for the oldest prototypes (we just decommisioned one a month ago or so) with Machine Learning computation unit tests being performed every ten minutes during that time period.
We are making the technology available for researchers to use on our LightOn Cloud cluster (https://cloud.lighton.ai/what-is-lighton-cloud/) and people can even rent one or several OPUs (LightOn Appliance: https://lighton.ai/lighton-appliance/ ).
I agree, we need to redesign our website.
Igor.
The Akronomicon: An Extreme-Scale Leaderboard of the Largest ML Models
Roughly speaking it looks like the nuclear power plants have three times as much installed capacity than solar and produce 6 times as much power to the grid. Since the nuclear power plants are near 100 percent capacity, I wonder what is the reason why wind has roughly a 50 percent production capability: is it maintenance or wind availability ?
Imaging Earth with the Moon
Hydrogen is not an energy source. It is not available in nature and has to be produced with .... power.
A reference page for Matrix Factorization
Since a great many algorithms behind collaborative filtering, clustering and so forth used by the likes of Google, Facebook, LinkedIn rely on advanced matrix factorization and this field is seeing some important growth. I just compiled a list of different recent algorithm that have sprung up as a result of faster convex optimization softwareand other field such as compressed sensing. It is...
Welcome to the Matrix Factorization Jungle
Are Perceptual Hashes an instance of Compressive Sensing ? http://nuit-blanche.blogspot.com/2011/06/are-perceptual-hash...
(ps: I write a small blog on CS).
Finding the source of the German E-Coli outbreak with datamining
500 euros (US$ 650) for whoever can solve this business model brain teaser
start from the bottom.
For mor eon the controversy with the Wired article: http://nuit-blanche.blogspot.com/2010/05/compressed-sensing-... http://nuit-blanche.blogspot.com/2010/03/why-compressed-sens...
The new reconstruction solvers: https://sites.google.com/site/igorcarron2/cs#reconstruction
Hardware that are implementing compressive sensing: https://sites.google.com/site/igorcarron2/compressedsensingh...
Ask HN: Your most awesome experience (UI/computation/language...)?
Here are three examples of what I call awesome experiences and while they are a little dated, I still cherish them. They are also very easy recalls as I can say exactly what changed at the moment I first used these models/processes:<p>* Back in the days, I used to write most of my graudate research with Fortran generated results... the process was excruciating. Then one day, I discovered...
Ask HN: What do you do to get into the "Zone"...
....and how much time does it take to get there ? Also do you use specific apps or specific processes to recall what you have done before ?
I have some experience in the process of doing the filtering you are mentioning as I write a blog on compressed sensing (http://nuit-blanche.blogspot.com/), a field that spans a wide range of communities that are, often, not talking to each other nor have the same culture.
I don't think Mendeley or any RSS feed services help much as they both target the same dead tree people who are generally about two years behind on the good stuff. Arxiv is good but not all the communities use arxiv either and so it really is a mix of different tools that yield the most results.
I have a high interest in expanding and automating my current processes and would be very much willing to talk to somebody who would want to spend some time on this.
I also agree with one of the posters, how much do you think people would be willing to pay for this service ?
Ask HN: Mindmaps for blog entries: Is there a code for that ?
I run a blog that is highly technical (http://nuit-blanche.blogspot.com) and I pay special attention to link back to previous entries relevant to new posts. However, in the past few months, I have run into the issue of trying to find back some entries but could not do that very well using the search engine that sits on top of the blog and I am wondering if the following piece of software...
Ask HN: Hundred of thousand of people stuck on the wrong continent.
Is any anybody interested in developing a solution that would help these travelers sutck on the wrong continent ?<p>For some academics, here is what I am proposing in my community:<p>http://nuit-blanche.blogspot.com/2010/04/cs-lets-make-lemonade-out-of-lemons.html<p>Could it be extended to a larger set of people ? Your thoughts.
Volcano Ash Cloud Advisories
Why do you go for a single piece of hardware when in fact, you could already do this on several GPGPUs and one CPU or use several CPUs on the cloud.?
It could probably be done using these techniques but the answer resides in the fact that if the computation of the phase transition is done very fast for a particular "measurement matrix" (i.e. a particular "sensor") then we could foresee its use in near-real time in instances where the "measurement matrix" would be time dependent (such as the atmosphere....). This would be important in that it would allow to figure out if these new "sensors" can be actually used (as opposed to being mere lab constructions).
Ask HN: What is your entry:"What would you do with 48 cores ?"
I am curious. Anybody else care to share their entries foe the 48 cores giveaway AMD contest( http://blogs.amd.com/work/2010/03/03/48-cores-contest/ ). The deadline is passed.Here is mine: http://nuit-blanche.blogspot.com/2010/03/what-would-you-do-with-48-cores.html