Back in 2011, Marc Andreesen announced that Software was eating the world while everyone was trying to make sense of the realities of the cloud versus brick and mortar businesses. Eight years later, Tarry Singh articulated how AI was eating software; a year before GPT-3 and Codex would give solid ground to this prediction. Fast forward two years later, we just witnessed how AI ate HPC and we believe those are the first steps towards how AI is eating Learning, Creative and Office work.
Let me explain.
At LightOn, we have been working on getting AI to be transformative for everyone. For that to happen, we used the Jean Zay French national supercomputer for two different yet somehow related reasons this past year. First, our LightOn’s Optical Processing Unit hardware was integrated into this top105 supercomputer. Even though LightOn’s hardware is analog and uses a technology currently unknown to supercomputing, there are several good reasons the future of computing will use this technology. Relatedly, in a co-design fashion, we also used the Jean Zay facility to implement and run code for the building of Large Language/Foundation Models that we believe are key to Transformative AI. In March, we trained the largest French language model ever called Auriga and made it available to everyone through our PAGnol demo.
In July, we launched the Muse API, making our language models available for business use. Initially released in private beta, Muse has quickly gained its first customers, and a public commercial version with five languages is to be released in early 2022. Some of these early customers are using this new AI to redefine SEO or the experience for website creation.
“True happiness comes from the joy of deeds well done, the zest of creating things new” Antoine de Saint-Exupéry
Eventually, a major impact of these Large Language Models trained on HPC infrastructures will be the ability for everyone to personally learn faster and for office workers worldwide to get the job done in a fashion never seen before.
If you are a start-up company or an individual starting a business around this promise, don’t hesitate to join the Muse Partnership program, and let’s start a discussion around how Muse can help you.
These models will also have the same effect in creative work and in the discovery process.
Stay tuned, the true AI revolution is really coming!
I am an algorithms researcher at Google (http://theory.stanford.edu/~rinap) and I am organizing this workshop on "Conceptual Understanding of Deep Learning" (details below). It's trying to understand the Brain/Mind as an algorithm from a mathematical/theoretical perspective. I believe that a mathematical/algorithmic approach for understanding the Mind is crucial and very much missing. I'd appreciate any help I can get with advertising this on your blog/mailing-lists/twitter.
Goal: How does the Brain/Mind (perhaps even an artificial one) work at an algorithmic level? While deep learning has produced tremendous technological strides in recent decades, there is an unsettling feeling of a lack of “conceptual” understanding of why it works and to what extent it will work in the current form. The goal of the workshop is to bring together theorists and practitioners to develop an understanding of the right algorithmic view of deep learning, characterizing the class of functions that can be learned, coming up with the right learning architecture that may (provably) learn multiple functions, concepts and remember them over time as humans do, theoretical understanding of language, logic, RL, meta learning and lifelong learning.
The speakers and panelists include Turing award winners Geoffrey Hinton, Leslie Valiant, and Godel Prize winner Christos Papadimitriou (full-details).
Panel Discussion: There will also be a panel discussion on the fundamental question of “Is there a mathematical model for the Mind?”. We will explore basic questions such as “Is there a provable algorithm that captures the essential capabilities of the mind?”, “How do we remember complex phenomena?”, “How is a knowledge graph created automatically?”, “How do we learn new concepts, function and action hierarchies over time?” and “Why do human decisions seem so interpretable?”
Please help advertise on mailing-lists/blog-posts and Retweet.
At LightOn, we build photonic hardware that performs random projections and it is nice to find a source of materials on the subject in one document. Here is a report comprehensively presenting how randomized algorithms are key to the future of computing:
Randomized algorithms have propelled advances in artificial intelligence and represent a foundational research area in advancing AI for Science. Future advancements in DOE Office of Science priority areas such as climate science, astrophysics, fusion, advanced materials, combustion, and quantum computing all require randomized algorithms for surmounting challenges of complexity, robustness, and scalability. This report summarizes the outcomes of that workshop, "Randomized Algorithms for Scientific Computing (RASC)," held virtually across four days in December 2020 and January 2021.
Progress usually comes from a steady technology bootstrap…until it doesn’t.
Take for instance the race for the $1,000 genome that started in the early 2000s. Initially, sequencing the human genome meant a race between the well-funded public and private sectors but more importantly, the resources for the first breakthrough ended up costing upwards of $450M. Yet despite all the economic promise of genome sequencing, had Moore’s law been applied, sequencing one full genome would still cost $100,000 today. However, once the goal became clearer to everyone, a diversity of technologies and challengers emerged. This intense competition eventually yielded a growth faster than Moore’s Law. The main takeaway is that one cannot rely on the steady progress of one specific technology alone to commoditize tools.
What does this have to do with the current state of silicon computing and the new demand for Large Language Models (LLMs)? Everything if you ask us and here is how.
Interestingly, much like the mass industrialization in the 1930s, the good folks at OpenAI are sketching new scaling laws for the industrialization of these larger models.
The sad truth is that extrapolating their findings to the training of a 10 Trillion parameters model involves a supercomputer runningcontinuously fortwo decades. The minimum capital expenditure of this adventure is estimated in the realm of several hundreds of million dollars.
Much like what happened in sequencing, while silicon improvement and architecture may achieve speedups in the following years, it is fair to say that, even with Moore’s law, no foreseeable technology can reasonably train a fully scaled-up GPT-4 and grab the economic value associated with it.
Rebooting silicon with a different physics, light, and NvNs
For a real breakthrough to occur, much like what happened in the sequencing story, different technologies need to be jointly optimized. In our case, this means performing co-design with new hardware and physics but also going rogue on full programmability.
LightOn’s photonic hardware can produce massively parallel matrix-vector multiplications with an equivalent of 2 trillion parameters “for free”: this is about one-fifth of the number of parameters needed for GPT-4. Next comes revisiting the programmability. Current LightOn’s technology keeps these weights fixed by design. Co-design means finding the algorithms for which CPUs and GPUs can perform some of the most intelligent computations and how LightOn’s massive Non-von Neumann (NvN) hardware can do the heavy lifting. We already published how we are replacing backpropagation, the workhorse of Deep Learning, with an algorithm that unleashes the full potential of our hardware in distributed training. We are also working similarly on an inference step that will take full advantage of the massive number of parameters at our disposal. This involved effort relies in a heavy part thanks to our access to ½ million GPU hours on some of France’s and Europe’s largest supercomputers.
And this is just the beginning. There is a vast untapped potential for repurposing large swaths of optical technologies directed primarily for entertainment and telecommunication into computing.
The road towards a $1,000 GPT-3
Based on the GPT-3 training cost estimates, achieving a $1,000 GPT-3 requires four orders of magnitude improvements. Much like what occurred in 2007 with the genome sequencing revolution, Moore’s law may take care of the first two orders of magnitude in the coming decade but the next two rely on an outburst of new efficient technologies — hardware and algorithms. It just so happens that GPT-3 has close to 100 layers, so achieving two orders of magnitude savings may arise faster than you can imagine. Stay tuned!
I gave a talk at #mathia2021 conference on March 9th, 2021 where I drew a parallel between the scaling laws that enabled industrialization in the 1920's and the new scaling laws in AI of the 2020's. AI is at its infancy and it needs to have guiding principles (as embedded in these empirical laws) and it also needs to develop new hardware. I showed how, in this context, LightOn can help unlock Transformative AI. Enjoy!
In the coming days, we'll be making another announcement but I wanted to first share a video we did recently. At LightOn, we don't build photonic computing hardware because it's fancy or cool (even though, it is cool) but because computing hardware is hitting the limits. I know what some say about Moore's law not being dead but the recent focus on Transformers and their attendant scaling laws makes it obvious that in order for more people to have access to these models, we need a new computing paradigm. Indeed not everyone can afford to spend a billion dollars in training these models. As Azeem was recently pointing out in one of his newsletters, this is how bad things will become:
The amazing thing is that we can start to compare the cost of training single AI models with the cost of building the physical fabs that make chips. TSMC’s state-of-the-art 3nm fabwill run to around $20bn when it is completed in two years. A fab like this may be competitive for 5-7 years, which means it’ll need to churn out $7-8m worth of chips every day before it pays back.
Implicit Neural Representations (sometimes also referred to coordinate-based representations) are a novel way to parameterize signals of all kinds. Conventional signal representations are usually discrete - for instance, images are discrete grids of pixels, audio signals are discrete samples of amplitudes, and 3D shapes are usually parameterized as grids of voxels, point clouds, or meshes. In contrast, Implicit Neural Representations parameterize a signal as a continuous function that maps the domain of the signal (i.e., a coordinate, such as a pixel coordinate for an image) to whatever is at that coordinate (for an image, an R,G,B color). Of course, these functions are usually not analytically tractable - it is impossible to "write down" the function that parameterizes a natural image as a mathematical formula. Implicit Neural Representations thus approximate that function via a neural network.
Why are they interesting?
Implicit Neural Representations have several benefits: First, they are not coupled to spatial resolution anymore, the way, for instance, an image is coupled to the number of pixels. This is because they are continuous functions! Thus, the memory required to parameterize the signal is independent of spatial resolution, and only scales with the complexity of the underyling signal. Another corollary of this is that implicit representations have "infinite resolution" - they can be sampled at arbitrary spatial resolutions.
This is immediately useful for a number of applications, such as super-resolution, or in parameterizing signals in 3D and higher dimensions, where memory requirements grow intractably fast with spatial resolution.
However, in the future, the key promise of implicit neural representations lie in algorithms that directly operate in the space of these representations. In other words: What's the "convolutional neural network" equivalent of a neural network operating on images represented by implicit representations? Questions like these offer a path towards a class of algorithms that are independent of spatial resolution!..........