July 7th, 2022 | Des chats et des fromages, ou De l’importance des modèles de langage nativement entraînés
Technical Blog
May 30th, 2022 | Of Cats and Cheese: The importance of natively trained language models
In this blog post we show that using an English language model plus translation layers can lead to generated text that does not take into account cultural and regional preferences. In contrast, a natively trained model avoids these pitfalls and consequently produces text that is more appropriate to the target audience.
April 12th, 2022 | LightOn publicly launches Muse, an API to VLM-4 Large Language Models trained natively in five European Languages
Paris-based LightOn today announced the public release of Muse, an API that leverages state-of-the-art AI to democratize Natural Language Processing (NLP) for everyone, everywhere.
December 28th, 2021 | LightOn Photonic coprocessor integrated into European AI Supercomputer
It is with immense pride and pleasure to announce that LightOn’s OPU has been installed in one of the world’s Top500 supercomputers as part of a pilot program with GENCI and IDRIS/CNRS.
October 7th, 2021 | Bringing containers to light
At LightOn, we keep the innovation going on all aspects of hardware and software, to better serve our users. We recently moved our cloud infrastructure into containers, with great features for LightOn Cloud users. We had fun, and wanted to share some of the details that stood out during this project.
At LightOn, we keep the innovation going on all aspects of hardware and software, to better serve our users. We recently moved our cloud infrastructure into containers, with great features for LightOn Cloud users. We had fun, and wanted to share some of the details that stood out during this project.June 17th, 2021 | LightOn Qore in a Nutshell
At LightOn, we recently announced a future quantum product called Qore. This blog post explains what this new product is and how it is a radical departure from traditional approaches to Quantum Computing. You may want to read the initial press release.
At LightOn, we recently announced a future quantum product called Qore. This blog post explains what this new product is and how it is a radical departure from traditional approaches to Quantum Computing. You may want to read the June 4th, 2021 | Summary of LightOn AI Meetup #14: “WeightWatcher: a Diagnostic Tool for Deep Neural Networks” – Invited Speaker: Charles Martin
It is about the one-year mark since we started our (virtual) LightOn AI Meetups, and to mark the anniversary 🥳, we had Charles Martin as a guest speaker. Charles is the Chief Scientist at Calculation Consulting and presented his work on WeightWatcher: a Diagnostic Tool for Deep Neural Networks, a Python package built around a series of papers.
It is about the one-year mark since we started our (virtual) LightOn AI Meetups, and to mark the anniversary 🥳, we had April 23rd, 2021 | Summary of LightOn AI meetup #13: “Efficient Neural Networks Training through Locality Sensitive Hashing” – Invited Speaker: Beidi Chen
Beidi Chen, a Postdoc Researcher at Stanford, was the guest of LightOn’s 13th AI Meetup and presented her work on SLIDE&MONGOOSE: LSH Frameworks for Efficient Neural Networks Training⚡ to appear at ICLR 2021. The 📺 recording of the meetup is on LightOn’s Youtube channel. Subscribe to the channel and subscribe to our Meetup to get notified of the next videos and events!
April 6th, 2021 | The $1,000 GPT-3
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.
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.March 24th, 2021 | Computing with Light: How LightOn intends to unlock Transformative AI
Igor Carron, our CEO and co-founder, gave a talk at #mathia2021 conference on March 9th, 2021 where he drew a parallel between the scaling laws that enabled industrialization in the 1920s and the new scaling laws in AI of the 2020s. AI is at its infancy and it needs to have guiding principles (as embedded in these empirical laws) while also needs to develop new hardware. Igor showed how, in this context, LightOn can help unlock Transformative AI. Enjoy!
Igor Carron, our CEO and co-founder, gave a talk at March 20th, 2021 | Summary of LightOn AI meetup #12: “Fast Graph Classification with LightOn Optical Processing Units” – Invited Speaker: Hashem Ghanem
At the 12th LightOn AI Meetup, we hosted Hashem Ghanem, a Ph.D. student at GIPSA-Lab, Grenoble, and IMB, Dijon, who presented his work on🕸️ Fast Graph Kernel with Optical Random Features ⚡ to appear at ICASSP 2021. The 📺 recording of the meetup is on LightOn’s Youtube channel. Subscribe to the channel and subscribe to our Meetup to get notified of the next videos and events!
At the 12th LightOn AI Meetup, we hosted Hashem Ghanem, a Ph.D. student at GIPSA-Lab, Grenoble, and IMB, Dijon, who presented his work on🕸️ Fast Graph Kernel with Optical Random Features ⚡ to appear at ICASSP 2021. The March 8th, 2021 | LightOn Appliance, a photonic co-processor sets a new pathway for Transformative AI computing
Several months ago, we released LightOn Appliance, the world’s most powerful photonic co-processor for AI and HPC. We originally had our technology available on the cloud with a nascent community, however, now it can be installed on-premises, as a 2U rackable unit.
Several months ago, we released February 19th, 2021 | Summary of LightOn AI meetup #11: “Reservoir Transformers: train faster with fewer parameters, and get better results” – Invited Speaker: Douwe Kiela
The 11th edition of the LightOn AI Meetup featured Douwe Kiela, Research Scientist at Facebook AI Research, who was presenting Reservoir Transformers 🛢️ The 📺 recording of the meetup is on LightOn’s Youtube channel. Subscribe to the channel and subscribe to our Meetup to get notified of the next videos and events!
January 18th, 2021 | Summary of LightOn AI meetup #10: “Rethinking Attention with Performers: Towards New Transformers’ Revolution” – Invited Speaker: Krzysztof Choromanski
For the 10th edition of the LightOn AI Meetup we had Krzysztof Choromanski, Research Scientist at Google Robotics NYC and Adjunct Assistant Professor at Columbia University, presenting his work on Rethinking Attention with Performers, which will appear as an oral presentation at ICLR 2021. Congratulations Krzysztof and co-authors!
December 8th, 2020 | LightOn at NeurIPS
December 4th, 2020 | Summary of LightOn AI meetup #9: “The Hardware Lottery: the cost of straying off the beaten path”. Invited Speaker: Sara Hooker
November 20th, 2020 | Summary of LightOn AI meetup #8: “Deep Reservoir Computing and Beyond: Fast and Lightweight Neural Networks”. Invited Speaker: Claudio Gallicchio
November 5th, 2020 | Ignorance is Bliss: Adversarial Robustness by Design with LightOn OPUs
October 23rd, 2020 | Summary of LightOn AI meetup #7: “Not all neural network architectures are created equal, some encode better priors than others”. Invited Speaker: Adam Gaier
October 13th, 2020 | Summary of LightOn AI meetup #6: “Slash the training cost of your models with random learning signals and without feedback”. Invited Speakers: Charlotte Frenkel and Martin Lefevbre
September 10th, 2020 | Five ways Photonic Computing can change the world
August 20th, 2020 | LightOn at HotChips 2020: The Three days AI Ate the Chips
July 20th, 2020 | LightOn Highlights from ICML 2020
June 19th, 2020 | Summary of LightOn AI meetup #5: “Handling Massive Data Rates with Localized Sketching”. Invited Speaker: Rakshith Srinivasa
June 11th, 2020 | An Easy Introduction to LightOnML
May 28th, 2020 | Photonic Computing for Massively Parallel AI is out!
May 14th, 2020 | Tackling Reinforcement Learning with the Aurora OPU
May 7th, 2020 | Optical Random Features Versus SARS-CoV-2 Glycoprotein Trajectory: Round#2
April 30th, 2020 | Random Projections did it again!
April 23rd, 2020 | How to Perform Supervised Random Projections with Light
April 14th, 2020 | Recover the Double Descent Curve with an OPU
April 7th, 2020 | Welcome to LightOn Cloud 2.0 featuring LightOn Aurora OPUs
April 1st, 2020 | Crash Course: Create your Own Image Classification Model with PyTorch
March 25th, 2020 | Accelerating SARS-COv2 Molecular Dynamics Studies with Optical Random Features
March 17th, 2020 | Lifting Data … and Washing Machines: Kernel Computations from Optical Random Features
February 20th, 2020 | Au Revoir Backprop! Bonjour Optical Transfer Learning!
January 15th, 2020 | Beyond Overfitting and Beyond Silicon: The double descent curve
December 17th, 2019 | LightOn’s AI Research Workshop — FoRM #4: The Future of Random Matrices. Thursday, December 19th, 2019
December 10th, 2019 | LightOn at #NeurIPS2019
December 7th, 2019 | Supervised Random Projections made Lighter
October 28th, 2019 | LightOn’s Summer Series #3 — How I Learned to Stop Worrying and Love Random Projections
August 20th, 2019 | LightOn’s Summer Series #2 — Optical Computing: a New Hope
August 9th, 2019 | LightOn’s Summer Series #1 — Faith No Moore: Silicon Will Not Scale Indefinitely
May 23rd, 2019 | LightOn’s Third Research Workshop
April 11th, 2019 | Shedding Light on the “Grand Débat”
