Graphcore at NeurIPS 2020

Connect with us at NeurIPS 2020

Graphcore is proud to be a Gold Sponsor at this year’s virtual edition of NeurIPS. 

Connect with us at our virtual booth to find out more about future directions of AI in 2021 and new approaches to machine intelligence with Graphcore Research. Our Expo talks will outline new innovations in training efficiency, approximate Bayesian computation and parallel training.

Plus, don’t miss Graphcore Research's poster session on Improving Neural Network Training in Low Dimensional Random Bases. The authors of the paper will discuss novel ways to train neural networks that could allow us to scale to substantially larger models in future.

Graphcore Schedule at NeurIPS 2020: 

Expo Talk: Training Neural Networks in Low-Dimensional Random Bases

Speaker: Frithjof Gressmann, Senior AI Researcher at Graphcore & Zach Eaton-Rosen, Research Scientist at Graphcore 
Tuesday, December 8, 2020 | 11:00 AM – 11:15 AM PST (7:00 PM - 7:15 PM GMT)

Expo Talk: Accelerating Approximate Bayesian Computation for COVID-19 Modelling

Speaker: Mario Michael Krell, Senior Machine Learning Engineer at Graphcore
Tuesday, December 8, 2020 | 11:30 AM – 11:45 AM PST (7:30 PM - 7:45 PM GMT)

Poster Session: Improving Neural Network Training in Low-Dimensional Random Bases

Paper Authors and Speakers: Frithjof Gressmann, Zach Eaton-Rosen and Carlo Luschi, Graphcore Research
Wednesday, December 9, 2020 | 9:00 AM – 11:00 AM PST (5:00 PM - 7:00 PM GMT)

Expo Talk: Graphcore Research Directions in AI for 2021

Speaker: Carlo Luschi, Director of Research at Graphcore 
Wednesday, December 9, 2020 | 11:30 AM – 11:45 AM PST (7:30 PM - 7:45 PM GMT)

Interested in learning more about our second generation IPU systems? Request a meeting by completing the form below and one of our AI experts will get back to you.