Paul Balanca

Joined

May 2023

Names

Paul Balanca (Preferred)

Emails

****@graphcore.ai (Confirmed)

,

****@gmail.com (Confirmed)

Personal Links

Career & Education History

Researcher

Graphcore (graphcore.ai)

2022 – 2024

Postdoc

Queen Mary University of London (qmul.ac.uk)

2015 – 2016

Postdoc

Technion - Israel Institute of Technology, Technion (technion.ac.il)

2013 – 2015

PhD student

Ecole Centrale Paris (ecp.fr)

2010 – 2013

Advisors, Relations & Conflicts

No relations added

Expertise

low precision

2021 – Present

deep learning

2015 – Present

Publications

Elucidating the Design Space of FP4 training
Robert Hu, Carlo Luschi, Paul Balanca
- Submitted to ICLR 2026
- Readers: Everyone
Scalify: scale propagation for efficient low-precision LLM training
Paul Balanca, Samuel Hosegood, Carlo Luschi, Andrew W Fitzgibbon
- WANT@ICML 2024 Poster
- Readers: Everyone
Training and inference of large language models using 8-bit floating point
Sergio P. Perez, Yan Zhang, James Briggs, Charlie Blake, Josh Levy-Kramer, Paul Balanca, Carlo Luschi, Stephen Barlow, Andrew W Fitzgibbon
- WANT@NeurIPS 2023 Oral
- Readers: Everyone
Training and inference of large language models using 8-bit floating point
Sergio P. Perez, Yan Zhang, James Briggs, Charlie Blake, Josh Levy-Kramer, Paul Balanca, Carlo Luschi, Stephen Barlow, Andrew W Fitzgibbon
- Submitted to ICLR 2024
- Readers: Everyone
Generating QM1B with PySCF$_{\text{IPU}}$
Alexander Mathiasen, Hatem Helal, Kerstin Klaeser, Paul Balanca, Josef Dean, Carlo Luschi, Dominique Beaini, Andrew W Fitzgibbon, Dominic Masters
- NeurIPS 2023 Datasets and Benchmarks Poster
- Readers: Everyone
Repurposing Density Functional Theory to Suit Deep Learning
Alexander Mathiasen, Hatem Helal, Paul Balanca, Kerstin Klaeser, Josef Dean, Carlo Luschi, Dominique Beaini, Andrew W Fitzgibbon, Dominic Masters
- SynS & ML @ ICML2023
- Readers: Everyone

Co-Authors