# Applying PQuantML to Large Architectures: Transformers for HEP and Time-Series Data

- Institution: CERN
- Speaker: Dino Liwen Cheng
- Year: 2026
- Language: English
- Source: [CERN](https://videos.cern.ch/record/3026652)
- Licence: [CERN](https://copyright.web.cern)
- Who can watch: Teens profiles and older (sources give no ratings; placed by Invisible Network from the source's own words)

**Project Description**:

The project studies how a jet tagging model with a transformer architecture trained on synthetic top quark pair production events can be compressed with \[PQuantML\]\[1\], a library for hardware-aware compression of neural networks during training, for inference on compute-constrained devices such as FPGAs, which are used for ultra-low-latency classification tasks such as jet flavour tagging.

**Supervisor**: Roope Niemi

\[1\]: https://pquantml.readthedocs.io/en/latest/ "PQuantML"

00:00:00 Slide 1

00:00:18 Slide 2

00:00:35 Slide 3

00:01:20 Slide 4

00:01:39 Slide 5

00:02:11 Slide 6

00:02:58 Slide 7

00:03:20 Slide 8

00:04:08 Slide 9

00:05:15 Slide 10

00:05:36 Slide 11

00:06:35 Slide 12

00:06:52 Slide 13

## Lectures

1. [Applying PQuantML to Large Architectures: Transformers for HEP and Time-Series Data](https://invisiblenetwork.ai/learn/cern/3026652/c7ad9bd10a80), 8 min (CERN's player)