Loading...
hidden

View Mobile version

Meta navigation

Startseite – Hochschule Luzern

Language selection and important links

  • Contents
  • Contact
  • Login
  • De
  • En
Search

Main navigation

School navigation

  • Engineering and Architecture
  • Business
  • Computer Science
  • Social Work
  • Art and Design
  • Music

Sub-navigation

  • Degree Programmes
  • Continuing Education
  • Research
  • International
  • Campus
  • About us
  • News

Sub-navigation

Breadcrumbs

  1. Research Research
  2. NLP Model and Embedding Toolchain NLP Model and Embedding Toolchain

NLP Model and Embedding Toolchain

The ABIZ Team is predominantly confronted with German text in industry projects. While having the hardware infrastructure to train state-of-the-art deep neural network models, we yet lack the software toolchain to efficiently train and finetune such models.

Brief information

School:

Computer Science and Information Technology

Status:

Completed

Period:

01.07.2019 - 31.12.2019

Overview

There have been quite a few groundbreaking advances in the field of Natural Language Processing over the past 2-3 years. Several new approaches and Machine Learning models with constantly increasing performance were proposed. During training, these models usually process a very large amount of unstructured text data to gain knowledge about the characteristics of a language, its words and its sentences. This initial training phase is, with modern deep learning models, computationally very expensive but has to be executed only once. For specific applications, the pre-trained base model can then be finetuned on a much smaller corpus in order to inoculate domain-specific knowledge. This approach called transfer learning is generally seen as one of the big methodical achievements that makes deep learning technology much more accessible for industry projects. Nowadays, most researchers release the source code of their models as well as pre-trained base models for the purpose of reproducibility as well as to facilitate further research. However, most of the state-of-the-art research is conducted with English text, and consequently the released pre-trained models are only useful for tasks that deal with English text. An efficient toolchain allows us to quickly test and apply state-of-the-art NLP models tailored to specific domains in our industry projects and to compute pretrained German models. Some applications require finetuning of such a model to domain-specific text such as job posts, CVs or reports from social welfare. Additionally, once we have computed high quality German models, we plan to release them to the public to facilitate more research.

hidden

Facts

Type of project

Forschung

Internal organisations involved
  • Algorithmic Business F&E
Funding
  • Forschungsfinanzierung allgemein
hidden

Persons involved: internal

Project manager
  • Tobias Mérinat
Member of project team
  • Andrin Bürli
  • Ramón Christen
  • Donnacha Daly
  • Alexander Denzler
  • Sita Mazumder
  • Luca Mazzola
  • Marc Pouly
  • Joel Salzmann
  • Florian Stalder
  • Pascal Wullschleger

Brief information

School:

Computer Science and Information Technology

Status:

Completed

Period:

07/01/2019 - 12/31/2019

Footer(s)

FH Zentralschweiz

Social media links

  •  Facebook
  •  Instagram
  •  Twitter
  •  LinkedIn
  •  YouTube
  •  Flickr

Contact

Logo Lucerne University of Applied Sciences and Arts

Lucerne University of Applied Sciences and Arts


Werftestrasse 4
6002 Luzern

+41 41 228 42 42

info@hslu.ch

Direct entry

  • A bachelor's degree –
  • A master's degree –
  • Prospective Students (Continuing & Executive Programmes)
  • Unternehmen & Institutionen
  • Media Relations
  • For Students
  • For Members of Staff

Quick link

  • People Finder
  • University Buildings & Campus Locations
  • News
  • Libraries
  • Events
  • Jobs & Karriere
  • Home
  • Hiring Rooms

Static links

  • Newsletter
  • Data protection notice
  • Publishing Acknowledgements
Logo Swissuniversities

QrCode

QrCode
We use cookies on this site to give you the best browsing experience. By continuing to navigate this site or closing this banner you accept this use of cookies. For more information please visit our privacy policy.
OK