WNNLP 2020 The 2nd IN5550 Teaching Workshop on Neural Natural - - PowerPoint PPT Presentation

▶

Sep 24, 2022 164 likes •296 views

WNNLP 2020 The 2nd IN5550 Teaching Workshop on Neural Natural Language Processing University of Oslo May 27 2020 WNNLP 2020 at a glance 15 accepted papers submitted by 21 students (17 MSc + 4 PhDs) 3 different thematic tracks:

SLIDE 1

– WNNLP 2020 – The 2nd IN5550 Teaching Workshop on Neural Natural Language Processing

University of Oslo

May 27 2020

SLIDE 2

WNNLP 2020 at a glance

◮ 15 accepted papers submitted by 21 students (17 MSc + 4 PhDs) ◮ 3 different thematic tracks: ◮ Negation Scope (Area Chair: Stephan Oepen) ◮ Sentiment Analysis (ACs: Jeremy Barnes & Lilja Øvrelid) ◮ Named Entity Recognition (AC: Erik Velldal) ◮ General PyTorch support by Vinit Ravishankar ◮ All papers peer-reviewed by two students + instructors. ◮ Workshop proceedings (132 pages): https://www.uio.no/studier/emner/matnat/ifi/IN5550/v20/ proceedings.pdf

SLIDE 3

Program

◮ Presentations from the four PhD fellows (15+5 min.) ◮ Ping-Han Hsieh: Targeted Sentiment Analysis for Norwegian Language ◮ Maja Buljan: A Replication Study in Negation Scope Resolution ◮ Awadelrahman Mohamedelsadig Ali Ahmed: Sequence to Sequence Learning for Named Entity Recognition ◮ Ole Magnus Holter: Toward multilingual Named Entity Recognition for Norwegian and English ◮ Awards:

◮ best paper ◮ outstanding reviews ◮ highest score across obligatory assignments

SLIDE 4

Awards: trophy t-shirts

SLIDE 5

Awards: trophy t-shirts

SLIDE 6

Awards

Highest scores across obligatory assignments ◮ Max. possible points across the 3 obligatory assignments: 36 (30+6)

SLIDE 7

Awards

Highest scores across obligatory assignments ◮ Max. possible points across the 3 obligatory assignments: 36 (30+6) ◮ Ole Magnus Holter: 35 points ◮ Zhenying Wu: 34 points

SLIDE 8

Awards

Highest scores across obligatory assignments ◮ Max. possible points across the 3 obligatory assignments: 36 (30+6) ◮ Ole Magnus Holter: 35 points ◮ Zhenying Wu: 34 points Outstanding reviews

SLIDE 9

Awards

Highest scores across obligatory assignments ◮ Max. possible points across the 3 obligatory assignments: 36 (30+6) ◮ Ole Magnus Holter: 35 points ◮ Zhenying Wu: 34 points Outstanding reviews ◮ Vemund Justnes ◮ Maria Singstad Paulsen

SLIDE 10

Awards

Highest scores across obligatory assignments ◮ Max. possible points across the 3 obligatory assignments: 36 (30+6) ◮ Ole Magnus Holter: 35 points ◮ Zhenying Wu: 34 points Outstanding reviews ◮ Vemund Justnes ◮ Maria Singstad Paulsen Best paper (PhD fellows not considered)

SLIDE 11

Awards

Highest scores across obligatory assignments ◮ Max. possible points across the 3 obligatory assignments: 36 (30+6) ◮ Ole Magnus Holter: 35 points ◮ Zhenying Wu: 34 points Outstanding reviews ◮ Vemund Justnes ◮ Maria Singstad Paulsen Best paper (PhD fellows not considered) ◮ Halvor Holhjem Bugge: Targeted Sentiment Analysis with Ensembles

SLIDE 12

Awards

Highest scores across obligatory assignments ◮ Max. possible points across the 3 obligatory assignments: 36 (30+6) ◮ Ole Magnus Holter: 35 points ◮ Zhenying Wu: 34 points Outstanding reviews ◮ Vemund Justnes ◮ Maria Singstad Paulsen Best paper (PhD fellows not considered) ◮ Halvor Holhjem Bugge: Targeted Sentiment Analysis with Ensembles ◮ Honourable mentions: ◮ Øystein Skauli: Model and Data Exploration with the NorNE Dataset ◮ Petter Mæhlum and Muhammad Asad Ali: Not Quite on Target: BIO Tag Variation and Lexicon Inclusion for Norwegian Targeted Sentiment Analysis