Rekordmange DIKU-publikationer accepteret ved to topkonferencer

Det er et højere antal end nogensinde før for DIKU ved samme konferencer.

De to konferencer placerer sig på hhv. 2. pladsen (NeurIPS) og 8. pladsen (EMNLP) over de mest indflydelsesrige konferencer inden for datalogi bredt set*. Det er således nogle af de vigtigste fora at få sin forskning publiceret i som datalogiforsker.

- Det er et rigtig godt resultat for DIKUs forskere og for dansk forskning generelt, understreger professor Christina Lioma, der står i spidsen for DIKUs Machine Learning forskningssektion.

Forskningen er udført i samarbejde med forskere fra en række forskningsinstitutioner som Google, DeepMind og Rigshospitalet og fra universiteter i blandt andet Grækenland, Italien, USA, Kina, Skotland, Israel og Sverige.

NeurIPS 2019 publikationer

• Fully Dynamic Consistent Facility Location
  Vincent Cohen-Addad (CNRS & Sorbonne Université), Niklas Oskar D Hjuler (University of Copenhagen), Nikos Parotsidis (University of Rome Tor Vergata), David Saulpic (Ecole normale supérieure), Chris Schwiegelshohn (Sapienza, University of Rome)
  
  
• U-Time: A Fully Convolutional Network for Time Series Segmentation Applied to Sleep Staging
  Mathias Perslev (University of Copenhagen), Michael H Jensen (University of Copehagen), Sune Darkner (University of Copenhagen, Denmark), Poul Jørgen Jennum (Danish Center for Sleep Medicine, Rigshospitalet), Christian Igel (University of Copenhagen)
  
  
• Comparing Unsupervised Word Translation Methods Step by Step
  Mareike Hartmann (University of Copenhagen), Yova Kementchedjhieva (University of Copenhagen), Anders Søgaard (University of Copenhagen)
  
  
• Nonstochastic Multiarmed Bandits with Unrestricted Delays
  Tobias Thune (University of Copenhagen), Nicolò Cesa-Bianchi (University of Milan), Yevgeny Seldin (University of Copenhagen)
  
  
• Connections Between Mirror Descent, Thompson Sampling and the Information Ratio
  Julian Zimmert (University of Copenhagen), Tor Lattimore (DeepMind)

EMNLP 2019 publikationer

• MultiFC: A Real-World Multi-Domain Dataset for Evidence-Based Fact Checking of Claims
  Isabelle Augenstein (University of Copenhagen), Christina Lioma (University of Copenhagen), Dongsheng Wang (University of Copenhagen), Lucas Chaves Lima (University of Copenhagen), Casper Hansen (University of Copenhagen), Christian Hansen (University of Copenhagen) and Jakob Grue Simonsen (University of Copenhagen)
  
  
• A systematic comparison of methods for low-resource dependency parsing on genuinely low-resource languages
  Clara Vania (University of Edinburgh), Yova Kementchedjhieva (University of Copenhagen), Anders Søgaard (University of Copenhagen) and Adam Lopez (University of Edinburgh)
  
  
• Lost in Evaluation: Misleading Benchmarks for Bilingual Dictionary Induction
  Yova Kementchedjhieva (University of Copenhagen), Mareike Hartmann (University of Copenhagen) and Anders Søgaard (University of Copenhagen)
  
  
• Adversarial Removal of Demographic Attributes Revisited
  Maria Barrett (University of Copenhagen), Yova Kementchedjhieva (University of Copenhagen), Yanai Elazar (Bar-Ilan University), Desmond Elliott (University of Copenhagen) and Anders Søgaard (University of Copenhagen)
  
  
• Analysing the Representational Geometry of Neural Language Encoders
  Mostafa Abdou (University of Copenhagen), Artur Kulmizev (Uppsala University), Felix Hill (DeepMind), Daniel M. Low (MIT) and Anders Søgaard (University of Copenhagen)
  
  
• Rewarding Coreference Resolvers for Being Consistent with World Knowledge
  Rahul Aralikatte (University of Copenhagen), Heather Lent (University of Copenhagen), Ana Valeria Gonzalez (University of Copenhagen), Daniel Herschcovich (University of Copenhagen), Chen Qiu (China University of Geosciences), Anders Sandholm (Google Inc.), Michael Ringaard (Google Inc.) and Anders Søgaard (University of Copenhagen)

---

* Baseret på konference H5-index>=12 af Google Scholar Metrics.