Location: Sala Àgora, Basement of Building B3 (entrance from Plaça Telecos), Campus Nord UPC, 08034 Barcelona.

Poster session

Deep learning researchers with previously published papers in top 2018 conferences that are not presented in the oral sessions were invited to submit a poster to the closing "torrons & cava" session, on Friday 21st at 1:05pm. These works had to match the general focus of DLBCN: scientists made in Barcelona, or research developed in the city.

  1. Laura Aina (UPF), "How to represent a word and predict it, too: improving tied architectures for language modelling". EMNLP 2018.
  2. Víctor Garcia Satorras (UVA), "Few-shot learning with graph neural networks". ICLR 2018.
  3. Andrés Mafla (CVC), "Single shot scene text retrieval". ECCV 2018.
  4. Sergio Oramas (Pandora), "Multimodal deep learning for music genre classification". ISMIR 2018.
  5. Junting Pan (Columbia), "Online action detection in untrimmed, streaming videos". ECCV 2018.
  6. Albert Pumarola (CSIC-UPC), "GANimation: anatomically-aware facial animation from a single image". ECCV 2018.
  7. Joan Serrà (Telefónica Research), "Overcoming catastrophic forgetting with hard attention to the task". ICML 2018.
  8. Ionut Teodor Sodoroc (UPF), "Comparatives, quantifiers, proportions: a multi-task model for the learning of quantities from vision". NAACL 2018.
  9. Dídac Surís (MIT CSAIL), "Jointly discovering visual objects and spoken words from raw sensory input". ECCV 2018.
  10. Yaxing Wang (CVC), "Transferring GANs: generating images from limited data". ECCV 2018.
  11. Chenshen Wu (CVC), "Memory replay GANs: learning to generate new categories without forgetting". NeurIPS 2018.
  12. Ana García del Molino (NTU), "PHD-GIFs: personalized highlight detection for automatic GIF creation". ACM-MM 2018.
  13. Ana García del Molino (NTU), "Predicting visual context for unsupervised event segmentation in continuous photo-streams". ACM-MM 2018.

Poster chairs: Joan Serrà (Telefónica Research) & Marta R. Costa-jussà (UPC-IDEAI).

Target publications: same as for the conference talks.

Kindly sponsored by Vilnyx.