Talks

Opening

Chair: Jordi Luque (Telefónica Research)

9:00 Opening remarks

• Daniel Marco (Government of Catalonia)
• Pierre-Louis Xech (Facebook Artificial Intelligence)
• Diego Perino (Telefónica Research)

Session 1: Computer Vision

Chairs: Petia Radeva (UB) & Maria Vanrell (CVC-UAB)

9:30 Aleix M. Martínez (OSU) - Keynote talk: How does deep learning work?

10:15 Cristian Canton (Facebook) - AI red teaming: risk and security in an AI driven world.

10:30 Patricia Vitoria (UPF) - ChromaGAN: adversarial picture colorization with semantic class distribution (WACV 2020).

10:45 Albert Pumarola (IRI-UPC) - 3DPeople: modeling the geometry of dressed humans (ICCV 2019).

Session 2: Deep Neural Networks

Chairs: Àgata Lapedriza (UOC/MIT) & Ferran Diego (Telefónica Research)

12:00 Joan Serrà (Dolby Labs) - Blow: a single-scale hyperconditioned flow for non-parallel raw-audio voice conversion (NeurIPS 2019).

12:15 Eric Arazo (DCU) - Unsupervised label noise modeling and loss correction (ICML 2019).

12:30 Víctor Garcia (UvA) - Combining generative and discriminative models for hybrid inference (NeurIPS 2019).

12:45 José A. Rodríguez (BBVA) - Modelling heterogeneous distributions with an uncountable mixture of asymmetric laplacians (NeurIPS 2019).

Session 3: Computer Vision

Chairs: Coloma Ballester (UPF) & Francesc Moreno (UPC/CSIC)

14:30 Giorgos Bouritsas (ICL) - Neural 3D morphable models: spiral convolutional networks for 3D shape representation learning and generation (ICCV 2019).

14:45 Míriam Bellver (BSC) - Budget-aware semi-supervised semantic and instance segmentation (CVPR 2019 Deepvision Workshop best paper award).

15:00 Alexander Gomez-Villa (UPF) - Convolutional neural networks can be deceived by visual illusions (CVPR 2019).

15:15 Carles Ventura (UOC) - RVOS: end-to-end recurrent network for video object segmentation (CVPR 2019).

15:30 Antonio M. López (CVC-UAB) - Exploring the limitations of behavior cloning for autonomous driving (ICCV 2019).

Session 4: Deep Neural Networks

Chairs: Carles Ventura (UOC) & Verónica Vilaplana (UPC)

16:45 Lluís Castrejón (MILA) - Improved conditional VRNNs for video prediction (ICCV 2019).

17:00 Víctor Campos (BSC) - How to initialize your network? Robust initialization for WeightNorm & ResNets (NeurIPS 2019).

17:15 Àgata Lapedriza (UOC/MIT) - Keynote talk: Emotion recognition.

Session 5: Natural Language

Chair: Cristina España i Bonet (DFKI) & Lluís Marquez (Amazon)

9:30 Marta R. Costa-jussà (UPC) - Keynote talk: Gender bias and natural language processing.

10:15 Gemma Boleda (UPF) - Short-term meaning shift: a distributional exploration (NAACL 2019).

10:30 Luis Espinosa-Anke (Cardiff University) - Seven: augmenting word embeddings with unsupervised relation vectors (COLING 2018).

10:45 Laura Aina (UPF) - Putting words in context: LSTM language models and lexical ambiguity (ACL 2019).

Session 6: Audio, Speech, and Music

Chairs: Joan Serrà (Dolby Labs) & Carlos Segura (Telefónica Research)

12:00 Miquel India (UPC) - Self multi-head attention for speaker recognition (INTERSPEECH 2019).

12:15 Alp Öktem (Col·lectivaT) - Prosodic phrase alignment for machine dubbing (INTERSPEECH 2019).

12:30 Santiago Pascual (UPC) - Learning problem-agnostic speech representations from multiple self-supervised tasks (INTERSPEECH 2019).

12:45 Olga Slizovskaia (UPF) - End-to-end sound source separation conditioned on instrument labels (ICASSP 2019).

Session 7: Multimodal

Chair: Gloria Haro (UPF)

14:30 Raul Gómez (Eurecat-UAB) - Exploring hate speech detection in multimodal publications (WACV 2020).

14:45 Adrià Recasens (MIT) - Learning words by drawing images (CVPR 2019).

15:00 Lluís Gómez (UAB) - Scene text visual question answering (ICCV 2019).

Session 8: Deep Reinforcement Learning

Chair: Gergely Neu (UPF)

15:15 Xavier Puig (MIT) - Synthesizing environment-aware activities via activity sketches (CVPR 2019).

15:30 Ignasi Clavera (Berkeley) - Learning to adapt in dynamic, real-world environments through meta-reinforcement learning (ICLR 2019).

15:45 Carlos Florensa (Berkeley) - Goal-conditioned imitation learning (NeurIPS 2019).

Chair: Xavi Giró (UPC/BSC)

16:30 Oriol Vinyals (Deepmind) - Keynote talk: Grandmaster level in StarCraft II using multi-agent reinforcement learning (Nature 2019).

Closing

Chair: Xavi Giró (UPC/BSC)

17:30 Closing remarks