Talks

Note: a global program view can be found here. The program is still subject to minor modifications.

Morning (9:00-14:00)

Location: Auditori - Edifici Vèrtex - Campus Nord UPC - Plaça Eusebi Güell 6, 08034 Barcelona [GMaps]

Keynote:

Chair: Cristian Canton (Meta)

9:00 - Laura Leal (NVIDIA)


Session 1:

Chair: Petia radeva (UB-CVC)

10:30 - Victor Garcia (Microsoft) - Equivariant Diffusion for Molecule Generation in 3D (ICML 2022)

10:45 - Eleonora Gualdoni (UPF) - What’s in a name? A large-scale computational study on how competition between names affects naming variation . (EMNLP 2022)

11:00 - Pedro Ramoneda (UPF - MTG) - Score Difficulty Analysis for Piano Performance Education based on Fingering (ICASSP 2022)

11:15 - Javier Romero (Meta) - AutoAvatar: Autoregressive Neural Fields for Dynamic Avatar Modeling (ECCV)

11:30 - Albert Mosella-Montoro (UPC - Epic Games) - SkinningNet: Two-Stream Graph Convolutional Neural Network for Skinning Prediction of Synthetic Characters (CVPR 2022)

11:45 - Àgata Lapedriza (UOC) - Interpreting Face Inference Models using Hierarchical Network Dissection (IJCV)


Session 2:

Chair: Coloma Ballester (UPF)

12:30 - Xavier Suau (Apple) - Self-Conditioning Pre-Trained Language Models (ICML 2022)

12:45 - Ariadna Sànchez (Amazon) - Unify and conquer: How phonetic feature representation affects Polyglot Text-to-Speech (TTS) (Interspeech 2022)

13:00 - Pau Rodríguez (Apple) - Multi-label Iterated Learning for Image Classification with Label Ambiguity (CVPR 2022)

13:15 - Adrià Recasens (DeepMind) - Multiagent off‑screen behavior prediction in football (Scientific Reports)

13:30 - Miriam Bellver (Amazon) - Human Body Measurement Estimation with Adversarial Augmentation (3DV 2022)

13:45 - Xavier Giró-i-Nieto (UPC - Amazon) - Towards Sign Language Translation & Production (CVPR 2021, CVPR 2022 & EMNLP WMT-SLT 2022)

Afternoon (15:30-18:00)

Session 3 (15:30 - 16:45)

Session 3.1: Computer Vision

Chair: Carles Ventura (UOC)Location: Sala d'actes

15:30 - Bhalaji Nagarajan (UB) - Learning Multi-Subset of Classes for Fine-Grained Food Recognition (ACMM 2022)

15:45 - Eric Arazo (Insight) - Embedding contrastive unsupervised features to cluster in-and out-of-distribution noise in corrupted image datasets (ECCV 2022)

16:00 - Debora Gil (UAB-CVC) - BronchoPose: an analysis of data and model configuration for vision-based bronchoscopy pose estimation (Computer Methods and Programs in Biomedicine)

16:15 - Alexandra Gomez-Villa (UAB-CVC) - Continually Learning Self-Supervised Representations with Projected Functional Regularization (CVPR 2022)

16:30 - Guo Wen (INRIA) - Multi-Person Extreme Motion Prediction (CVPR 2022)

Session 3.2: Audio & Speech

Chair: Toni Bonafonte (Amazon)Location: Aula VS218

15:30 - Enric Gusó (Eurecat) - On Loss Functions and Evaluation Metrics for Music Source Separation (ICASSP 2022)

15:45 - Ramon Sanabria (The University of Edinburgh) - Analyzing Acoustic Word Embeddings from Pre-trained Self-supervised Models (EMNLP 2021)

16:00 - Adria Mallol-Ragolta (Ausburg) - Multi-Type Outer Product-Based Fusion of Respiratory Sounds for Detecting COVID-19 (Interspeech 2022)

16:15 - Ioannis Tsiamas (UPC) - SHAS: Approaching optimal Segmentation for End-to-End Speech Translation (Interspeech 2022)

16:30 - Guillermo Cámbara (Amazon) - Recycle Your Wav2Vec2 Codebook: A Speech Perceiver for Keyword Spotting (COLING 2022)

Session 3.3: Machine & Reinforcement Learning

Chair: Xavier Giró-i-Nieto (Amazon)Location: Aula VS208

15:30 - Álvaro Parafita (UB) - Estimand-Agnostic Causal Query Estimation With Deep Causal Graphs (IEEE Acceess)

15:45 - Alfonso Amayuelas (UC Santa Barbara) - Neural Method for Logical Reasoning over Knowledge Graphs (ICLR 2022)

16:00 - Axel Brando (BSC) - Deep Non-Crossing Quantiles through the Partial Derivative (AISTATS 2022)

16:15 - Francisco Roldan Sanchez (Insight) - Dexterous Robotic Manipulation using Deep Reinforcement Learning and Knowledge Transfer for Complex Sparse Reward-based Tasks (Journal of Knowledge Engineering)

16:30 - Lorenzo Steccanella (UPF) - State Representation Learning for Goal-Conditioned Reinforcement Learning (ECML-PKDD 2022)

Session 3.4: NLP

Chair: Xavi Suau (Apple)Location: Aula VS217

15:30 - Gerard Serra (Batou Research) - A programmable interface for creative exploration (NeurIPS W 2022)

15:45 - Javier Ferrando (UPC) - Measuring the Mixing of Contextual Information in the Transformer (EMNLP 2022)

16:00 - Eleonora Gualdoni (UPF) - Woman or tennis player? Visual typicality and lexical frequency affect variation in object naming (CogSci 2022)

16:15 - Aitor Gonzalez-Agirre (BSC) - MarIA: Spanish Language Models (SEPLN 2022)

16:30 - Oriol Domingo Roig (Batou Research) - OCCGEN: Selection of Real-world Multilingual Parallel Data Balanced in Gender within Occupations (NeurIPS 2022)


Session 4: (17:15 - 18:00)

Session 4.1: Multimodal

Chair: Amaia Salvador (Microsoft)Location: Sala d'actes

17:15 - Juan Montesinos (UPF) - VoViT: Low Latency Graph-based Audio-Visual Voice Separation Transformer (ECCV 2022)

17:30 - Ginger Delmas (UPC-IRI) - PoseScript: 3D Human Poses from Natural Language (ECCV 2022)

17.45 - Margarita Geleta (UC Berkeley) - Pixinwav: Residual Steganography for Hiding Pixels in Audio (ICASSP 2022)

Session 4.2: Life Sciences

Chair: Àgata Lapedriza (UOC-MIT)Location: Aula VS208

17:15 - Kaisar Kushibar (UB) - Layer Ensembles: A Single-Pass Uncertainty Estimation in Deep Learning for Segmentation (MICCAI 2022)

17:30 - Adrian Galdran (UPF-Adelaide) - Test Time Transform Prediction for Open Set Histopathological Image Recognition (MICCAI 2022)

17:45 - Jose Elias Yauri (CVC-UAB) - Recognition of the Mental Workloads of Pilots in the Cockpit Using EEG Signals (MDPI 2022)

Session 4.3: 3D

Chair: Míriam Bellver (Amazon)Location: Aula VS217

17:15 - Pol Caselles Rico (Crisalix) - SIRA: Relightable Avatars from a Single Image (WACV 2022)

17:30 - Enric Corona (UPC-IRI) - Learned Vertex Descent: A New Direction for 3D Human Model Fitting (ECCV 2022)

17:45 - Jianxiong Shen (UPC-IRI) - Conditional-Flow NeRF: Accurate 3D Modelling with Reliable Uncertainty Quantification (ECCV 2022)