| OCTOBER 29TH. | EVENT |
|---|---|
| 08:00 | Welcome and Agenda Presentation |
| 08:15 | Spotlight Presentation of Workshop Papers. 5 minutes each: Paper 1: “Curriculum Learning for Multi-Task Classification of Visual Attributes” Nikolaos Sarafianos; Theodore Giannakopoulos; Christophoros Nikou; Ioannis Kakadiaris Paper 2: “Zero-Shot Learning posed as a Missing Data Problem” Bo Zhao; Botong Wu; Tianfu Wu; Yizhou Wang Paper 3: “Deep Modality Invariant Adversarial Network for Shared Representation Learning” Kuniaki Saito; Yusuke Mukuta; Yoshitaka Ushiku; Tatsuya Harada Paper 4: “Discrepancy-based networks for unsupervised domain adaptation: a comparative study” Gabriela Csurka; Fabien Baradel; Boris Chidlovskii; Stephane Clinchant Paper 5: “Adaptive SVM+: Learning with Privileged Information for Domain Adaptation” Nikolaos Sarafianos; Michalis Vrigkas; Ioannis Kakadiaris Poster 6: “Deep Depth Domain Adaptation: A Case Study” Novi Patricia; Fabio M. Carlucci; Barbara Caputo Paper 7: “Deep Domain Adaptation by Geodesic Distance Minimization” Yifei Wang; Wen Li; Dengxin Dai; Luc Van Gool Paper 8: “Unsupervised Cross-Domain Image Hashing with Adversarial Learning” Liu Liu*; Hairong Qi Paper 9: “Inferring Human Activities Using Robust Privileged Probabilistic Learning” Michalis Vrigkas; Evangelos Kazakos; Christophoros Nikou; Ioannis Kakadiaris |
| 09:00 | Invited Talk 1: “Adaptive Deep Learning at UC Berkeley and Nexar”, Prof. Trevor Darrell Learning of layered or “deep” representations has provided significant advances in computer vision in recent years, but has traditionally been limited to fully supervised settings with very large amounts of training data. New results in adversarial adaptive representation learning show how such methods can also excel when learning across modalities and domains. I’ll present recent long-term recurrent network models that learn visuomotor policies including end to end driving from large scale crowdsourced datasets. The models leverage recently released datasets from Nexar; I’ll review these and will also present the recent Nexar recognition challenge winning entries. |
| 09:45 | Poster Session and Coffee Break |
| 11:15 | Oral 1 – Best Paper Award winner: “Generating Visual Representations for Zero-Shot Classification” Maxime Bucher; Stephane Herbin; Frederic Jurie |
| 11:35 | Oral 2 – Honorable Mention paper: “Exploiting Convolution Filter Patterns for Transfer Learning” Mehmet Aygün; Yusuf Aytar; Hazim Ekenel |
| 11:50 | Invited Talk 2: “Unsupervised Pixel-level Domain Adaptation with Generative Adversarial Networks”, Dr. Dilip Krishnan Collecting well-annotated image datasets to train modern machine learning algorithms is prohibitively expensive for many tasks. One appealing alternative is rendering synthetic data where ground-truth annotations are generated automatically. Unfortunately, models trained purely on rendered images often fail to generalize to real images. To address this shortcoming, prior work introduced unsupervised domain adaptation algorithms that attempt to map representations between the two domains or learn to extract features that are domain-invariant. In this work, we present a new approach that learns, in an unsupervised manner, a transformation in the pixel space from one domain to the other. Our generative adversarial network (GAN)-based method adapts source-domain images to appear as if drawn from the target domain. Our approach not only produces plausible samples, but also outperforms the state-of-the-art on a number of unsupervised domain adaptation scenarios by large margins. Finally, we demonstrate that the adaptation process generalizes to object classes unseen during training. |
| 12:35 | Lunch |
| 14:00 | Invited Talk 3: “Domain Adaptation for Visual Applications”, Dr. Gabriela Csurka The aim of my talk will be to give an overview of visual domain adaptation methods. The first part will focus on the tendencies used by historical shallow methods. Then I will discuss the effect of the success of deep convolutional architectures on the field of domain adaption showing that domain adaptation can benefit from these architectures in various manner. At the end of my talk I will briefly discuss domain adaptation for applications beyond image categorization. |
| 14:45 | VisDA Challenge Introduction |
| 14:55 | VisDA Classification Challenge: Honorable Mention Talk |
| 15:05 | VisDA Classification Challenge: Runner-Up Talk |
| 15:15 | VisDA Classification Challenge: Winner Talk |
| 15:25 | Afternoon Break |
| 16:00 | VisDA Segmentation Challenge: Honorable Mention Talk |
| 16:10 | VisDA Segmentation Challenge: Runner-Up Talk |
| 16:20 | VisDA Segmentation Challenge: Winner Talk |
| 16:30 | Best paper announcement and Workshop closing |