Dealing with Novelty in Open Worlds: DNOW

Jan 4th, held in conjunction with WACV 2022



Overview

Computer vision algorithms are often developed inside a closed-world paradigm, for example, recognizing objects from a fixed set of categories. However, the "world" is open and constantly and dynamically changes. As a result, when the “world” changes, these computer vision algorithms are unable to detect the change and then continue to perform their tasks with incorrect and sometimes misleading predictions. In this workshop, we aim to facilitate research directions that aim to operate well in the open world while maintaining performance in the “closed” world. Many real-world applications considered at WACV must deal with changing worlds where a variety of novelty is introduced (e.g., new classes of objects).


Examples

Let's consider the following motivational examples:

  • An autonomous vehicle may not recognize an overturned truck
  • A visual recognition system used in entertainment might not fairly recognize people of different races
  • A fashion recommendation system might not produce satisfactory recommendation results for "new-arrival" clothes.
  • An e-commerce system might not check new-products that are posted in reviews or uploaded for sale.

Impact

Addressing novelties in open worlds have broad societal impacts. For example,

  • It improves the safety and robustness of vision systems, e.g., flagging an alert when seeing an unknown object in autonomous driving
  • It helps interdisciplinary research, e.g., discovering novel species of biological organisms
  • It helps mitigate bias and promotes fairness in AI or machine learning applications, e.g., detecting and carefully handling underrepresented subpopulations rather than blindly making predictions about them.

Outcomes

We expect that our workshop will

  • offer new insights to the audience with respect to new challenges and opportunities when studying computer vision systems in the open-world
  • give voice to the need for more attention to open-world paradigms and will continue the discussion, from previous such workshops/seminars, on the formalization of metrics and datasets in this space
  • provide a platform to exchange ideas among people that have different backgrounds and that come from different fields
  • bridge the gaps of academic research experiments and the requirements of real-world applications
  • explore mechanisms to measure competence at recognizing and dealing with novelty. This is especially important since the distribution of the test data (and presumably the evaluation date) will by definition, differ from the training data. This introduces challenges for evaluation as well as achieving competence.

Speakers: TBD




Organizers

Please contact Pulkit or Anubhav with any questions: pulkit [at] umd [dot] edu / anubhav [at] umd [dot] edu


Pulkit Kumar
University of Maryland

Anubhav
University of Maryland

Shu Kong
Carnegie Mellon University

Terrance Boult
University of Colorado, Colorado Spring

Bill Ferguson
Raytheon BBN Technologies

Abhinav Shrivastava
 
University of Maryland




Call for Papers

As this workshop aims to draw attention to more realistic setups and explorations of broad vision problems in a real open world, there is a “call for papers", with authors having the choice for short-unpublished papers or full papers which will be published as part of the WACV workshops. Both will be reviewed by an expert panel.

Submission Instructions

Authors are required to use the same author kit as provided by WACV, 2022. All submissions will be handled through CMT

Important Dates

  • Paper Submission Deadline: Oct 11, 2021 Oct 18, 2021 11:59PM PST
  • Acceptance Notification: Nov 5, 2021 Nov 9, 2021
  • Camera Ready Submission: Nov 19, 2021

Program Schedule

Time (HST, UTC-10)
Event
Title/Presenter
Links
0930 - 0945
Opening remarks
Pulkit Kumar, Anubhav, Abhinav Shrivastava UMD
Dealing With Novelty in Open Worlds
0945 - 1015
Invited Talk #1
Thomas G. Deitterich, OSU
Initial Experiences with Multi-Task Novelty Detection
1015 - 1045
Invited Talk #2
Anthony Hoogs, Kitware
Activity-based Novelty and Anomaly Detection in Videos
1045 - 1100
Accepted Paper Talk #1
Marvin Klingner, Technische Universität Braunschweig
Unsupervised BatchNorm Adaptation (UBNA): A Domain Adaptation Method for Semantic Segmentation Without Using Source Domain Representations
1100 - 1115
Accepted Paper Talk #2
Ashutosh Agarwal, IIT Delhi
Anay Majee, Intel
Attention Guided Cosine Margin to Overcome Class-Imbalance in Few-Shot Road Object Detection
1115 - 1130
Accepted Paper Talk #3
Narjes Askarian, Monash University
Inductive Biases for Low Data VQA: A Data Augmentation Approach
1130 - 1145
Accepted Paper Talk #4
Yimeng Li, George Mason University
Uncertainty Aware Proposal Segmentation for Unknown Object Detection
1145 - 1300
Lunch Break
1300 - 1330
Invited Talk #3
Carl Vondrick, Columbia University
Robust Perception with Natural Supervision
1330 - 1400
Invited Talk #4
Stella Yu, UC Berkeley
Actionable Representation Learning for Open-World Vision
1400 - 1415
Accepted Paper Talk #5
David Patrick and Michael Geyer,
Reconstructive Training for Real-World Robustness in Image Classification
1415 - 1430
Accepted Paper Talk #6
Yevhen Kuznietsov, KU Leuven
Towards Unsupervised Online Domain Adaptation for Semantic Segmentation
1430 - 1500
Coffee Break
1500 - 1515
Accepted Paper Talk #7
Badri Petro,
Auto QA : The Question Is Not Only What, but Also Where
1515 - 1530
Accepted Paper Talk #8
Badri Petro,
VQuAD: Video Question Answering Diagnostic Dataset
1415 - 1430
Invited Talk #5
Terrance Boult, Mohsen Jafarzadeh UCCS
Open World Learning of New Classes