Incremental Learning of Object Detectors without Catastrophic Forgetting
Konstantin Shmelkov,Cordelia Schmid,Karteek Alahari +2 more
- pp 3420-3429
Reads0
Chats0
TLDR
In this paper, a loss function is proposed to balance the interplay between predictions on the new classes and a new distillation loss which minimizes the discrepancy between responses for old classes from the original and the updated networks.Abstract:
Despite their success for object detection, convolutional neural networks are ill-equipped for incremental learning, i.e., adapting the original model trained on a set of classes to additionally detect objects of new classes, in the absence of the initial training data. They suffer from “catastrophic forgetting”–an abrupt degradation of performance on the original set of classes, when the training objective is adapted to the new classes. We present a method to address this issue, and learn object detectors incrementally, when neither the original training data nor annotations for the original classes in the new training set are available. The core of our proposed solution is a loss function to balance the interplay between predictions on the new classes and a new distillation loss which minimizes the discrepancy between responses for old classes from the original and the updated networks. This incremental learning can be performed multiple times, for a new set of classes in each step, with a moderate drop in performance compared to the baseline network trained on the ensemble of data. We present object detection results on the PASCAL VOC 2007 and COCO datasets, along with a detailed empirical analysis of the approach.read more
Citations
More filters
Journal ArticleDOI
Review of deep learning: concepts, CNN architectures, challenges, applications, future directions
Laith Alzubaidi,Jinglan Zhang,Amjad J. Humaidi,Ayad Q. Al-Dujaili,Ye Duan,Omran Al-Shamma,José Santamaría,Mohammed A. Fadhel,Muthana Al-Amidie,Laith Farhan +9 more
TL;DR: In this paper, a comprehensive survey of the most important aspects of DL and including those enhancements recently added to the field is provided, and the challenges and suggested solutions to help researchers understand the existing research gaps.
Journal ArticleDOI
A continual learning survey: Defying forgetting in classification tasks.
Matthias Delange,Rahaf Aljundi,Marc Masana,Sarah Parisot,Xu Jia,Ales Leonardis,Greg Slabaugh,Tinne Tuytelaars +7 more
TL;DR: This work focuses on task incremental classification, where tasks arrive sequentially and are delineated by clear boundaries and study the influence of model capacity, weight decay and dropout regularization, and the order in which the tasks are presented, and qualitatively compare methods in terms of required memory, computation time and storage.
Proceedings ArticleDOI
PackNet: Adding Multiple Tasks to a Single Network by Iterative Pruning
Arun Mallya,Svetlana Lazebnik +1 more
TL;DR: In this article, a method for adding multiple tasks to a single deep neural network while avoiding catastrophic forgetting is presented, which exploits redundancies in large deep networks to free up parameters that can then be employed to learn new tasks.
Proceedings ArticleDOI
Large Scale Incremental Learning
TL;DR: This work found that the last fully connected layer has a strong bias towards the new classes, and this bias can be corrected by a linear model, and with two bias parameters, this method performs remarkably well on two large datasets.
Book ChapterDOI
Memory Aware Synapses: Learning What (not) to Forget
TL;DR: This paper argues that, given the limited model capacity and the unlimited new information to be learned, knowledge has to be preserved or erased selectively and proposes a novel approach for lifelong learning, coined Memory Aware Synapses (MAS), which computes the importance of the parameters of a neural network in an unsupervised and online manner.
References
More filters
Proceedings ArticleDOI
Deep Residual Learning for Image Recognition
TL;DR: In this article, the authors proposed a residual learning framework to ease the training of networks that are substantially deeper than those used previously, which won the 1st place on the ILSVRC 2015 classification task.
Proceedings Article
ImageNet Classification with Deep Convolutional Neural Networks
TL;DR: The state-of-the-art performance of CNNs was achieved by Deep Convolutional Neural Networks (DCNNs) as discussed by the authors, which consists of five convolutional layers, some of which are followed by max-pooling layers, and three fully-connected layers with a final 1000-way softmax.
Book ChapterDOI
Microsoft COCO: Common Objects in Context
Tsung-Yi Lin,Michael Maire,Serge Belongie,James Hays,Pietro Perona,Deva Ramanan,Piotr Dollár,C. Lawrence Zitnick +7 more
TL;DR: A new dataset with the goal of advancing the state-of-the-art in object recognition by placing the question of object recognition in the context of the broader question of scene understanding by gathering images of complex everyday scenes containing common objects in their natural context.
Proceedings ArticleDOI
Rich Feature Hierarchies for Accurate Object Detection and Semantic Segmentation
TL;DR: RCNN as discussed by the authors combines CNNs with bottom-up region proposals to localize and segment objects, and when labeled training data is scarce, supervised pre-training for an auxiliary task, followed by domain-specific fine-tuning, yields a significant performance boost.
Journal ArticleDOI
The Pascal Visual Object Classes (VOC) Challenge
TL;DR: The state-of-the-art in evaluated methods for both classification and detection are reviewed, whether the methods are statistically different, what they are learning from the images, and what the methods find easy or confuse.