Learning image representations using synthetic data allows training neural networks without some of the concerns associated with real images, such as privacy and bias.
The recent rapid advances in machine learning technologies largely depend on the vast richness of data available today, in terms of both the quantity and the rich content contained within.
It is thus better to tailor synthetic pre-training data to a specific downstream task, for best performance.
Physical adversarial attacks apply carefully crafted adversarial perturbations onto real objects to maliciously alter the prediction of object classifiers or detectors.
Deep convolutional networks have recently achieved great success in video recognition, yet their practical realization remains a challenge due to the large amount of computational resources required to achieve robust recognition.
As the base dataset and unlabeled dataset are from different domains, projecting the target images in the class-domain of the base dataset with a fixed pretrained model might be sub-optimal.
The regional-to-local attention includes two steps: first, the regional self-attention extract global information among all regional tokens and then the local self-attention exchanges the information among one regional token and the associated local tokens via self-attention.
Specifically, given a video segment, a multi-modal policy network is used to decide what modalities should be used for processing by the recognition model, with the goal of improving both accuracy and efficiency.
1 code implementation • • Assaf Arbelle, Sivan Doveh, Amit Alfassy, Joseph Shtok, Guy Lev, Eli Schwartz, Hilde Kuehne, Hila Barak Levi, Prasanna Sattigeri, Rameswar Panda, Chun-Fu Chen, Alex Bronstein, Kate Saenko, Shimon Ullman, Raja Giryes, Rogerio Feris, Leonid Karlinsky
In this work, we focus on the task of Detector-Free WSG (DF-WSG) to solve WSG without relying on a pre-trained detector.
Ranked #1 on Phrase Grounding on Visual Genome
To this end, we propose a dual-branch transformer to combine image patches (i. e., tokens in a transformer) of different sizes to produce stronger image features.
Ranked #419 on Image Classification on ImageNet
Tremendous progress has been made in visual representation learning, notably with the recent success of self-supervised contrastive learning methods.
Second, to effectively transfer knowledge, we develop a dynamic block swapping method by randomly replacing the blocks in the lower-precision student network with the corresponding blocks in the higher-precision teacher network.
no code implementations • 20 Nov 2020 • Ulrich Finkler, Michele Merler, Rameswar Panda, Mayoore S. Jaiswal, Hui Wu, Kandan Ramakrishnan, Chun-Fu Chen, Minsik Cho, David Kung, Rogerio Feris, Bishwaranjan Bhattacharjee
Neural Architecture Search (NAS) is a powerful tool to automatically design deep neural networks for many tasks, including image classification.
In recent years, a number of approaches based on 2D or 3D convolutional neural networks (CNN) have emerged for video action recognition, achieving state-of-the-art results on several large-scale benchmark datasets.
no code implementations • 23 Jun 2020 • Rameswar Panda, Michele Merler, Mayoore Jaiswal, Hui Wu, Kandan Ramakrishnan, Ulrich Finkler, Chun-Fu Chen, Minsik Cho, David Kung, Rogerio Feris, Bishwaranjan Bhattacharjee
The typical way of conducting large scale NAS is to search for an architectural building block on a small dataset (either using a proxy set from the large dataset or a completely different small scale dataset) and then transfer the block to a larger dataset.
Current state-of-the-art models for video action recognition are mostly based on expensive 3D ConvNets.
Ranked #78 on Action Recognition on Something-Something V2 (using extra training data)
We propose a new method to create compact convolutional neural networks (CNNs) by exploiting sparse convolutions.
The proposed approach demonstrates improvement of model efficiency and performance on both object recognition and speech recognition tasks, using popular architectures including ResNet and ResNeXt.
In contrast, we argue that it is essential to prune neurons in the entire neuron network jointly based on a unified goal: minimizing the reconstruction error of important responses in the "final response layer" (FRL), which is the second-to-last layer before classification, for a pruned network to retrain its predictive power.