Skip to search formSkip to main contentSkip to account menu

ALBERT: A Lite BERT for Self-supervised Learning of Language Representations

This work presents two parameter-reduction techniques to lower memory consumption and increase the training speed of BERT, and uses a self-supervised loss that focuses on modeling inter-sentence coherence.
View PDF on arXiv (opens in a new tab)

Joint segmentation and classification of human actions in video

This method is based on a discriminative temporal extension of the spatial bag-of-words model that has been very popular in object recognition and is performed robustly within a multi-class SVM framework whereas the inference over the segments is done efficiently with dynamic programming.
View on IEEE (opens in a new tab)

Self-Paced Learning with Diversity

This work proposes an approach called self-paced learning with diversity (SPLD) which formalizes the preference for both easy and diverse samples into a general regularization term, independent of the learning objective, and thus can be easily generalized into various learning tasks.
View Paper (opens in a new tab)

Beyond Gaussian Pyramid: Multi-skip Feature Stacking for action recognition

This work proposes a novel feature enhancing technique called Multi-skIp Feature Stacking (MIFS), which stacks features extracted using a family of differential filters parameterized with multiple time skips and encodes shift-invariance into the frequency space and proves that MIFS enhances the learnability of differential-based features exponentially.
View on IEEE (opens in a new tab)

Hidden Two-Stream Convolutional Networks for Action Recognition

This paper presents a novel CNN architecture that implicitly captures motion information between adjacent frames and directly predicts action classes without explicitly computing optical flow, and significantly outperforms the previous best real-time approaches.
View PDF on arXiv (opens in a new tab)

Deep Local Video Feature for Action Recognition

Experimental results on the HMDB51 and UCF101 datasets show that a simple maximum pooling on the sparsely sampled local features leads to significant performance improvement.
View on IEEE (opens in a new tab)

Double Fusion for Multimedia Event Detection

This paper introduces a fusion scheme named double fusion, which combines early fusion and late fusion together to incorporate their advantages and results are reported on TRECVID MED 2010 and 2011 data sets.
View via Publisher (opens in a new tab)

Extending Semi-supervised Learning Methods for Inductive Transfer Learning

A new transfer learning method is developed, namely COITL, by extending the co-training method in semi-supervised learning by pointing out that many semi- supervised learning methods can be extended for inductive transfer learning, if the step of labeling an unlabeled instance is replaced by re-weighting a diff-distribution instance.
View on IEEE (opens in a new tab)

Multimedia classification and event detection using double fusion

This paper introduces a fusion scheme named double fusion, which simply combines early fusion and late fusion together to incorporate their advantages, and reports the best reported results to date.
View on Springer (opens in a new tab)

Viral Video Style: A Closer Look at Viral Videos on YouTube

The proposed method is unique in that it is the first attempt to incorporate video metadata into the peak day prediction, and outperforms the state-of-the-art methods, with statistically significant differences.
View on ACM (opens in a new tab)
By clicking accept or continuing to use the site, you agree to the terms outlined in our Privacy Policy (opens in a new tab), Terms of Service (opens in a new tab), and Dataset License (opens in a new tab)