Arrow Research search

Author name cluster

Hua He

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

3 papers
2 author rows

Possible papers

3

ICRA Conference 2025 Conference Paper

MJPR: Multi-Modal Joint Predictive Representation in Deep Reinforcement Learning

  • Zehan Wang
  • Ziming He
  • Zijia Wang
  • Hua He
  • Beiya Yang
  • Haobin Shi

Multi-modal reinforcement learning (RL) has been brought into focus due to its ability to provide complementary information from different sensors, enriching observations of agents. However, the introduction of multi-modal highdimensional observations brings challenges to sample efficiency. There is a lack of research on how to efficiently obtain multi-modal latent states while encouraging them to generate complementary information. To address this, we propose a representation learning method, Multi-modal Joint Predictive Representation (MJPR), which utilizes multi-modal interactive information to predict future latent states. The joint prediction method achieves the representation training for modalities and promotes each modality to generate complementary information related to predictions of each other. In addition, we introduce multi-modal loss balancing to prompt training equilibrium and cross-modal contrastive learning (CMCL) to align the modalities for effective modal interaction. We establish the multi-modal environments in the Deepmind Control suite (DMC) and Webots and compare our method with current RL representation methods. Experimental results show that MJPR outperforms state-of-the-art methods by an average of 12. 0% on six subtasks in DMC environments. It outperforms advanced methods by 16. 7% and 55. 4% in simple tasks and complex tasks of Webots environment, respectively. Moreover, ablation experiments are established in the DMC environment to verify the importance of each module to MJPR.

Details

ICLR Conference 2020 Conference Paper

Network Deconvolution

  • Chengxi Ye
  • Matthew S. Evanusa
  • Hua He
  • Anton Mitrokhin
  • Tom Goldstein
  • James A. Yorke
  • Cornelia Fermüller
  • Yiannis Aloimonos

Convolution is a central operation in Convolutional Neural Networks (CNNs), which applies a kernel to overlapping regions shifted across the image. However, because of the strong correlations in real-world image data, convolutional kernels are in effect re-learning redundant data. In this work, we show that this redundancy has made neural network training challenging, and propose network deconvolution, a procedure which optimally removes pixel-wise and channel-wise correlations before the data is fed into each layer. Network deconvolution can be efficiently calculated at a fraction of the computational cost of a convolution layer. We also show that the deconvolution filters in the first layer of the network resemble the center-surround structure found in biological neurons in the visual regions of the brain. Filtering with such kernels results in a sparse representation, a desired property that has been missing in the training of neural networks. Learning from the sparse representation promotes faster convergence and superior results without the use of batch normalization. We apply our network deconvolution operation to 10 modern neural network models by replacing batch normalization within each. Extensive experiments show that the network deconvolution operation is able to deliver performance improvement in all cases on the CIFAR-10, CIFAR-100, MNIST, Fashion-MNIST, Cityscapes, and ImageNet datasets.

Details

EAAI Journal 2012 Journal Article

Learning to predict ice accretion on electric power lines

  • Ashkan Zarnani
  • Petr Musilek
  • Xiaoyu Shi
  • Xiaodi Ke
  • Hua He
  • Russell Greiner

Ice accretion on power transmission and distribution lines is one of the major causes of power grid outages in northern regions. While such icing events are rare, they are very costly. Thus, it would be useful to predict how much ice will accumulate. Many current ice accretion forecasting systems use precipitation-type prediction and physical ice accretion models. These systems are based on expert knowledge and experimentations. An alternative strategy is to learn the patterns of ice accretion based on observations of previous events. This paper presents two different forecasting systems that are obtained by applying the learning algorithm of Support Vector Machines to the outputs of a Numerical Weather Prediction model. The first forecasting system relies on an icing model, just as the previous algorithms do. The second system learns an effective forecasting model directly from meteorological features. We use a rich data set of eight different icing events (from 2002 to 2008) to empirically compare the performance of the various ice accretion forecasting systems. Several experiments are conducted to investigate the effectiveness of the forecasting algorithms. Results indicate that the proposed forecasting system is significantly more accurate than other state-of-the-art algorithms.

Details DOI