18 May 2020 in IEEE Transactions on Multimedia
Vision and language understanding is one of the most fundamental and difficult tasks in Multimedia Intelligence. Simultaneously Visual Question Answering (VQA) is even more challenging since it requires complex reasoning steps to the correct answer. To achieve this, Neural Module Network (NMN) and its variants rely on parsing the natural language question into a module (layout i.e., a problem-solving program). In particular, this process follows a feedforward encoder-decoder pipeline: the encoder embeds the question into a static vector and the decoder generates the layout. However, we argue that such conventional encoder-decoder neglects the dynamic nature of question comprehension (i.e., we should attend to different words from step to step) and per-module intermediate results (i.e., we should discard module performing badly) in the reasoning steps. In this paper, we present a novel NMN, called Self-Adaptive Neural Module Transformer (SANMT), which adaptively adjusts both of the question feature encoding and the layout decoding by considering intermediate Q&A results. Specifically, we encode the intermediate results with the given question features by a novel transformer module to generate dynamic question feature embedding which evolves over reasoning steps. Besides, the transformer utilizes the intermediate results from each reasoning step to guide subsequent layout arrangement. Extensive experimental evaluations demonstrate the superiority of the proposed SANMT over NMN and its variants on four challenging benchmarks, including CLEVR, CLEVR-CoGenT, VQAv1.0, and VQAv2.0 (on average the relative improvement over NMN are 1.5, 2.3, 0.7 and 0.5 points with respect to accuracy).
28 September 2018 in IEEE Transactions on Circuits and Systems for Video Technology
In this paper, we present novel sharp attention networks by adaptively sampling feature maps from convolutional neural networks for person re-identification (re-ID) problems. Due to the introduction of sampling-based attention models, the proposed approach can adaptively generate sharper attention-aware feature masks. This greatly differs from the gating-based attention mechanism that relies on soft gating functions to select the relevant features for person re-ID. In contrast, the proposed sampling-based attention mechanism allows us to effectively trim irrelevant features by enforcing the resultant feature masks to focus on the most discriminative features. It can produce sharper attentions that is more assertive in localizing subtle features relevant to re-identifying people across cameras. For this purpose, a differentiable Gumbel-Softmax sampler is employed to approximate the Bernoulli sampling to train the sharp attention networks. Extensive experimental evaluations demonstrate the superiority of this new sharp attention model for person re-ID over other related existing, published state-of-the-art works on three challenging benchmarks, including CUHK03, Market-1501, and DukeMTMC-reID.
11 January 2016 in Computer Assisted Language Learning
This paper reports on task interaction, task success and word learning among second language (L2) learners of different levels of word recognition from speech (WRS) proficiency who used a CALL application previously shown to be effective in the development of L2 WRS. Participants (N = 65) were categorised into three levels of L2 WRS proficiency according to their baseline ability to accurately recognise the phonological form of words in connected speech. All participants undertook the same web-based intervention task which involved reconstructing spoken text into the written form in a number of stages. In terms of task interaction, lower WRS proficiency participants chose to listen to stimulus material more often and changed their efforts at reconstruction significantly more frequently than participants of higher WRS proficiency levels. In relation to task success, lower WRS proficiency learners were significantly less accurate overall and less immediate in their arrival at peak reconstruction accuracy than higher WRS proficiency participants. A comparison of gain scores also showed that participants of different L2 WRS proficiency experienced significantly different word learning outcomes after using the application, with moderate WRS proficiency participants significantly outperforming lower WRS proficiency level learners in relation to academic word learning. A synthesis of the results is used to recommend design features of CALL for WRS development which will more adequately cater to learners of different L2 WRS proficiency levels. The paper concludes with some avenues for future research and some implications for the integration of computer-mediated L2 WRS development as a component of L2 listening instruction.
26 December 2022 in Briefings in Bioinformatics
The time since deposition (TSD) of a bloodstain, i.e., the time of a bloodstain formation is an essential piece of biological evidence in crime scene investigation. The practical usage of some existing microscopic methods (e.g., spectroscopy or RNA analysis technology) is limited, as their performance strongly relies on high-end instrumentation and/or rigorous laboratory conditions. This paper presents a practically applicable deep learning-based method (i.e., BloodNet) for efficient, accurate, and costless TSD inference from a macroscopic view, i.e., by using easily accessible bloodstain photos. To this end, we established a benchmark database containing around 50,000 photos of bloodstains with varying TSDs. Capitalizing on such a large-scale database, BloodNet adopted attention mechanisms to learn from relatively high-resolution input images the localized fine-grained feature representations that were highly discriminative between different TSD periods. Also, the visual analysis of the learned deep networks based on the Smooth Grad-CAM tool demonstrated that our BloodNet can stably capture the unique local patterns of bloodstains with specific TSDs, suggesting the efficacy of the utilized attention mechanism in learning fine-grained representations for TSD inference. As a paired study for BloodNet, we further conducted a microscopic analysis using Raman spectroscopic data and a machine learning method based on Bayesian optimization. Although the experimental results show that such a new microscopic-level approach outperformed the state-of-the-art by a large margin, its inference accuracy is significantly lower than BloodNet, which further justifies the efficacy of deep learning techniques in the challenging task of bloodstain TSD inference. Our code is publically accessible via https://github.com/shenxiaochenn/BloodNet. Our datasets and pre-trained models can be freely accessed via https://figshare.com/articles/dataset/21291825.
31 May 2019 in ACM Transactions on Multimedia Computing, Communications, and Applications
In this article, we propose a novel deep Siamese architecture based on a convolutional neural network (CNN) and multi-level similarity perception for the person re-identification (re-ID) problem. According to the distinct characteristics of diverse feature maps, we effectively apply different similarity constraints to both low-level and high-level feature maps during training stage. Due to the introduction of appropriate similarity comparison mechanisms at different levels, the proposed approach can adaptively learn discriminative local and global feature representations, respectively, while the former is more sensitive in localizing part-level prominent patterns relevant to re-identifying people across cameras. Meanwhile, a novel strong activation pooling strategy is utilized on the last convolutional layer for abstract local-feature aggregation to pursue more representative feature representations. Based on this, we propose final feature embedding by simultaneously encoding original global features and discriminative local features. In addition, our framework has two other benefits: First, classification constraints can be easily incorporated into the framework, forming a unified multi-task network with similarity constraints. Second, as similarity-comparable information has been encoded in the network’s learning parameters via back-propagation, pairwise input is not necessary at test time. That means we can extract features of each gallery image and build an index in an off-line manner, which is essential for large-scale real-world applications. Experimental results on multiple challenging benchmarks demonstrate that our method achieves splendid performance compared with the current state-of-the-art approaches.