TY - JOUR A1 - Long, Xiang A1 - de Melo, Gerard A1 - He, Dongliang A1 - Li, Fu A1 - Chi, Zhizhen A1 - Wen, Shilei A1 - Gan, Chuang T1 - Purely attention based local feature integration for video classification JF - IEEE Transactions on Pattern Analysis and Machine Intelligence N2 - Recently, substantial research effort has focused on how to apply CNNs or RNNs to better capture temporal patterns in videos, so as to improve the accuracy of video classification. In this paper, we investigate the potential of a purely attention based local feature integration. Accounting for the characteristics of such features in video classification, we first propose Basic Attention Clusters (BAC), which concatenates the output of multiple attention units applied in parallel, and introduce a shifting operation to capture more diverse signals. Experiments show that BAC can achieve excellent results on multiple datasets. However, BAC treats all feature channels as an indivisible whole, which is suboptimal for achieving a finer-grained local feature integration over the channel dimension. Additionally, it treats the entire local feature sequence as an unordered set, thus ignoring the sequential relationships. To improve over BAC, we further propose the channel pyramid attention schema by splitting features into sub-features at multiple scales for coarse-to-fine sub-feature interaction modeling, and propose the temporal pyramid attention schema by dividing the feature sequences into ordered sub-sequences of multiple lengths to account for the sequential order. Our final model pyramidxpyramid attention clusters (PPAC) combines both channel pyramid attention and temporal pyramid attention to focus on the most important sub-features, while also preserving the temporal information of the video. We demonstrate the effectiveness of PPAC on seven real-world video classification datasets. Our model achieves competitive results across all of these, showing that our proposed framework can consistently outperform the existing local feature integration methods across a range of different scenarios. KW - Feature extraction KW - Convolution KW - Computational modeling KW - Plugs KW - Three-dimensional displays KW - Task analysis KW - Two dimensional displays KW - Video classification KW - action recognition KW - attention mechanism KW - computer vision KW - Algorithms KW - Neural Networks KW - Computer Y1 - 2020 U6 - https://doi.org/10.1109/TPAMI.2020.3029554 SN - 0162-8828 SN - 1939-3539 SN - 2160-9292 VL - 44 IS - 4 SP - 2140 EP - 2154 PB - Inst. of Electr. and Electronics Engineers CY - Los Alamitos ER - TY - JOUR A1 - Pan, Yuanwei A1 - Ma, Xuehua A1 - Liu, Chuang A1 - Xing, Jie A1 - Zhou, Suqiong A1 - Parshad, Badri A1 - Schwerdtle, Tanja A1 - Li, Wenzhong A1 - Wu, Aiguo A1 - Haag, Rainer T1 - Retinoic acid-loaded dendritic polyglycerol-conjugated gold nanostars for targeted photothermal therapy in breast cancer stem cells JF - ACS nano N2 - The existence of cancer stem cells (CSCs) poses a major obstacle for the success of current cancer therapies, especially the fact that non-CSCs can spontaneously turn into CSCs, which lead to the failure of the treatment and tumor relapse. Therefore, it is very important to develop effective strategies for the eradication of the CSCs. In this work, we have developed a CSCs-specific targeted, retinoic acid (RA)-loaded gold nanostars-dendritic polyglycerol (GNSs-dPG) nanoplatform for the efficient eradication of CSCs. The nanocomposites possess good biocompatibility and exhibit effective CSCs-specific multivalent targeted capability due to hyaluronic acid (HA) decorated on the multiple attachment sites of the bioinert dendritic polyglycerol (dPG). With the help of CSCs differentiation induced by RA, the self-renewal of breast CSCs and tumor growth were suppressed by the high therapeutic efficacy of photothermal therapy (PTT) in a synergistic inhibitory manner. Moreover, the stemness gene expression and CSC-driven tumorsphere formation were significantly diminished. In addition, the in vivo tumor growth and CSCs were also effectively eliminated, which indicated superior anticancer activity, effective CSCs suppression, and prevention of relapse. Taken together, we developed a CSCs-specific targeted, RA-loaded GNSs-dPG nanoplatform for the targeted eradication of CSCs and for preventing the relapse. KW - cancer stem cells KW - dendritic polyglycerol KW - gold nanostars KW - retinoic acid KW - photothermal therapy Y1 - 2021 U6 - https://doi.org/10.1021/acsnano.1c05452 SN - 1936-0851 SN - 1936-086X VL - 15 IS - 9 SP - 15069 EP - 15084 PB - American Chemical Society CY - Washington ER - TY - JOUR A1 - Cook, Katherine V. A1 - Li, Chuang A1 - Cai, Haiyuan A1 - Krumholz, Lee R. A1 - Hambright, K. David A1 - Paerl, Hans W. A1 - Steffen, Morgan M. A1 - Wilson, Alan E. A1 - Burford, Michele A. A1 - Grossart, Hans-Peter A1 - Hamilton, David P. A1 - Jiang, Helong A1 - Sukenik, Assaf A1 - Latour, Delphine A1 - Meyer, Elisabeth I. A1 - Padisak, Judit A1 - Qin, Boqiang A1 - Zamor, Richard M. A1 - Zhu, Guangwei T1 - The global Microcystis interactome JF - Limnology and oceanography N2 - Bacteria play key roles in the function and diversity of aquatic systems, but aside from study of specific bloom systems, little is known about the diversity or biogeography of bacteria associated with harmful cyanobacterial blooms (cyanoHABs). CyanoHAB species are known to shape bacterial community composition and to rely on functions provided by the associated bacteria, leading to the hypothesized cyanoHAB interactome, a coevolved community of synergistic and interacting bacteria species, each necessary for the success of the others. Here, we surveyed the microbiome associated with Microcystis aeruginosa during blooms in 12 lakes spanning four continents as an initial test of the hypothesized Microcystis interactome. We predicted that microbiome composition and functional potential would be similar across blooms globally. Our results, as revealed by 16S rRNA sequence similarity, indicate that M. aeruginosa is cosmopolitan in lakes across a 280 degrees longitudinal and 90 degrees latitudinal gradient. The microbiome communities were represented by a wide range of operational taxonomic units and relative abundances. Highly abundant taxa were more related and shared across most sites and did not vary with geographic distance, thus, like Microcystis, revealing no evidence for dispersal limitation. High phylogenetic relatedness, both within and across lakes, indicates that microbiome bacteria with similar functional potential were associated with all blooms. While Microcystis and the microbiome bacteria shared many genes, whole-community metagenomic analysis revealed a suite of biochemical pathways that could be considered complementary. Our results demonstrate a high degree of similarity across global Microcystis blooms, thereby providing initial support for the hypothesized Microcystis interactome. Y1 - 2019 U6 - https://doi.org/10.1002/lno.11361 SN - 0024-3590 SN - 1939-5590 VL - 65 SP - S194 EP - S207 PB - Wiley CY - Hoboken ER -