Lisa M. Brown

Bio: Lisa M. Brown is an academic researcher from University of Washington. The author has contributed to research in topics: Video tracking & Object (computer science). The author has an hindex of 45, co-authored 144 publications receiving 8042 citations. Previous affiliations of Lisa M. Brown include Albert Einstein College of Medicine & IBM.

Through the wall: extracellular vesicles in Gram-positive bacteria, mycobacteria and fungi

Abstract: Extracellular vesicles (EVs) are produced by all domains of life In Gram-negative bacteria, EVs are produced by the pinching off of the outer membrane; however, how EVs escape the thick cell walls of Gram-positive bacteria, mycobacteria and fungi is still unknown Nonetheless, EVs have been described in a variety of cell-walled organisms, including Staphylococcus aureus, Mycobacterium tuberculosis and Cryptococcus neoformans These EVs contain varied cargo, including nucleic acids, toxins, lipoproteins and enzymes, and have important roles in microbial physiology and pathogenesis In this Review, we describe the current status of vesiculogenesis research in thick-walled microorganisms and discuss the cargo and functions associated with EVs in these species

Moments in Time Dataset: One Million Videos for Event Understanding

Abstract: We present the Moments in Time Dataset, a large-scale human-annotated collection of one million short videos corresponding to dynamic events unfolding within three seconds. Modeling the spatial-audio-temporal dynamics even for actions occurring in 3 second videos poses many challenges: meaningful events do not include only people, but also objects, animals, and natural phenomena; visual and auditory events can be symmetrical in time (“opening” is “closing” in reverse), and either transient or sustained. We describe the annotation process of our dataset (each video is tagged with one action or activity label among 339 different classes), analyze its scale and diversity in comparison to other large-scale video datasets for action recognition, and report results of several baseline models addressing separately, and jointly, three modalities: spatial, temporal and auditory. The Moments in Time dataset, designed to have a large coverage and diversity of events in both visual and auditory modalities, can serve as a new challenge to develop models that scale to the level of complexity and abstract reasoning that a human processes on a daily basis.

Defect structure of metal‐organic chemical vapor deposition‐grown epitaxial (0001) GaN/Al2O3

Abstract: Defect structures were investigated by transmission electron microscopy for GaN/Al2O3 (0001) epilayers grown by metal‐organic chemical vapor deposition using a two‐step process. The defect structures, including threading dislocations, partial dislocation bounding stacking faults, and inversion domains, were analyzed by diffraction contrast, high‐resolution imaging, and convergent beam diffraction. GaN film growth was initiated at 600 °C with a nominal 20 nm nucleation layer. This was followed by high‐temperature growth at 1080 °C. The near‐interfacial region of the films consists of a mixture of cubic and hexagonal GaN, which is characterized by a high density of stacking faults bounded by Shockley and Frank partial dislocations. The near‐interfacial region shows a high density of inversion domains. Above ∼0.5 μm thickness, the film consists of isolated threading dislocations of either pure edge, mixed, or pure screw character with a total density of ∼7×108 cm−2. The threading dislocation reduction in the...

Domain-adversarial training of neural networks

Abstract: We introduce a new representation learning approach for domain adaptation, in which data at training and test time come from similar but different distributions. Our approach is directly inspired by the theory on domain adaptation suggesting that, for effective domain transfer to be achieved, predictions must be made based on features that cannot discriminate between the training (source) and test (target) domains. The approach implements this idea in the context of neural network architectures that are trained on labeled data from the source domain and unlabeled data from the target domain (no labeled target-domain data is necessary). As the training progresses, the approach promotes the emergence of features that are (i) discriminative for the main learning task on the source domain and (ii) indiscriminate with respect to the shift between the domains. We show that this adaptation behaviour can be achieved in almost any feed-forward model by augmenting it with few standard layers and a new gradient reversal layer. The resulting augmented architecture can be trained using standard backpropagation and stochastic gradient descent, and can thus be implemented with little effort using any of the deep learning packages. We demonstrate the success of our approach for two distinct classification problems (document sentiment analysis and image classification), where state-of-the-art domain adaptation performance on standard benchmarks is achieved. We also validate the approach for descriptor learning task in the context of person re-identification application.

Unsupervised Domain Adaptation by Backpropagation

Abstract: Top-performing deep architectures are trained on massive amounts of labeled data. In the absence of labeled data for a certain task, domain adaptation often provides an attractive option given that labeled data of similar nature but from a different domain (e.g. synthetic images) are available. Here, we propose a new approach to domain adaptation in deep architectures that can be trained on large amount of labeled data from the source domain and large amount of unlabeled data from the target domain (no labeled target-domain data is necessary). As the training progresses, the approach promotes the emergence of "deep" features that are (i) discriminative for the main learning task on the source domain and (ii) invariant with respect to the shift between the domains. We show that this adaptation behaviour can be achieved in almost any feed-forward model by augmenting it with few standard layers and a simple new gradient reversal layer. The resulting augmented architecture can be trained using standard backpropagation. Overall, the approach can be implemented with little effort using any of the deep-learning packages. The method performs very well in a series of image classification experiments, achieving adaptation effect in the presence of big domain shifts and outperforming previous state-of-the-art on Office datasets.

Unsupervised Domain Adaptation by Backpropagation

Abstract: Top-performing deep architectures are trained on massive amounts of labeled data. In the absence of labeled data for a certain task, domain adaptation often provides an attractive option given that labeled data of similar nature but from a different domain (e.g. synthetic images) are available. Here, we propose a new approach to domain adaptation in deep architectures that can be trained on large amount of labeled data from the source domain and large amount of unlabeled data from the target domain (no labeled target-domain data is necessary). As the training progresses, the approach promotes the emergence of "deep" features that are (i) discriminative for the main learning task on the source domain and (ii) invariant with respect to the shift between the domains. We show that this adaptation behaviour can be achieved in almost any feed-forward model by augmenting it with few standard layers and a simple new gradient reversal layer. The resulting augmented architecture can be trained using standard back propagation. Overall, the approach can be implemented with little effort using any of the deep-learning packages. The method performs very well in a series of image classification experiments, achieving adaptation effect in the presence of big domain shifts and outperforming previous state-of-the-art on Office datasets.