Machine learning algorithms are often vulnerable to adversarial examples that have imperceptible alterations from the original counterparts but can fool the state-of-the-art models. It is helpful to evaluate or even improve the robustness of these models by exposing the maliciously crafted adversarial examples. In this paper, we present TextFooler, a simple but strong baseline to generate adversarial text. By applying it to two fundamental natural language tasks, text classification and textual entailment, we successfully attacked three target models, including the powerful pre-trained BERT, and the widely used convolutional and recurrent neural networks. We demonstrate three advantages of this framework: (1) effective—it outperforms previous attacks by success rate and perturbation rate, (2) utility-preserving—it preserves semantic content, grammaticality, and correct types classified by humans, and (3) efficient—it generates adversarial text with computational complexity linear to the text length.1
Tomographic phase microscopy (TPM) is an emerging optical microscopic technique for bioimaging. TPM uses digital holographic measurements of complex scattered fields to reconstruct three-dimensional refractive index (RI) maps of cells with diffraction-limited resolution by solving inverse scattering problems. In this paper, we review the developments of TPM from the fundamental physics to its applications in bioimaging. We first provide a comprehensive description of the tomographic reconstruction physical models used in TPM. The RI map reconstruction algorithms and various regularization methods are discussed. Selected TPM applications for cellular imaging, particularly in hematology, are reviewed. Finally, we examine the limitations of current TPM systems, propose future solutions, and envision promising directions in biomedical research.
Single cell trapping increasingly serves as a key manipulation technique in single cell analysis for many cutting-edge cell studies. Due to their inherent advantages, microfluidic devices have been widely used to enable single cell immobilization. To further improve the single cell trapping efficiency, this paper reports on a passive hydrodynamic microfluidic device based on the "least flow resistance path" principle with geometry optimized in line with corresponding cell types. Different from serpentine structure, the core trapping structure of the micro-device consists of a series of concatenated T and inverse T junction pairs which function as bypassing channels and trapping constrictions. This new device enhances the single cell trapping efficiency from three aspects: (1) there is no need to deploy very long or complicated channels to adjust flow resistance, thus saving space for each trapping unit; (2) the trapping works in a "deterministic" manner, thus saving a great deal of cell samples; and (3) the compact configuration allows shorter flowing path of cells in multiple channels, thus increasing the speed and throughput of cell trapping. The mathematical model of the design was proposed and optimization of associated key geometric parameters was conducted based on computational fluid dynamics (CFD) simulation. As a proof demonstration, two types of PDMS microfluidic devices were fabricated to trap HeLa and HEK-293T cells with relatively significant differences in cell sizes. Experimental results showed 100% cell trapping and 90% single cell trapping over 4 × 100 trap sites for these two cell types, respectively. The space saving is estimated to be 2-fold and the cell trapping speed enhancement to be 3-fold compared to previously reported devices. This device can be used for trapping various types of cells and expanded to trap cells in the order of tens of thousands on 1-cm(2) scale area, as a promising tool to pattern large-scale single cells on specific substrates and facilitate on-chip cellular assay at the single cell level.
We combine generative adversarial network (GAN) with light microscopy to achieve deep learning super-resolution under a large field of view (FOV). By appropriately adopting prior microscopy data in an adversarial training, the neural network can recover a high-resolution, accurate image of new specimen from its single low-resolution measurement. Its capacity has been broadly demonstrated via imaging various types of samples, such as USAF resolution target, human pathological slides, fluorescence-labelled fibroblast cells, and deep tissues in transgenic mouse brain, by both wide-field and light-sheet microscopes. The gigapixel, multi-color reconstruction of these samples verifies a successful GAN-based single image super-resolution procedure. We also propose an image degrading model to generate low resolution images for training, making our approach free from the complex image registration during training dataset preparation. After a welltrained network being created, this deep learning-based imaging approach is capable of recovering a large FOV (~95 mm 2 ), high-resolution (~1.7 μm) image at high speed (within 1 second), while not necessarily introducing any changes to the setup of existing microscopes.
Successful evidence-based medicine (EBM) applications rely on answering clinical questions by analyzing large medical literature databases. In order to formulate a well-defined, focused clinical question, a framework called PICO is widely used, which identifies the sentences in a given medical text that belong to the four components: Participants/Problem (P), Intervention (I), Comparison (C) and Outcome (O). In this work, we present a Long Short-Term Memory (LSTM) neural network based model to automatically detect PICO elements. By jointly classifying subsequent sentences in the given text, we achieve state-of-the-art results on PICO element classification compared to several strong baseline models. We also make our curated data public as a benchmarking dataset so that the community can benefit from it.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.