Word embeddings have become the basic building blocks for several natural language processing and information retrieval tasks. Pre-trained word embeddings are used in several downstream applications as well as for constructing representations for sentences, paragraphs and documents. Recently, there has been an emphasis on further improving the pre-trained word vectors through post-processing algorithms. One such area of improvement is the dimensionality reduction of the word embeddings. Reducing the size of word embeddings through dimensionality reduction can improve their utility in memory constrained devices, benefiting several real-world applications. In this work, we present a novel algorithm that effectively combines PCA based dimensionality reduction with a recently proposed post-processing algorithm, to construct word embeddings of lower dimensions. Empirical evaluations on 12 standard word similarity benchmarks show that our algorithm reduces the embedding dimensionality by 50%, while achieving similar or (more often) better performance than the higher dimension embeddings.

In tissues, cells reside in confining microenvironments, which may mechanically restrict the ability of a cell to double in size as it prepares to divide. How confinement affects cell cycle progression remains unclear. We show that cells progressed through the cell cycle and proliferated when cultured in hydrogels exhibiting fast stress relaxation but were mostly arrested in the G0/G1 phase of the cell cycle when cultured in hydrogels that exhibit slow stress relaxation. In fast-relaxing gels, activity of stretch-activated channels (SACs), including TRPV4, promotes activation of the phosphatidylinositol 3-kinase (PI3K)/Akt pathway, which in turn drives cytoplasmic localization of the cell cycle inhibitor p27Kip1, thereby allowing S phase entry and proliferation. Cell growth during G1 activated the TRPV4-PI3K/Akt-p27Kip1 signaling axis, but growth is inhibited in the confining slow-relaxing hydrogels. Thus, in confining microenvironments, cells sense when growth is sufficient for division to proceed through a growth-responsive signaling axis mediated by SACs.

In this paper, we observe that semi-structured tabulated text is ubiquitous; understanding them requires not only comprehending the meaning of text fragments, but also implicit relationships between them. We argue that such data can prove as a testing ground for understanding how we reason about information. To study this, we introduce a new dataset called INFOTABS, comprising of human-written textual hypotheses based on premises that are tables extracted from Wikipedia info-boxes. Our analysis shows that the semi-structured, multi-domain and heterogeneous nature of the premises admits complex, multi-faceted reasoning. Experiments reveal that, while human annotators agree on the relationships between a table-hypothesis pair, several standard modeling strategies are unsuccessful at the task, suggesting that reasoning about tables can pose a difficult modeling challenge.

The glycocalyx is a coating of protein and sugar on the surface of all living cells. Dramatic perturbations to the composition and structure of the glycocalyx are frequently observed in aggressive cancers. However, tools to experimentally mimic and model the cancer-specific glycocalyx remain limited. Here, we develop a genetically encoded toolkit to engineer the chemical and physical structure of the cellular glycocalyx. By manipulating the glycocalyx structure, we are able to switch the adhesive state of cells from strongly adherent to fully detached. Surprisingly, we find that a thick and dense glycocalyx with high O-glycan content promotes cell survival even in a suspended state, characteristic of circulating tumor cells during metastatic dissemination. Our data suggest that glycocalyx-mediated survival is largely independent of receptor tyrosine kinase and mitogen activated kinase signaling. While anchorage is still required for proliferation, we find that cells with a thick glycocalyx can dynamically attach to a matrix scaffold, undergo cellular division, and quickly disassociate again into a suspended state. Together, our technology provides a needed toolkit for engineering the glycocalyx in glycobiology and cancer research.

We present a feature vector formation technique for documents -Sparse Composite Document Vector (SCDV) -which overcomes several shortcomings of the current distributional paragraph vector representations that are widely used for text representation. In SCDV, word embeddings are clustered to capture multiple semantic contexts in which words occur. They are then chained together to form document topic-vectors that can express complex, multi-topic documents. Through extensive experiments on multi-class and multi-label classification tasks, we outperform the previous state-of-the-art method, NTSG (Liu et al., 2015a). We also show that SCDV embeddings perform well on heterogeneous tasks like Topic Coherence, context-sensitive Learning and Information Retrieval. Moreover, we achieve significant reduction in training and prediction times compared to other representation methods. SCDV achieves best of both worlds -better performance with lower time and space complexity.

While neural models show remarkable accuracy on individual predictions, their internal beliefs can be inconsistent across examples. In this paper, we formalize such inconsistency as a generalization of prediction error. We propose a learning framework for constraining models using logic rules to regularize them away from inconsistency. Our framework can leverage both labeled and unlabeled examples and is directly compatible with off-the-shelf learning schemes without model redesign. We instantiate our framework on natural language inference, where experiments show that enforcing invariants stated in logic can help make the predictions of neural models both accurate and consistent.

No abstract

Reasoning about tabular information presents unique challenges to modern NLP approaches which largely rely on pre-trained contextualized embeddings of text. In this paper, we study these challenges through the problem of tabular natural language inference. We propose easy and effective modifications to how information is presented to a model for this task. We show via systematic experiments that these strategies substantially improve tabular inference performance.