Nanophotonics, the field that merges photonics and nanotechnology, has in recent years revolutionized the field of optics by enabling the manipulation of light–matter interactions with subwavelength structures. However, despite the many advances in this field, the design, fabrication and characterization has remained widely an iterative process in which the designer guesses a structure and solves the Maxwell’s equations for it. In contrast, the inverse problem, i.e., obtaining a geometry for a desired electromagnetic response, remains a challenging and time-consuming task within the boundaries of very specific assumptions. Here, we experimentally demonstrate that a novel Deep Neural Network trained with thousands of synthetic experiments is not only able to retrieve subwavelength dimensions from solely far-field measurements but is also capable of directly addressing the inverse problem. Our approach allows the rapid design and characterization of metasurface-based optical elements as well as optimal nanostructures for targeted chemicals and biomolecules, which are critical for sensing, imaging and integrated spectroscopy applications.

Purpose To develop a deep learning reconstruction approach to improve the reconstruction speed and quality of highly undersampled variable-density single-shot fast spin-echo imaging by using a variational network (VN), and to clinically evaluate the feasibility of this approach. Materials and Methods Imaging was performed with a 3.0-T imager with a coronal variable-density single-shot fast spin-echo sequence at 3.25 times acceleration in 157 patients referred for abdominal imaging (mean age, 11 years; range, 1-34 years; 72 males [mean age, 10 years; range, 1-26 years] and 85 females [mean age, 12 years; range, 1-34 years]) between March 2016 and April 2017. A VN was trained based on the parallel imaging and compressed sensing (PICS) reconstruction of 130 patients. The remaining 27 patients were used for evaluation. Image quality was evaluated in an independent blinded fashion by three radiologists in terms of overall image quality, perceived signal-to-noise ratio, image contrast, sharpness, and residual artifacts with scores ranging from 1 (nondiagnostic) to 5 (excellent). Wilcoxon tests were performed to test the hypothesis that there was no significant difference between VN and PICS. Results VN achieved improved perceived signal-to-noise ratio (P = .01) and improved sharpness (P < .001), with no difference in image contrast (P = .24) and residual artifacts (P = .07). In terms of overall image quality, VN performed better than did PICS (P = .02). Average reconstruction time ± standard deviation was 5.60 seconds ± 1.30 per section for PICS and 0.19 second ± 0.04 per section for VN. Conclusion Compared with the conventional parallel imaging and compressed sensing reconstruction (PICS), the variational network (VN) approach accelerates the reconstruction of variable-density single-shot fast spin-echo sequences and achieves improved overall image quality with higher perceived signal-to-noise ratio and sharpness. © RSNA, 2018 Online supplemental material is available for this article.

These authors contributed equally to this workOur visual perception of our surroundings is ultimately limited by the diffraction-limit, which stipulates that optical information smaller than roughly half the illumination wavelength is not retrievable. Over the past decades, many breakthroughs have led to unprecedented imaging capabilities beyond the diffraction-limit, with applications in biology and nanotechnology. In this context, nano-photonics has revolutionized the field of optics in recent years by enabling the manipulation of light-matter interaction with subwavelength structures (1-3). However, despite the many advances in this field, its impact and penetration in our daily life has been hindered by a convoluted and iterative process, cycling through modeling, nanofabrication and nano-characterization. The fundamental reason is the fact that not only the prediction of the optical response is very time consuming and requires solving Maxwell's equations with dedicated numerical packages (4-6).But, more significantly, the inverse problem, i.e. designing a nanostructure with an on-demand optical response, is currently a prohibitive task even with the most advanced numerical tools due to the high nonlinearity of the problem (7-8). Here, we harness the power of Deep Learning, a new path in modern machine learning, and show its ability to predict the geometry of nanostructures based solely on their far-field response. This approach also addresses in a direct way the currently inaccessible inverse problem breaking the ground for on-demand design of optical response with applications such as sensing, imaging and also for Plasmon's mediated cancer thermotherapy.While computer science has been harnessed to address the diffraction limit in imaging and characterization on one hand (super-resolution techniques such as PALM and STORM techniques and more (9-12)) and to assist with the design process on the other hand (13-19) to date no computational technique is capable of addressing both aspects in an integrated manner.Here, we present an integrated deep learning (DL) approach and show how deep neural networks

Recent state-of-the-art natural language understanding models, such as BERT and XLNet, score a pair of sentences (A and B) using multiple cross-attention operations – a process in which each word in sentence A attends to all words in sentence B and vice versa. As a result, computing the similarity between a query sentence and a set of candidate sentences, requires the propagation of all query-candidate sentence-pairs throughout a stack of cross-attention layers. This exhaustive process becomes computationally prohibitive when the number of candidate sentences is large. In contrast, sentence embedding techniques learn a sentence-to-vector mapping and compute the similarity between the sentence vectors via simple elementary operations. In this paper, we introduce Distilled Sentence Embedding (DSE) – a model that is based on knowledge distillation from cross-attentive models, focusing on sentence-pair tasks. The outline of DSE is as follows: Given a cross-attentive teacher model (e.g. a fine-tuned BERT), we train a sentence embedding based student model to reconstruct the sentence-pair scores obtained by the teacher model. We empirically demonstrate the effectiveness of DSE on five GLUE sentence-pair tasks. DSE significantly outperforms several ELMO variants and other sentence embedding methods, while accelerating computation of the query-candidate sentence-pairs similarities by several orders of magnitude, with an average relative degradation of 4.6% compared to BERT. Furthermore, we show that DSE produces sentence embeddings that reach state-of-the-art performance on universal sentence representation benchmarks. Our code is made publicly available at https://github.com/microsoft/Distilled-Sentence-Embedding.

Language models that utilize extensive selfsupervised pre-training from unlabeled text, have recently shown to significantly advance the state-of-the-art performance in a variety of language understanding tasks. However, it is yet unclear if and how these recent models can be harnessed for conducting text-based recommendations. In this work, we introduce RecoBERT, a BERT-based approach for learning catalog-specialized language models for text-based item recommendations. We suggest novel training and inference procedures for scoring similarities between pairs of items, that don't require item similarity labels. Both the training and the inference techniques were designed to utilize the unlabeled structure of textual catalogs, and minimize the discrepancy between them. By incorporating four scores during inference, RecoBERT can infer text-based item-to-item similarities more accurately than other techniques. In addition, we introduce a new language understanding task for wine recommendations using similarities based on professional wine reviews. As an additional contribution, we publish annotated recommendations dataset crafted by human wine experts. Finally, we evaluate Re-coBERT and compare it to various state-of-theart NLP models on wine and fashion recommendations tasks.

We present a novel model for the problem of ranking a collection of documents according to their semantic similarity to a source (query) document. While the problem of document-todocument similarity ranking has been studied, most modern methods are limited to relatively short documents or rely on the existence of "ground-truth" similarity labels. Yet, in most common real-world cases, similarity ranking is an unsupervised problem as similarity labels are unavailable. Moreover, an ideal model should not be restricted by documents' length. Hence, we introduce SDR, a self-supervised method for document similarity that can be applied to documents of arbitrary length. Importantly, SDR can be effectively applied to extremely long documents, exceeding the 4, 096 maximal token limit of Longformer. Extensive evaluations on large documents datasets show that SDR significantly outperforms its alternatives across all metrics. To accelerate future research on unlabeled long document similarity ranking, and as an additional contribution to the community, we herein publish two humanannotated test-sets of long documents similarity evaluation. The SDR code and datasets are publicly available 1 .

Recently, there has been an increasing number of studies applying machine learning techniques for the design of nanostructures. Most of these studies train a deep neural network (DNN) to approximate the highly nonlinear function of the underlying physical mapping between spectra and nanostructures. At the end of training, the DNN allows an on-demand design of nanostructures, i.e., the model can infer nanostructure geometries for desired spectra. While these approaches have presented a new paradigm, they are limited in the complexity of the structures proposed, often bound to parametric geometries. Here we introduce spectra2pix, which is a DNN trained to generate 2D images of the target nanostructures. By predicting an image, our model architecture is not limited to a closed set of nanostructure shapes, and can be trained for the design of a much wider space of geometries. We show, for the first time, to the best of our knowledge, a successful generalization ability, by designing completely unseen shapes of geometries. We attribute the successful generalization to the ability of a pixel-wise architecture to learn local properties of the meta-material, therefore mimicking faithfully the underlying physical process. Importantly, beyond synthetical data, we show our model generalization capability on real experimental data.

When training neural models, it is common to combine multiple loss terms. The balancing of these terms requires considerable human effort and is computationally demanding. Moreover, the optimal trade-off between the loss term can change as training progresses, especially for adversarial terms. In this work, we generalize the Adam optimization algorithm to handle multiple loss terms. The guiding principle is that for every layer, the gradient magnitude of the terms should be balanced. To this end, the Multi-Term Adam (MTAdam) computes the derivative of each loss term separately, infers the first and second moments per parameter and loss term, and calculates a first moment for the magnitude per layer of the gradients arising from each loss. This magnitude is used to continuously balance the gradients across all layers, in a manner that both varies from one layer to the next and dynamically changes over time. Our results show that training with the new method leads to fast recovery from suboptimal initial loss weighting and to training outcomes that match conventional training with the prescribed hyperparameters of each method. MTAdam code can be found here 1 .