Samsung (United Kingdom)

Scikit-learn is an increasingly popular machine learning li- brary. Written in Python, it is designed to be simple and efficient, accessible to non-experts, and reusable in various contexts. In this paper, we present and discuss our design choices for the application programming interface (API) of the project. In particular, we describe the simple and elegant interface shared by all learning and processing units in the library and then discuss its advantages in terms of composition and reusability. The paper also comments on implementation details specific to the Python ecosystem and analyzes obstacles faced by users and developers of the library.

Ivana Balazevic, Carl Allen, Timothy Hospedales. Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP). 2019.

Domain generalization (DG) is the challenging and topical problem of learning models that generalize to novel testing domains with different statistics than a set of known training domains. The simple approach of aggregating data from all source domains and training a single deep neural network end-to-end on all the data provides a surprisingly strong baseline that surpasses many prior published methods. In this paper we build on this strong baseline by designing an episodic training procedure that trains a single deep network in a way that exposes it to the domain shift that characterises a novel domain at runtime. Specifically, we decompose a deep network into feature extractor and classifier components, and then train each component by simulating it interacting with a partner who is badly tuned for the current domain. This makes both components more robust, ultimately leading to our networks producing state-of-the-art performance on three DG benchmarks. Furthermore, we consider the pervasive workflow of using an ImageNet trained CNN as a fixed feature extractor for downstream recognition tasks. Using the Visual Decathlon benchmark, we demonstrate that our episodic-DG training improves the performance of such a general purpose feature extractor by explicitly training a feature for robustness to novel problems. This shows that DG training can benefit standard practice in computer vision.

This overview portrays the evolution of orthogonal frequency division multiplexing (OFDM) research. The amelioration of powerful multicarrier OFDM arrangements with multiple-input multiple-output (MIMO) systems has numerous benefits, which are detailed in this treatise. We continue by highlighting the limitations of conventional detection and channel estimation techniques designed for multiuser MIMO OFDM systems in the so-called rank-deficient scenarios, where the number of users supported or the number of transmit antennas employed exceeds the number of receiver antennas. This is often encountered in practice, unless we limit the number of users granted access in the base station's or radio port's coverage area. Following a historical perspective on the associated design problems and their state-of-the-art solutions, the second half of this treatise details a range of classic multiuser detectors (MUDs) designed for MIMO-OFDM systems and characterizes their achievable performance. A further section aims for identifying novel cutting-edge genetic algorithm (GA)-aided detector solutions, which have found numerous applications in wireless communications in recent years. In an effort to stimulate the cross pollination of ideas across the machine learning, optimization, signal processing, and wireless communications research communities, we will review the broadly applicable principles of various GA-assisted optimization techniques, which were recently proposed also for employment in multiuser MIMO OFDM. In order to stimulate new research, we demonstrate that the family of GA-aided MUDs is capable of achieving a near-optimum performance at the cost of a significantly lower computational complexity than that imposed by their optimum maximum-likelihood (ML) MUD aided counterparts. The paper is concluded by outlining a range of future research options that may find their way into next-generation wireless systems.

BACKGROUND: First-generation molecular profiles for human breast cancers have enabled the identification of features that can predict therapeutic response; however, little is known about how the various data types can best be combined to yield optimal predictors. Collections of breast cancer cell lines mirror many aspects of breast cancer molecular pathobiology, and measurements of their omic and biological therapeutic responses are well-suited for development of strategies to identify the most predictive molecular feature sets. RESULTS: We used least squares-support vector machines and random forest algorithms to identify molecular features associated with responses of a collection of 70 breast cancer cell lines to 90 experimental or approved therapeutic agents. The datasets analyzed included measurements of copy number aberrations, mutations, gene and isoform expression, promoter methylation and protein expression. Transcriptional subtype contributed strongly to response predictors for 25% of compounds, and adding other molecular data types improved prediction for 65%. No single molecular dataset consistently out-performed the others, suggesting that therapeutic response is mediated at multiple levels in the genome. Response predictors were developed and applied to TCGA data, and were found to be present in subsets of those patient samples. CONCLUSIONS: These results suggest that matching patients to treatments based on transcriptional subtype will improve response rates, and inclusion of additional features from other profiling data types may provide additional benefit. Further, we suggest a systems biology strategy for guiding clinical trials so that patient cohorts most likely to respond to new therapies may be more efficiently identified.

PURPOSE: One of the main challenges of lung cancer research is identifying patients at high risk for recurrence after surgical resection. Simple, accurate, and reproducible methods of evaluating individual risks of recurrence are needed. EXPERIMENTAL DESIGN: Based on a combined analysis of time-to-recurrence data, censoring information, and microarray data from a set of 138 patients, we selected statistically significant genes thought to be predictive of disease recurrence. The number of genes was further reduced by eliminating those whose expression levels were not reproducible by real-time quantitative PCR. Within these variables, a recurrence prediction model was constructed using Cox proportional hazard regression and validated via two independent cohorts (n = 56 and n = 59). RESULTS: After performing a log-rank test of the microarray data and successively selecting genes based on real-time quantitative PCR analysis, the most significant 18 genes had P values of <0.05. After subsequent stepwise variable selection based on gene expression information and clinical variables, the recurrence prediction model consisted of six genes (CALB1, MMP7, SLC1A7, GSTA1, CCL19, and IFI44). Two pathologic variables, pStage and cellular differentiation, were developed. Validation by two independent cohorts confirmed that the proposed model is significantly accurate (P = 0.0314 and 0.0305, respectively). The predicted median recurrence-free survival times for each patient correlated well with the actual data. CONCLUSIONS: We have developed an accurate, technically simple, and reproducible method for predicting individual recurrence risks. This model would potentially be useful in developing customized strategies for managing lung cancer.

Fifth generation cellular systems will be deployed in the microwave and millimeterwave (mmWave) frequency bands (i.e., between 0.5100 GHz). Propagation characteristics at these bands have a fundamental impact on each aspect of the cellular architecture, ranging from equipment design to real-time performance in the field. While we have a reasonable understanding of the propagation characteristics at microwave (<; 6 GHz) frequencies, the same cannot be said for mmWave. This article explains key differences in the propagation characteristics between the microwave and mmWave bands, and further gives examples of how these differences impact 5G system design.

We aim to learn a domain generalizable person re-identification (ReID) model. When such a model is trained on a set of source domains (ReID datasets collected from different camera networks), it can be directly applied to any new unseen dataset for effective ReID without any model updating. Despite its practical value in real-world deployments, generalizable ReID has seldom been studied. In this work, a novel deep ReID model termed Domain-Invariant Mapping Network (DIMN) is proposed. DIMN is designed to learn a mapping between a person image and its identity classifier, i.e., it produces a classifier using a single shot. To make the model domain-invariant, we follow a meta-learning pipeline and sample a subset of source domain training tasks during each training episode. However, the model is significantly different from conventional meta-learning methods in that: (1) no model updating is required for the target domain, (2) different training tasks share a memory bank for maintaining both scalability and discrimination ability, and (3) it can be used to match an arbitrary number of identities in a target domain. Extensive experiments on a newly proposed large-scale ReID domain generalization benchmark show that our DIMN significantly outperforms alternative domain generalization or meta-learning methods.

3D Morphable Face Models are a powerful tool in computer vision. They consist of a PCA model of face shape and colour information and allow to reconstruct a 3D face from a single 2D image. 3D Morphable Face Models are used for 3D head pose estimation, face analysis, face recognition, and, more recently, facial landmark detection and tracking. However, they are not as widely used as 2D methods - the process of building and using a 3D model is much more involved. In this paper, we present the Surrey Face Model, a multi-resolution 3D Morphable Model that we make available to the public for non-commercial purposes. The model contains different mesh resolution levels and landmark point annotations as well as metadata for texture remapping. Accompanying the model is a lightweight open-source C++ library designed with simplicity and ease of integration as its foremost goals. In addition to basic functionality, it contains pose estimation and face frontalisation algorithms. With the tools presented in this paper, we aim to close two gaps. First, by offering different model resolution levels and fast fitting functionality, we enable the use of a 3D Morphable Model in time-critical applications like tracking. Second, the software library makes it easy for the community to adopt the 3D Morphable Face Model in their research, and it offers a public place for collaboration.

Fas ligand (FasL), perforin, TNF-alpha, IL-1, and NO have been considered as effector molecule(s) leading to beta cell death in autoimmune diabetes. However, the real culprit(s) in beta cell destruction have long been elusive, despite intense investigation. We and others have demonstrated that FasL is not a major effector molecule in autoimmune diabetes, and previous inability to transfer diabetes to Fas-deficient nonobese diabetic (NOD)-lpr mice was due to constitutive FasL expression on lymphocytes from these mice. Here, we identified IFN-gamma/TNF-alpha synergism as the final effector molecules in autoimmune diabetes of NOD mice. A combination of IFN-gamma and TNF-alpha, but neither cytokine alone, induced classical caspase-dependent apoptosis in insulinoma and pancreatic islet cells. IFN-gamma treatment conferred susceptibility to TNF-alpha-induced apoptosis on otherwise resistant insulinoma cells by STAT1 activation followed by IFN regulatory factor (IRF)-1 induction. IRF-1 played a central role in IFN-gamma/TNF-alpha-induced cytotoxicity because inhibition of IRF-1 induction by antisense oligonucleotides blocked IFN-gamma/TNF-alpha-induced cytotoxicity, and transfection of IRF-1 rendered insulinoma cells susceptible to TNF-alpha-induced cytotoxicity. STAT1 and IRF-1 were expressed in pancreatic islets of diabetic NOD mice and colocalized with apoptotic cells. Moreover, anti-TNF-alpha Ab inhibited the development of diabetes after adoptive transfer. Taken together, our results indicate that IFN-gamma/TNF-alpha synergism is responsible for autoimmune diabetes in vivo as well as beta cell apoptosis in vitro and suggest a novel signal transduction in IFN-gamma/TNF-alpha synergism that may have relevance in other autoimmune diseases and synergistic anti-tumor effects of the two cytokines.

Facial affect analysis aims to create new types of human–computer interactions by enabling computers to better understand a person’s emotional state in order to provide ad hoc help and interactions. Since discrete emotional classes (such as anger, happiness, sadness and so on) are not representative of the full spectrum of emotions displayed by humans on a daily basis, psychologists typically rely on dimensional measures, namely valence (how positive the emotional display is) and arousal (how calming or exciting the emotional display looks like). However, while estimating these values from a face is natural for humans, it is extremely difficult for computer-based systems and automatic estimation of valence and arousal in naturalistic conditions is an open problem. Additionally, the subjectivity of these measures makes it hard to obtain good quality data. Here we introduce a novel deep neural network architecture to analyse facial affect in naturalistic conditions with a high level of accuracy. The proposed network integrates face alignment and jointly estimates both categorical and continuous emotions in a single pass, making it suitable for real-time applications. We test our method on three challenging datasets collected in naturalistic conditions and show that our approach outperforms all previous methods. We also discuss caveats regarding the use of this tool, and ethical aspects that must be considered in its application. The annotation of the visual signs of emotions can be important for psychological studies and even human–computer interactions. Instead of only ascribing discrete emotions, Toisoul and colleagues use a single neural network that predicts emotional labels on a spectrum of valence and arousal without separate face-alignment steps.

Smart textiles consist of discrete devices fabricated from-or incorporated onto-fibres. Despite the tremendous progress in smart textiles for lighting/display applications, a large scale approach for a smart display system with integrated multifunctional devices in traditional textile platforms has yet to be demonstrated. Here we report the realisation of a fully operational 46-inch smart textile lighting/display system consisting of RGB fibrous LEDs coupled with multifunctional fibre devices that are capable of wireless power transmission, touch sensing, photodetection, environmental/biosignal monitoring, and energy storage. The smart textile display system exhibits full freedom of form factors, including flexibility, bendability, and rollability as a vivid RGB lighting/grey-level-controlled full colour display apparatus with embedded fibre devices that are configured to provide external stimuli detection. Our systematic design and integration strategies are transformational and provide the foundation for realising highly functional smart lighting/display textiles over large area for revolutionary applications on smart homes and internet of things (IoT).

In 2023, the NCCN Guidelines for Hepatobiliary Cancers were divided into 2 separate guidelines: Hepatocellular Carcinoma and Biliary Tract Cancers. The NCCN Guidelines for Biliary Tract Cancers provide recommendations for the evaluation and comprehensive care of patients with gallbladder cancer, intrahepatic cholangiocarcinoma, and extrahepatic cholangiocarcinoma. The multidisciplinary panel of experts meets at least on an annual basis to review requests from internal and external entities as well as to evaluate new data on current and emerging therapies. These Guidelines Insights focus on some of the recent updates to the NCCN Guidelines for Biliary Tract Cancers as well as the newly published section on principles of molecular testing.

Abstract Securing the chemical and physical stabilities of electrode/solid‐electrolyte interfaces is crucial for the use of solid electrolytes in all‐solid‐state batteries. Directly probing these interfaces during electrochemical reactions would significantly enrich the mechanistic understanding and inspire potential solutions for their regulation. Herein, the electrochemistry of the lithium/Li 7 La 3 Zr 2 O 12 ‐electrolyte interface is elucidated by probing lithium deposition through the electrolyte in an anode‐free solid‐state battery in real time. Lithium plating is strongly affected by the geometry of the garnet‐type Li 7 La 3 Zr 2 O 12 (LLZO) surface, where nonuniform/filamentary growth is triggered particularly at morphological defects. More importantly, lithium‐growth behavior significantly changes when the LLZO surface is modified with an artificial interlayer to produce regulated lithium depositions. It is shown that lithium‐growth kinetics critically depend on the nature of the interlayer species, leading to distinct lithium‐deposition morphologies. Subsequently, the dynamic role of the interlayer in battery operation is discussed as a buffer and seed layer for lithium redistribution and precipitation, respectively, in tailoring lithium deposition. These findings broaden the understanding of the electrochemical lithium‐plating process at the solid‐electrolyte/lithium interface, highlight the importance of exploring various interlayers as a new avenue for regulating the lithium‐metal anode, and also offer insight into the nature of lithium growth in anode‐free solid‐state batteries.

While there is clarity on the wide range of applications that are to be supported by 5G cellular communications, and standardization of 5G has now started in 3GPP, there is no conclusion yet on the detailed design of the overall 5G RAN. This article provides a comprehensive overview of the 5G RAN design guidelines, key design considerations, and functional innovations as identified and developed by key players in the field.1 It depicts the air interface landscape that is envisioned for 5G, and elaborates on how this will likely be harmonized and integrated into an overall 5G RAN, in the form of concrete control and user plane design considerations and architectural enablers for network slicing, supporting independent business-driven logical networks on a common infrastructure. The article also explains key functional design considerations for the 5G RAN, highlighting the difference to legacy systems such as LTE-A and the implications of the overall RAN design.

The therapeutic potential of stem cells in chronic obstructive pulmonary disease is not well known although stem cell therapy is effective in models of other pulmonary diseases. We tested the capacities of bone marrow cells (BMCs), mesenchymal stem cells (MSCs), and conditioned media of MSCs (MSC-CM) to repair cigarette smoke-induced emphysema. Inbred female Lewis rats were exposed to cigarette smoke for 6 mo and then received BMCs, MSCs, or MSC-CM from male Lewis rats. For 2 mo after injection, the BMC treatment gradually alleviated the cigarette smoke-induced emphysema and restored the increased mean linear intercept. The BMC treatment significantly increased cell proliferation and the number of small pulmonary vessels, reduced apoptotic cell death, attenuated the mean pulmonary arterial pressure, and inhibited muscularization in small pulmonary vessels. However, only a few male donor cells were detected from 1 day to 1 mo after BMC administration. The MSCs and cell-free MSC-CM also induced the repair of emphysema and increased the number of small pulmonary vessels. Our data show that BMC, MSCs, and MSC-CM treatment repaired cigarette smoke-induced emphysema. The repair activity of these treatments is consistent with a paracrine effect rather than stem cell engraftment because most of the donor cells disappeared and because cell-free MSC-CM also induced the repair.

We aim to learn deep person re-identification (ReID) models that are robust against noisy training data. Two types of noise are prevalent in practice: (1) label noise caused by human annotator errors and (2) data outliers caused by person detector errors or occlusion. Both types of noise pose serious problems for training ReID models, yet have been largely ignored so far. In this paper, we propose a novel deep network termed DistributionNet for robust ReID. Instead of representing each person image as a feature vector, DistributionNet models it as a Gaussian distribution with its variance representing the uncertainty of the extracted features. A carefully designed loss is formulated in DistributionNet to unevenly allocate uncertainty across training samples. Consequently, noisy samples are assigned large variance/uncertainty, which effectively alleviates their negative impacts on model fitting. Extensive experiments demonstrate that our model is more effective than alternative noise-robust deep models. The source code is available at: https://github.com/TianyuanYu/DistributionNet.

Abstract Semantic embedding of knowledge graphs has been widely studied and used for prediction and statistical analysis tasks across various domains such as Natural Language Processing and the Semantic Web. However, less attention has been paid to developing robust methods for embedding OWL (Web Ontology Language) ontologies, which contain richer semantic information than plain knowledge graphs, and have been widely adopted in domains such as bioinformatics. In this paper, we propose a random walk and word embedding based ontology embedding method named , which encodes the semantics of an OWL ontology by taking into account its graph structure, lexical information and logical constructors. Our empirical evaluation with three real world datasets suggests that benefits from these three different aspects of an ontology in class membership prediction and class subsumption prediction tasks. Furthermore, often significantly outperforms the state-of-the-art methods in our experiments.

BACKGROUND AND PURPOSE: A number of investigations support the theory that the elevated plasma homocyst(e)ine is associated with occlusive vascular disease. The aim of this study is to examine whether moderate hyperhomocyst(e)inemia is an independent risk factor for cerebral infarction. In addition, we examined the association between plasma homocyst(e)ine and the severity of cerebral atherosclerosis. METHODS: We conducted a hospital-based case-control study with 140 male controls and 78 male patients with nonfatal cerebral infarction, aged between 39 and 82 years. Plasma homocyst(e)ine levels were analyzed in 218 subjects. Fifty-five patients were evaluated for cerebral vascular stenosis by MR angiography. RESULTS: The mean plasma level of homocyst(e)ine was higher in cases than in controls (11.8+/-5.6 versus 9.6+/-4.1 micromol/L; P=0.002). The proportion of subjects with moderate hyperhomocyst(e)inemia was significantly higher in cases than in controls (16.7% versus 5.0%; P=0.004). Based on the logistic regression model, the odds ratio of the highest 5% of homocyst(e)ine levels in control group was 4.17 (95% confidence interval, 3.71 to 4. 71)(P=0.0001). After additional adjustment for total cholesterol, hypertension, smoking, diabetes, and age, the odds ratio was 1.70 (95% confidence interval, 1.48 to 1.95) (P=0.0001). The plasma homocyst(e)ine levels of patients having vessels with 3 or 2 stenosed sites were significantly higher than those of patients having vessels with 1 stenosed site or normal vessels (14.6+/-1.4, 11.0+/-1.4 versus 7.8+/-1.5, 8.9+/-1.4 micromol/L respectively; P<0. 02). Multiple logistic regression analysis revealed that moderate hyperhomocyst(e)ienemia was significantly associated with the number of stenosed vessels (P=0.001). CONCLUSIONS: These findings suggest that moderate hyperhomocyst(e)inemia is an independent risk factor for cerebral infarction and may predict the severity of cerebral atherosclerosis in patients with cerebral infarction.