scholarly journals Joint Attention Neural Model for Demand Prediction in Online Marketplaces

2020 ◽  
Vol 1 ◽  
pp. 6
Author(s):  
Ashish Gupta ◽  
Rishabh Mehrotra
Keyword(s):  
Inventory Management ◽  
Large Scale ◽  
Contextual Information ◽  
Neural Model ◽  
Superior Performance ◽  
Real World Data ◽  
Demand Prediction ◽  
Online Marketplaces ◽  
Textual Content ◽  
The Impact

As an increasing number of consumers rely on online marketplaces to purchase goods from, demand prediction becomes an important problem for suppliers to inform their pricing and inventory management decisions. Business volatility and the complexity of factors influence demand, which makes it a harder quantity to predict. In this paper, we consider the case of an online classified marketplace and propose a joint multi-modal neural model for demand prediction. The proposed neural model incorporates a number of factors including product description information (title, description, images), contextual information (geography, similar products) and historic interest to predict demand. Large-scale experiments on real-world data demonstrate superior performance over established baselines. Our experiments highlight the importance of considering, quantifying and leveraging the textual content of products and image quality for enhanced demand prediction. Finally, we quantify the impact of the different factors in predicting demand.

scholarly journals PathFlow-MixMatch for Whole Slide Image Registration: An Investigation of a Segment-Based Scalable Image Registration Method

2020 ◽  
Author(s):  
Joshua J. Levy ◽  
Christopher R. Jackson ◽  
Christian C. Haudenschild ◽  
Brock C. Christensen ◽  
Louis J. Vaickus
Keyword(s):  
Image Registration ◽  
Large Scale ◽  
Digital Pathology ◽  
Superior Performance ◽  
Registration Method ◽  
Target Registration Error ◽  
Alignment Algorithms ◽  
Slide Image ◽  
Segment Matching ◽  
The Impact

AbstractImage registration involves finding the best alignment between different images of the same object. In these tasks, the object in question is viewed differently in each of the images (e.g. different rotation or light conditions, etc.). In digital pathology, image registration aligns correspondent regions of tissue from different stereotactic viewpoints (e.g. subsequent deeper sections of the same tissue). These comparisons are important for histological analysis and can facilitate previously unavailable manipulations, such as 3D tissue reconstruction and cell-level alignment of immunohistochemical (IHC) and special stains. Several benchmarks have been established for evaluating image registration techniques for histological tissue; however, little work has evaluated the impact of scaling registration techniques to Giga-Pixel Whole Slide Images (WSI), which are large enough for significant memory limitations, and contain recurrent patterns and deformations that hinder traditional alignment algorithms. Furthermore, as tissue sections often contain multiple, discrete, smaller tissue fragments, it is unnecessary to align an entire image when the bulk of the image is background whitespace and tissue fragments’ orientations are often agnostic of each other. We present a methodology for circumventing large-scale image registration issues in histopathology and accompanying software. By removing background pixels, parsing the slide into discrete tissue segments, and matching, orienting and registering smaller segment pairs, we recovered registrations with lower Target Registration Error (TRE) when compared to utilizing the unmanipulated WSI. We tested our technique by having a pathologist annotate landmarks from 13 pairs of differently stained liver biopsy slides, performing WSI and segment-based registration techniques, and comparing overall TRE. Preliminary results demonstrate superior performance of registering segment pairs versus registering WSI (difference of median TRE of 44 pixels, p<0.001). Segment matching within WSI is an effective solution for histology image registration but requires further testing and validation to ensure its viability for stain translation and 3D histology analysis.

Evaluating the Effects of Transportation—Land-Use Policies on Housing Values and Household Welfare

1980 ◽  
Vol 12 (7) ◽  
pp. 747-764 ◽  
Author(s):  
A Anas
Keyword(s):  
Simulation Model ◽  
Discrete Choice ◽  
Transportation Planning ◽  
Large Scale ◽  
Travel Demand ◽  
Choice Theory ◽  
Property Values ◽  
Computationally Efficient ◽  
Demand Prediction ◽  
The Impact

In a previous article published in this journal (Anas, 1979a), a simulation model developed by the author was used to examine the impact of transit investment on property values in an urban transportation corridor that had a completely centralized employment distribution. The present paper examines the effect of rail-transit investment in the context of various scenarios which deal with urban employment decentralization, housing distribution, transportation pricing, and income composition. From these simulations it appears that under a variety of assumptions regarding urban change the taxation of short-run differential changes in property values caused by transit investment can raise only a small portion of the cost of typical transit investments. The distinctive feature of the simulation model is that it is consistent with the discrete-choice theory of travel demand currently used in transportation planning and travel-demand prediction. But whereas the state of the art in transportation planning ignores the simultaneity of transportation changes and price changes in the housing market, the model developed here is a first attempt to deal with these effects by incorporating discrete-choice theory into a Walrasian market-equilibration procedure. In addition to being a theoretical alternative to the classical bid-rent model, still made use of by urban economists, the new approach is computationally efficient and suitable for large-scale simulation.

scholarly journals Damage detection using in-domain and cross-domain transfer learning

2021 ◽  
Author(s):  
Zaharah A. Bukhsh ◽  
Nils Jansen ◽  
Aaqib Saeed
Keyword(s):  
Damage Detection ◽  
Learning Strategies ◽  
Transfer Learning ◽  
Large Scale ◽  
Visual Inspection ◽  
Superior Performance ◽  
Cross Domain ◽  
Domain Transfer ◽  
Domain Transfer Learning ◽  
The Impact

AbstractWe investigate the capabilities of transfer learning in the area of structural health monitoring. In particular, we are interested in damage detection for concrete structures. Typical image datasets for such problems are relatively small, calling for the transfer of learned representation from a related large-scale dataset. Past efforts of damage detection using images have mainly considered cross-domain transfer learning approaches using pre-trained ImageNet models that are subsequently fine-tuned for the target task. However, there are rising concerns about the generalizability of ImageNet representations for specific target domains, such as for visual inspection and medical imaging. We, therefore, evaluate a combination of in-domain and cross-domain transfer learning strategies for damage detection in bridges. We perform comprehensive comparisons to study the impact of cross-domain and in-domain transfer, with various initialization strategies, using six publicly available visual inspection datasets. The pre-trained models are also evaluated for their ability to cope with the extremely low-data regime. We show that the combination of cross-domain and in-domain transfer persistently shows superior performance specially with tiny datasets. Likewise, we also provide visual explanations of predictive models to enable algorithmic transparency and provide insights to experts about the intrinsic decision logic of typically black-box deep models.

Effectiveness of chest pain centre accreditation on the management of acute coronary syndrome: a retrospective study using a national database

2020 ◽  
pp. bmjqs-2020-011491
Author(s):  
Pengfei Sun ◽  
Jianping Li ◽  
Weiyi Fang ◽  
Xi Su ◽  
Bo Yu ◽  
...  
Keyword(s):  
Acute Coronary Syndrome ◽  
Retrospective Study ◽  
Chest Pain ◽  
Hospital Mortality ◽  
Large Scale ◽  
National Database ◽  
Real World Data ◽  
Coronary Syndrome ◽  
Pain Centre ◽  
The Impact

BackgroundLarge-scale real-world data to evaluate the impact of chest pain centre (CPC) accreditation on acute coronary syndrome (ACS) emergency care in heavy-burden developing countries like China are rare.MethodsThis study is a retrospective study based on data from the Hospital Quality Monitoring System (HQMS) database. This study included emergency patients admitted with ACS to hospitals that uploaded clinical data continuously to the database from 2013 to 2016. Propensity score matching was used to compare hospitals with and without CPC accreditation during this period. A longitudinal self-contrast comparison design with mixed-effects models was used to compare management of ACS before and after accreditation.ResultsA total of 798 008 patients with ACS from 746 hospitals were included in the analysis. After matching admission date, hospital levels and types and adjusting for possible covariates, patients with ACS admitted to accredited CPCs had lower in-hospital mortality (OR=0.70, 95% CI 0.53 to 0.93), shorter length of stay (LOS; adjusted multiplicative effect=0.89, 95% CI 0.84 to 0.94) and more percutaneous coronary intervention (PCI) procedures (OR=3.53, 95% CI 2.20 to 5.66) than patients admitted in hospitals without applying for CPC accreditation. Furthermore, when compared with the ‘before accreditation’ group only in accredited CPCs, the in-hospital mortality and LOS decreased and the usage of PCI were increased in both ‘accreditation’ (for in-hospital mortality: OR=0.86, 95% CI 0.79 to 0.93; for LOS: 0.94, 95% CI 0.93 to 0.95; for PCI: OR=1.22, 95% CI 1.18 to 1.26) and ‘after accreditation’ groups (for in-hospital mortality: OR=0.90, 95% CI 0.84 to 0.97; for LOS: 0.89, 95% CI 0.89 to 0.90; for PCI: OR=1.36, 95% CI 1.33 to 1.39). The significant benefits of decreased in-hospital mortality, reduced LOS and increased PCI usage were also observed for patients with acute myocardial infarction.ConclusionsCPC accreditation is associated with better management and in-hospital clinical outcomes of patients with ACS. CPC establishment and accreditation should be promoted and implemented in countries with high levels of ACS.

scholarly journals Flexible Charging of Electric Vehicles: Results of a Large-Scale Smart Charging Demonstration

2021 ◽  
Vol 12 (2) ◽  
pp. 82
Author(s):  
Pieter C. Bons ◽  
Aymeric Buatois ◽  
Friso Schuring ◽  
Frank Geerts ◽  
Robert van den Hoed
Keyword(s):  
Large Scale ◽  
Peak Load ◽  
Real World Data ◽  
Night Time ◽  
Electricity Grid ◽  
Charging Station ◽  
Wide Range ◽  
Smart Charging ◽  
Charging Current ◽  
The Impact

Flexible charging can be applied to avoid peak loads on the electricity grid by curbing demand of electric vehicle chargers as well as matching charging power with availability of sustainable energy. This paper presents results of a large-scale demonstration project “Flexpower” where time-dependent charging profiles are applied to 432 public charging stations in the city of Amsterdam between November 2019 and March 2020. The charging current on Flexpower stations is reduced during household peak consumption hours (18:00–21:00), increased during the night-time, and dynamically linked to solar intensity levels during the day. The results show that the EV contribution to the grid peak load can be reduced by 1.2 kW per charging station with very limited user impact. The increased charging current during sunny conditions does not lead to a significantly higher energy transfer during the day because of lack of demand and technical limitations in the vehicles. A simulation model is presented based on empirical power measurements over a wide range of conditions combining the flexibility provided by simulations with the power of real-world data. The model was validated by comparing aggregated results to actual measurements and was used to evaluate the impact of different smart charging profiles in the Amsterdam context.

scholarly journals Differences between Acute Exacerbations of Idiopathic Pulmonary Fibrosis and Other Interstitial Lung Diseases

Diagnostics ◽  
2021 ◽  
Vol 11 (9) ◽  
pp. 1623
Author(s):  
Paola Faverio ◽  
Anna Stainer ◽  
Sara Conti ◽  
Fabiana Madotto ◽  
Federica De Giacomi ◽  
...  
Keyword(s):  
Prognostic Factors ◽  
Idiopathic Pulmonary Fibrosis ◽  
Pulmonary Fibrosis ◽  
Lung Diseases ◽  
Large Scale ◽  
Interstitial Lung Diseases ◽  
Differential Count ◽  
University Hospitals ◽  
Real World Data ◽  
The Impact

Interstitial lung diseases (ILDs) comprise a wide group of pulmonary parenchymal disorders. These patients may experience acute respiratory deteriorations of their respiratory condition, termed “acute exacerbation” (AE). The incidence of AE-ILD seems to be lower than idiopathic pulmonary fibrosis (IPF), but prognosis and prognostic factors are largely unrecognized. We retrospectively analyzed a cohort of 158 consecutive adult patients hospitalized for AE-ILD in two Italian university hospitals from 2009 to 2016. Patients included in the analysis were divided into two groups: non-IPF (62%) and IPF (38%). Among ILDs included in the non-IPF group, the most frequent diagnoses were non-specific interstitial pneumonia (NSIP) (42%) and connective tissue disease (CTD)-ILD (20%). Mortality during hospitalization was significantly different between the two groups: 19% in the non-IPF group and 43% in the IPF group. AEs of ILDs are difficult-to-predict events and are burdened by relevant mortality. Increased inflammatory markers, such as neutrophilia on the differential blood cell count (HR 1.02 (CI 1.01–1.04)), the presence of pulmonary hypertension (HR 1.85 (CI 1.17–2.92)), and the diagnosis of IPF (HR 2.31 (CI 1.55–3.46)), resulted in negative prognostic factors in our analysis. Otherwise, lymphocytosis on the differential count seemed to act as a protective prognostic factor (OR 0.938 (CI 0.884–0.995)). Further prospective, large-scale, real-world data are needed to support and confirm the impact of our findings.

scholarly journals Evaluating the Impact of Uveitis on Visual Field Progression Using Large-Scale Real-World Data

2019 ◽  
Vol 207 ◽  
pp. 144-150 ◽  
Author(s):  
Xiaoxuan Liu ◽  
Stephen R. Kelly ◽  
Giovanni Montesano ◽  
Susan R. Bryan ◽  
Robert J. Barry ◽  
...  
Keyword(s):  
Visual Field ◽  
Real World ◽  
Large Scale ◽  
Real World Data ◽  
World Data ◽  
Visual Field Progression ◽  
The Impact

Navigating COVID-19 with emotional intelligence

2020 ◽  
Vol 66 (8) ◽  
pp. 810-820
Author(s):  
Mubashir Majid Baba
Keyword(s):  
Emotional Intelligence ◽  
Large Scale ◽  
Demographic Variable ◽  
Total Sample ◽  
Demographic Variables ◽  
Faculty Members ◽  
Superior Performance ◽  
Primary Data ◽  
Secondary Sources ◽  
The Impact

Background: Coronavirus disease 2019 (COVID-19) is an infectious disease caused by a newly discovered coronavirus. COVID-19 has affected educational systems worldwide, leading to the widespread closure of schools, colleges and universities. The COVID-19 pandemic is also having a dramatic impact on societies and economies around the world. With various measures of lockdowns and social distancing in place, it becomes important to understand emotional intelligence of faculty members working in institutions of higher learning on a large scale in this pandemic. Aim: The purpose of this article is to examine the perception of faculty members toward their emotional intelligence during COVID-19 and to study the impact of demographic variables on their emotional intelligence. Method: The data collected were analyzed using descriptive and inferential statistics. The data for the study were collected through both the primary and secondary sources. Online questionnaires were used to gather the primary data. The measuring items used for the study were sourced from existing validated scales and literature. Descriptive statistics was employed to know the descriptive information across various demographic variables on a total sample of 683. The various demographic variables, which were considered for the study, were gender and designation. Results: The results revealed that the faculty members perceived their emotional intelligence at an above-average level in the present pandemic, that is, COVID-19. The results also revealed that the perception of the respondent faculty members toward their emotional intelligence from different universities and states is more or less the same and also the demographic variable gender has a significant impact on emotional intelligence during the present pandemic. Conclusion: Besides having theoretical implications that open pathways for conducting further research, the findings of the study may serve as a reference for service practitioners in designing strategies that will ensure superior performance of faculty members in higher educational institutions during the pandemic.

scholarly journals Recommendation with Multi-Source Heterogeneous Information

2018 ◽  
Author(s):  
Li Gao ◽  
Hong Yang ◽  
Jia Wu ◽  
Chuan Zhou ◽  
Weixue Lu ◽  
...  
Keyword(s):  
Social Networks ◽  
Contextual Information ◽  
Representation Learning ◽  
Network Embedding ◽  
Real World Data ◽  
Heterogeneous Information ◽  
Coupled Neural Networks ◽  
Deep Network ◽  
Low Dimensional ◽  
Textual Content

Network embedding has been recently used in social network recommendations by embedding low-dimensional representations of network items for recommendation. However, existing item recommendation models in social networks suffer from two limitations. First, these models partially use item information and mostly ignore important contextual information in social networks such as textual content and social tag information. Second, network embedding and item recommendations are learned in two independent steps without any interaction. To this end, we in this paper consider item recommendations based on heterogeneous information sources. Specifically, we combine item structure, textual content and tag information for recommendation. To model the multi-source heterogeneous information, we use two coupled neural networks to capture the deep network representations of items, based on which a new recommendation model Collaborative multi-source Deep Network Embedding (CDNE for short) is proposed to learn different latent representations. Experimental results on two real-world data sets demonstrate that CDNE can use network representation learning to boost the recommendation performance.

Continuous surveillance of UK coastline using EO data to monitor coastal change impact

2021 ◽  
Author(s):  
Martin Jones ◽  
Andreas Payo. Garcia
Keyword(s):  
Land Use ◽  
Land Cover ◽  
Spatial Resolution ◽  
Large Scale ◽  
Contextual Information ◽  
Water Dynamics ◽  
Coastal Change ◽  
Flooding Risk ◽  
The Uk ◽  
The Impact

&lt;p&gt;The UK coast is under&amp;#160; increasing risk due to coastal change, cliffs are collapsing endangering houses near the coast and of the 12,400 km of&amp;#160; coastline, 2,500 km present a flooding risk. Constant monitoring is necessary in order to keep coastal evolution under surveillance and to adapt the measures to mitigate the impact of coastal change. Earth Observation technology is unique in that it has now been available for over 25 years and currently there is a range of satellites both civil and commercial that are constantly viewing our coast. Satellite imagery provides large scale observation at a high spatial resolution with an average revisit time of 5 days for most missions. Temporal and spatial resolution are key components to provide a continuous monitoring service of a coast. Using the balance of ever increasing resolution coupled to a range of innovative techniques that make full use of the spectral signatures being captured enables us to recreate the coastal boundary to a high degree of reliability over complete national coastlines.&lt;/p&gt;&lt;p&gt;Our developed methodology combines different types of products to completely characterize the different coastal environments. The land/sea boundary is used to monitor changes along the coast and combine with a backshore land use, land cover classification map, we are able to bring contextual information on coastal vulnerability and their erosive potential. Our LiuJezek_CoastL processor extracts the instantaneous land/sea boundary from all satellite observations available and provides a vector line which represents the coast morphology depending on sea level at the time of the acquisition. This line is then corrected from all water dynamics such as waves, tidal level to create shorelines at a reference datum height. The error in positioning the shoreline is relaint on beach slopes, for example in the case of cliffs or civil works along the coast compared to long shelfing beaches. Our backshore classification, provides land use and land cover information which can correct the shoreline position according to the features present along the coast.&lt;/p&gt;

Sign in / Sign up

Export Citation Format

Share Document