<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Booking.com Data Science on Medium]]></title>
        <description><![CDATA[Stories by Booking.com Data Science on Medium]]></description>
        <link>https://medium.com/@datascience.blog?source=rss-1aecbb50db31------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*fKAKI9Xl8A4CG6GpTuHeQQ.png</url>
            <title>Stories by Booking.com Data Science on Medium</title>
            <link>https://medium.com/@datascience.blog?source=rss-1aecbb50db31------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Mon, 16 Nov 2026 15:44:15 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@datascience.blog/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[A fundamental problem of hypothesis testing with finite inventory in e-commerce]]></title>
            <link>https://booking.ai/a-fundamental-problem-of-hypothesis-testing-with-finite-inventory-in-e-commerce-892bfa751d72?source=rss-1aecbb50db31------2</link>
            <guid isPermaLink="false">https://medium.com/p/892bfa751d72</guid>
            <category><![CDATA[publication]]></category>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[statistics]]></category>
            <dc:creator><![CDATA[Booking.com Data Science]]></dc:creator>
            <pubDate>Thu, 29 Oct 2020 11:12:49 GMT</pubDate>
            <atom:updated>2020-10-29T11:12:49.667Z</atom:updated>
            <content:encoded><![CDATA[<p>Published in Applied Stochastic Models in Business and Industry by Dennis Bohle, Alexander Marynych, Matthias Meiners</p><p>Paper: <a href="https://onlinelibrary.wiley.com/doi/epdf/10.1002/asmb.2574">https://onlinelibrary.wiley.com/doi/epdf/10.1002/asmb.2574</a></p><p>We draw attention to a problem that is often overlooked or ignored by companies practicing hypothesis testing (A/B testing) in online environments. We show that conducting experiments on limited inventory that is shared between variants in the experiment can lead to high false‐positive rates since the core assumption of independence between the groups is violated. We provide a detailed analysis of the problem in a simplified setting whose parameters are informed by realistic scenarios. The setting we consider is a two‐dimensional (2D) random walk in a semiinfinite strip. It is rich enough to take a finite inventory into account, but is at the same time simple enough to allow for a closed form of the false‐positive probability. We prove that high false‐positive rates can occur, and develop tools that are suitable to help design adequate tests in follow‐up work. Our results also show that high false‐negative rates may occur. The proofs rely on a functional limit theorem for the 2D random walk in a semiinfinite strip.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=892bfa751d72" width="1" height="1" alt=""><hr><p><a href="https://booking.ai/a-fundamental-problem-of-hypothesis-testing-with-finite-inventory-in-e-commerce-892bfa751d72">A fundamental problem of hypothesis testing with finite inventory in e-commerce</a> was originally published in <a href="https://booking.ai/">Booking.com ML &amp; DS Blog</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Recommending Accommodation Filters with Online Learning]]></title>
            <link>https://booking.ai/recommending-accommodation-filters-with-online-learning-f035ef4cd65b?source=rss-1aecbb50db31------2</link>
            <guid isPermaLink="false">https://medium.com/p/f035ef4cd65b</guid>
            <category><![CDATA[fraud-detection]]></category>
            <category><![CDATA[online-learning]]></category>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[recsys]]></category>
            <category><![CDATA[machine-learning]]></category>
            <dc:creator><![CDATA[Booking.com Data Science]]></dc:creator>
            <pubDate>Thu, 01 Oct 2020 18:57:23 GMT</pubDate>
            <atom:updated>2020-11-05T23:16:28.468Z</atom:updated>
            <content:encoded><![CDATA[<p>Presented in the Workshop on Online Recommender Systems and User Modeling <a href="https://orsum.inesctec.pt/orsum2020/">ORSUM</a> ACM <a href="https://recsys.acm.org/recsys20/">RecSys 2020</a> by <a href="https://medium.com/u/fe16abb58a96">Lucas Bernardi</a>, <a href="https://medium.com/u/fb1bbdbdb049">Pablo Estevez</a>, Matias Eidis and Eqbal Osama</p><p><a href="http://ceur-ws.org/Vol-2715/paper3.pdf">Paper</a></p><h3><strong>Abstract</strong></h3><p>Online Accommodations Platforms match guests searching for accommodation with hospitality service providers. A fundamental characteristic of efficient platforms is the ability to satisfy the needs and preferences of the guests. To achieve this goal, a common search tool is the Results Filtering capability which allows users to refine query results with explicit criteria. However, as supply grows and diversifies, more filtering options become available, reaching hundreds of different criteria for one query, and making it hard for customers to find the ones that are relevant to them. In this work we present the implementation of an Accommodation Filters Recommender System addressing this issue. The problem poses several challenges around recommendations feedback, user experience constraints, and non stationarity among others. We provide an end-to-end description of the System, discuss implementation issues and provide techniques to address them including a large scale distributed online learning architecture. The solution was validated through several Online Controlled Experiments performed in Booking.com, a top Online Travel Agency serving millions of daily users, showing statistically significant results on various user behaviour metrics indicating a strong positive effect on User Engagement.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=f035ef4cd65b" width="1" height="1" alt=""><hr><p><a href="https://booking.ai/recommending-accommodation-filters-with-online-learning-f035ef4cd65b">Recommending Accommodation Filters with Online Learning</a> was originally published in <a href="https://booking.ai/">Booking.com ML &amp; DS Blog</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Beyond algorithms: Ranking at scale at Booking.com]]></title>
            <link>https://booking.ai/beyond-algorithms-ranking-at-scale-at-booking-com-b1f43c1a2875?source=rss-1aecbb50db31------2</link>
            <guid isPermaLink="false">https://medium.com/p/b1f43c1a2875</guid>
            <category><![CDATA[platform-engineering]]></category>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[system-architecture]]></category>
            <category><![CDATA[rankings]]></category>
            <category><![CDATA[machine-learning]]></category>
            <dc:creator><![CDATA[Booking.com Data Science]]></dc:creator>
            <pubDate>Thu, 24 Sep 2020 11:16:13 GMT</pubDate>
            <atom:updated>2020-09-24T15:49:10.400Z</atom:updated>
            <content:encoded><![CDATA[<p>Published in Recsys 2020, ComplexRec by <a href="https://medium.com/u/b5dc69191d6c">Themis Mavridis</a>, <a href="https://medium.com/u/c53f15c03356">Soraya Hausl</a>, <a href="https://medium.com/u/d9f68e451424">Andrew Mende</a>, <a href="https://medium.com/u/bfbcaee084d5">Roberto Pagano</a></p><p>Paper: <a href="http://www.toinebogers.com/workshops/complexrec2020/Mavridis.pdf">http://www.toinebogers.com/workshops/complexrec2020/Mavridis.pdf</a></p><p>Booking.com is one of the world’s largest online travel companies where millions of customers find their accommodation through an extensive and diverse list of properties including hotels, apartments, guest houses, and more. Our customers heavily rely on the ranking of search results in order to find a property that satisfies their preferences and criteria. While most of the Machine Learning literature focuses on the algorithmic aspect of ranking, not much has been published about the intricacies of developing a ranking that is Machine Learning powered and delivers business impact in a commercial environment. In this work, we describe the various challenges we faced while developing a large-scale Machine Learning powered ranking, including signal definition and feature representation complexity in the modelling aspect, information leakage and speed of innovation in experimentation for online evaluation, and response time minimization around serving at scale in production. We display an overview of the business gains of the step changes in the Machine Learning ranking of Booking.com. Our main conclusion is that there are multiple aspects that need to be carefully addressed for the creation of a Machine Learned ranker in a large-scale commercial setting</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=b1f43c1a2875" width="1" height="1" alt=""><hr><p><a href="https://booking.ai/beyond-algorithms-ranking-at-scale-at-booking-com-b1f43c1a2875">Beyond algorithms: Ranking at scale at Booking.com</a> was originally published in <a href="https://booking.ai/">Booking.com ML &amp; DS Blog</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Efficient Image Gallery Representations at Scale Through Multi-Task Learning]]></title>
            <link>https://booking.ai/efficient-image-gallery-representations-at-scale-through-multi-task-learning-1ac099a97d47?source=rss-1aecbb50db31------2</link>
            <guid isPermaLink="false">https://medium.com/p/1ac099a97d47</guid>
            <category><![CDATA[deep-learning]]></category>
            <category><![CDATA[computer-vision]]></category>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[neural-networks]]></category>
            <category><![CDATA[publication]]></category>
            <dc:creator><![CDATA[Booking.com Data Science]]></dc:creator>
            <pubDate>Tue, 28 Jul 2020 09:01:36 GMT</pubDate>
            <atom:updated>2020-07-28T09:01:36.521Z</atom:updated>
            <content:encoded><![CDATA[<p>Published in SIGIR ‘20 by Benjamin Gutelman, <a href="https://medium.com/u/15902baf85f0">Pavel Levin</a></p><p>Paper: <a href="https://dl.acm.org/doi/abs/10.1145/3397271.3401433">https://dl.acm.org/doi/abs/10.1145/3397271.3401433</a></p><p>Image galleries provide a rich source of diverse information about a product which can be leveraged across many recommendation and retrieval applications. We study the problem of building a universal image gallery encoder through multi-task learning (MTL) approach and demonstrate that it is indeed a practical way to achieve generalizability of learned representations to new downstream tasks. Additionally, we analyze the relative predictive performance of MTL-trained solutions against optimal and substantially more expensive solutions, and find signals that MTL can be a useful mechanism to address sparsity in low-resource binary tasks.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=1ac099a97d47" width="1" height="1" alt=""><hr><p><a href="https://booking.ai/efficient-image-gallery-representations-at-scale-through-multi-task-learning-1ac099a97d47">Efficient Image Gallery Representations at Scale Through Multi-Task Learning</a> was originally published in <a href="https://booking.ai/">Booking.com ML &amp; DS Blog</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Cross-lingual intent classification in a low resource industrial setting]]></title>
            <link>https://booking.ai/cross-lingual-intent-classification-in-a-low-resource-industrial-setting-37f9f88cf87b?source=rss-1aecbb50db31------2</link>
            <guid isPermaLink="false">https://medium.com/p/37f9f88cf87b</guid>
            <category><![CDATA[neural-networks]]></category>
            <category><![CDATA[computer-vision]]></category>
            <category><![CDATA[deep-learning]]></category>
            <category><![CDATA[ai-applications]]></category>
            <category><![CDATA[machine-learning]]></category>
            <dc:creator><![CDATA[Booking.com Data Science]]></dc:creator>
            <pubDate>Fri, 18 Oct 2019 15:12:24 GMT</pubDate>
            <atom:updated>2020-10-29T11:00:55.641Z</atom:updated>
            <content:encoded><![CDATA[<p>Published in EMNLP 2019 by <a href="https://medium.com/u/214a2c2ffa0d">Talaat Khalil</a>, <a href="https://medium.com/u/a02a256c9009">Kornel Kielczewski</a>, George Chouliaras, Amina Keldibek and Maarten Versteegh</p><p>This paper explores different approaches to multilingual intent classification in a low resource setting. Recent advances in multilingual text representations promise cross-lingual transfer for classifiers. We investigate the potential for this transfer in an applied industrial setting and compare to multilingual classification using machine translated text. Our results show that while the recently developed methods show promise, practical application calls for a combination of techniques for useful results.</p><p>Link: <a href="https://www.aclweb.org/anthology/D19-1676/">https://www.aclweb.org/anthology/D19-1676/</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=37f9f88cf87b" width="1" height="1" alt=""><hr><p><a href="https://booking.ai/cross-lingual-intent-classification-in-a-low-resource-industrial-setting-37f9f88cf87b">Cross-lingual intent classification in a low resource industrial setting</a> was originally published in <a href="https://booking.ai/">Booking.com ML &amp; DS Blog</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Combining Context Features in Sequence-Aware Recommender Systems]]></title>
            <link>https://booking.ai/combining-context-features-in-sequence-aware-recommender-systems-b6691bdd2f2d?source=rss-1aecbb50db31------2</link>
            <guid isPermaLink="false">https://medium.com/p/b6691bdd2f2d</guid>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[recurrent-neural-network]]></category>
            <category><![CDATA[fraud-detection]]></category>
            <category><![CDATA[recommender-systems]]></category>
            <category><![CDATA[predictive-analytics]]></category>
            <dc:creator><![CDATA[Booking.com Data Science]]></dc:creator>
            <pubDate>Wed, 28 Aug 2019 18:49:04 GMT</pubDate>
            <atom:updated>2019-08-28T18:49:04.000Z</atom:updated>
            <content:encoded><![CDATA[<p>presented in RecSys 2019 Late-breaking Results, 16th-20th September 2019, Copenhagen, Denmark by Sarai Mizrachi and <a href="https://medium.com/u/15902baf85f0">Pavel Levin</a></p><p><strong>Paper</strong>: <a href="http://ceur-ws.org/Vol-2431/paper3.pdf">http://ceur-ws.org/Vol-2431/paper3.pdf</a></p><p>There are several important design choices that machine learning practitioners need to make when incorporating predictors into RNN-based contextual recommender systems. A great deal of currently reported findings about these decisions focus on the setting where predicted items take on values from the space of sequence items. This work provides an empirical evaluation of some straightforward approaches of dealing with such problems on a real-world large scale prediction problem from the travel domain, where predicted entities do not live in the space of sequence items.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=b6691bdd2f2d" width="1" height="1" alt=""><hr><p><a href="https://booking.ai/combining-context-features-in-sequence-aware-recommender-systems-b6691bdd2f2d">Combining Context Features in Sequence-Aware Recommender Systems</a> was originally published in <a href="https://booking.ai/">Booking.com ML &amp; DS Blog</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[150 Successful Machine Learning Models: 6 Lessons Learned at Booking.com]]></title>
            <link>https://booking.ai/150-successful-machine-learning-models-6-lessons-learned-at-booking-com-681e09107bec?source=rss-1aecbb50db31------2</link>
            <guid isPermaLink="false">https://medium.com/p/681e09107bec</guid>
            <category><![CDATA[experiment]]></category>
            <category><![CDATA[predictive-analytics]]></category>
            <category><![CDATA[recommender-systems]]></category>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[data-science]]></category>
            <dc:creator><![CDATA[Booking.com Data Science]]></dc:creator>
            <pubDate>Mon, 12 Aug 2019 12:33:52 GMT</pubDate>
            <atom:updated>2022-02-27T14:06:18.489Z</atom:updated>
            <content:encoded><![CDATA[<p>Published in KDD 2019 by <a href="https://medium.com/u/fe16abb58a96">Lucas Bernardi</a>, <a href="https://medium.com/u/b5dc69191d6c">Themis Mavridis</a> and <a href="https://medium.com/u/fb1bbdbdb049">Pablo Estevez</a></p><p><strong>Paper</strong>: <a href="https://dl.acm.org/doi/10.1145/3292500.3330744">https://dl.acm.org/doi/10.1145/3292500.3330744</a></p><p>Booking.com is the world’s largest online travel agent where millions of guests find their accommodation and millions of accommodation providers list their properties including hotels, apartments, bed and breakfasts, guest houses, and more. During the last years we have applied Machine Learning to improve the experience of our customers and our business. While most of the Machine Learning literature focuses on the algorithmic or mathematical aspects of the field, not much has been published about how Machine Learning can deliver meaningful impact in an industrial environment where commercial gains are paramount. We conducted an analysis on about 150 successful customer facing applications of Machine Learning, developed by dozens of teams in Booking.com, exposed to hundreds of millions of users worldwide and validated through rigorous Randomized Controlled Trials. Following the phases of a Machine Learning project we describe our approach, the many challenges we found, and the lessons we learned while scaling up such a complex technology across our organization. Our main conclusion is that an iterative, hypothesis driven process, integrated with other disciplines was fundamental to build 150 successful products enabled by Machine Learning.</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2FDEjyUkbVVHA%3Ffeature%3Doembed&amp;display_name=YouTube&amp;url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DDEjyUkbVVHA&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2FDEjyUkbVVHA%2Fhqdefault.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=youtube" width="854" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/4251fc9cb7b771f11a419499e6f7f0eb/href">https://medium.com/media/4251fc9cb7b771f11a419499e6f7f0eb/href</a></iframe><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=681e09107bec" width="1" height="1" alt=""><hr><p><a href="https://booking.ai/150-successful-machine-learning-models-6-lessons-learned-at-booking-com-681e09107bec">150 Successful Machine Learning Models: 6 Lessons Learned at Booking.com</a> was originally published in <a href="https://booking.ai/">Booking.com ML &amp; DS Blog</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Mediation Analysis in Online Experiments at Booking.com: Disentangling Direct and Indirect Effects]]></title>
            <link>https://booking.ai/mediation-analysis-in-online-experiments-at-booking-com-disentangling-direct-and-indirect-effects-f27207d51f02?source=rss-1aecbb50db31------2</link>
            <guid isPermaLink="false">https://medium.com/p/f27207d51f02</guid>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[experimentation]]></category>
            <category><![CDATA[publication]]></category>
            <dc:creator><![CDATA[Booking.com Data Science]]></dc:creator>
            <pubDate>Thu, 01 Nov 2018 14:55:40 GMT</pubDate>
            <atom:updated>2018-11-20T09:25:08.231Z</atom:updated>
            <content:encoded><![CDATA[<p>Published in CODE 2018 by <a href="https://medium.com/u/56d5f68b6745">Tolga Oztan</a>, Zoé van Havre, Caio Gomes, <a href="https://medium.com/u/c79c818be3ab">Lukas Vermeer</a></p><p>Paper: <a href="https://arxiv.org/abs/1810.12718">https://arxiv.org/abs/1810.12718</a></p><p>Online experimentation is at the core of Booking.com’s customer-centric product development. While randomised controlled trials are a powerful tool for estimating the overall effects of product changes on business metrics, they often fall short in explaining the mechanism of change. This becomes problematic when decision-making depends on being able to distinguish between the direct effect of a treatment on some outcome variable and its indirect effect via a mediator variable. In this paper, we demonstrate the need for mediation analyses in online experimentation, and use simulated data to show how these methods help identify and estimate direct causal effect. Failing to take into account all confounders can lead to biased estimates, so we include sensitivity analyses to help gauge the robustness of estimates to missing causal factors.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=f27207d51f02" width="1" height="1" alt=""><hr><p><a href="https://booking.ai/mediation-analysis-in-online-experiments-at-booking-com-disentangling-direct-and-indirect-effects-f27207d51f02">Mediation Analysis in Online Experiments at Booking.com: Disentangling Direct and Indirect Effects</a> was originally published in <a href="https://booking.ai/">Booking.com ML &amp; DS Blog</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Learning to Match]]></title>
            <link>https://booking.ai/learning-to-match-37c11270c4f4?source=rss-1aecbb50db31------2</link>
            <guid isPermaLink="false">https://medium.com/p/37c11270c4f4</guid>
            <category><![CDATA[experimentation]]></category>
            <category><![CDATA[publication]]></category>
            <category><![CDATA[machine-learning]]></category>
            <dc:creator><![CDATA[Booking.com Data Science]]></dc:creator>
            <pubDate>Thu, 14 Jun 2018 13:57:26 GMT</pubDate>
            <atom:updated>2018-11-01T15:08:03.755Z</atom:updated>
            <content:encoded><![CDATA[<p>Published in WSDM 2018 by <a href="https://medium.com/u/b5dc69191d6c">Themis Mavridis</a>, <a href="https://medium.com/u/fb1bbdbdb049">Pablo Estevez</a>, <a href="https://medium.com/u/fe16abb58a96">Lucas Bernardi</a></p><p><strong>Paper: </strong><a href="https://arxiv.org/pdf/1802.03102.pdf">https://arxiv.org/pdf/1802.03102.pdf</a></p><p>Booking.com is a virtual two-sided marketplace where guests and accommodation providers are the two distinct stakeholders. They meet to satisfy their respective and different goals. Guests want to be able to choose accommodations from a huge and diverse inventory, fast and reliably within their requirements and constraints. Accommodation providers desire to reach a reliable and large market that maximizes their revenue. Finding the best accommodation for the guests, a problem typically addressed by the recommender systems community, and finding the best audience for the accommodation providers, are key pieces of a good platform. This work describes how Booking.com extends such approach, enabling the guests themselves to find the best accommodation by helping them to discover their needs and restrictions, what the market can actually offer, reinforcing good decisions, discouraging bad ones, etc. turning the platform into a decision process advisor, as opposed to a decision maker. Booking.com implements this idea with hundreds of Machine Learned Models, all of them validated through rigorous Randomized Controlled Experiments. We further elaborate on model types, techniques, methodological issues and challenges that we have faced.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=37c11270c4f4" width="1" height="1" alt=""><hr><p><a href="https://booking.ai/learning-to-match-37c11270c4f4">Learning to Match</a> was originally published in <a href="https://booking.ai/">Booking.com ML &amp; DS Blog</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Toward a full-scale neural machine translation in production: the Booking.com use case]]></title>
            <link>https://booking.ai/toward-a-full-scale-neural-machine-translation-in-production-the-booking-com-use-case-4ee47da0a562?source=rss-1aecbb50db31------2</link>
            <guid isPermaLink="false">https://medium.com/p/4ee47da0a562</guid>
            <category><![CDATA[nlp]]></category>
            <category><![CDATA[publication]]></category>
            <dc:creator><![CDATA[Booking.com Data Science]]></dc:creator>
            <pubDate>Thu, 14 Jun 2018 13:56:52 GMT</pubDate>
            <atom:updated>2018-11-01T15:06:43.856Z</atom:updated>
            <content:encoded><![CDATA[<p>Published in MT Summit XVI by <a href="https://medium.com/u/15902baf85f0">Pavel Levin</a>, <a href="https://medium.com/u/cea5c6562631">Nishikant Dhanuka</a>, Talaat Khalil, Fedor Kovalev, Maxim Khalilov</p><p><strong>Paper</strong>: <a href="https://arxiv.org/pdf/1709.05820.pdf">https://arxiv.org/pdf/1709.05820.pdf</a></p><p>While some remarkable progress has been made in neural machine translation (NMT) research, there have not been many reports on its development and evaluation in practice. This paper tries to fill this gap by presenting some of our findings from building an in-house travel domain NMT system in a large scale E-commerce setting. The three major topics that we cover are optimization and training (including different optimization strategies and corpus sizes), handling real-world content and evaluating results.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=4ee47da0a562" width="1" height="1" alt=""><hr><p><a href="https://booking.ai/toward-a-full-scale-neural-machine-translation-in-production-the-booking-com-use-case-4ee47da0a562">Toward a full-scale neural machine translation in production: the Booking.com use case</a> was originally published in <a href="https://booking.ai/">Booking.com ML &amp; DS Blog</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>