Maximize your impact with
the Yusp Deep Learning Modules

Benefit from the world-class Yusp personalization engine, combined with the Deep Learning Module, or use the Deep Learning as an Ensemble with your existing tech stack. Incorporate decades of scientific expertise across several domains.

Two ways to harness the power of deep learning

existing-stack Keep your existing stack and add the Yusp Deep Learning Module

Buy Yusp’s Deep Learning solution and integrate it into your personalization system and tech stack. We offer Deep Learning as a Managed Service (SaaS), or as an on-premise deployment with a perpetual license, complete with source code and training for your team.

Deep learning technology is a beast that’s difficult to tame. Its complexity demands massive computational capacity. Data scientists competent enough to handle it are few and far between. Failing to meet the requirements of productivization, most deep learning projects get abandoned or remain in pilot forever. 

If you want to avoid such pitfalls, you now have the opportunity to integrate Yusp’s tried and tested deep learning technology into your existing personalization system. The package includes source code and perpetual license, plus training and customization options to boot.

replace-engine Replace your engine with Yusp Personalization + Deep Learning

Get ahead of the curve by implementing a personalization engine turbocharged with a Deep Learning Module. Our engine won A/B tests versus other vendors 42 times over 10 years, and now has a state-of-the-art Deep Learning add-on to go further.

Users’ preferences and goals often vary from one site visit to another. In a process called hyper-personalization, our deep learning recommendation algorithms capture user intent on the spot, leveraging complex neural network-based AI and real-time data. These are highly optimized implementations of the algorithm GRU4Rec, which was developed by Gravity R&D with researchers from Telefonica and Netflix. 

On offer as an additional module to Yusp since 2019, deep learning-based recommendations can improve main business KPIs by 15 to 20 percent compared to conventional collaborative filtering algorithms. Ideal for e-businesses with large product or content catalogs.


The Yusp success powered by

Gravity is a forerunner in applying deep learning for personalization and has a technology partnership with NVIDIA since it has been selected for NVIDIA’s Inception Program in 2020.

Our unique solution provides proven high performance on a cost-effective basis, which is scalable and applicable in multiple industries. The algorithm portfolio is our proprietary know-how, while the solution utilizes NVIDIA’s specific particular hardware and software technologies.

Our technology partnership has recently been leveraged to the next level as Gravity has been selected as a member of the NVIDIA ISV Program in 2021.

Proven track record and success with the Yusp deep learning algorithms

Uplift in Gross Merchandising value

Significant increase in revenue generated by recommendations, Turkey’s largest marketplace, compared their existing personalization solutions with Yusp.

  • Yusp by Gravity R&D covers an omnichannel personalized experience across our marketing channels, helping our visitors find the most relevant products and then convert to customers. With a dedicated customer support team, they are flexible to work with us on personalizing

    Öykü Kulaksız Sağlam

    Marketing Analytics Manager

    arrow_back_ios arrow_forward_ios

Get in Touch to Learn More

Our colleagues will guide you through the most important questions and help you get started with our Proof of Concept period.

All your questions about Deep Learning, answered

  • We have another recommendation system. Is it possible to complement it with Yusp’s Deep Learning technology?

    Yes. You can subscribe to the Yusp Deep Learning Module as a managed service, or if you prefer you can purchase the standalone Deep Learning Module – including Yusp’s unique Deep Learning algorithms, source code and perpetual license – and we’ll train you and go away, or if you want we'll help you integrate it seamlessly into your existing recommendation system. The new era for Deep Learning as part of your ensemble personalization strategy has arrived, and it is meant to fit your approach.

    The Deep Learning algorithms will become your own, leveraging what we believe is the world's best architectures, enabling your team to get up the learning curve in Deep Learning, customizing to the use cases that deliver highest performance and CX.

    And the best part is it can run on CPUs in production, so costs are kept low and ROI comes quickly for your organization - and we’re always ready to provide specialized training or support, should you need any.

  • How can I use the Yusp Deep Learning Module for my business?

    There are two ways to leverage Deep Learning, depending on your business objectives, in-house capabilities and tech roadmap.

    If you already have a recommendation system and a thorough knowledge of your algorithms and data, you can add Yusp’s Deep Learning Module, either as a SaaS managed service or as a perpetual license including source code. This way, you can integrate Deep Learning algorithms with your existing algorithm portfolio, recommendation engine and technology architecture, but still gain the benefits of having Deep Learning as part of an ensemble solution.

    If your Board is looking for an overhaul -- as in going with a top vendor with both a world class personalization engine as well as State-of-the-Art Deep Learning capability -- you can replace your existing recommendation engine with Yusp and we'll include the Deep Learning Module, an extension of the default setup, on a single subscription basis. For traditional retailers who want to compete with today's leading marketplaces or video streamers, this will get you there. We'll guarantee based on a benchmark assessment a significant increase in top line growth while also decreasing your operational costs.

    For either of these deep learning-related projects – Yusp’s Deep Learning experts are ready to help.

  • Deep Learning technology by Yusp sounds impressive, but what’s in it for me?

    Yusp Enterprise lowers the threshold of upfront investment to a historic low.

    As a result, integrating Deep Learning into the personalization engines can be ROI positive for the world's leading marketplaces as well as traditional retailers who lack the in-house skills to deploy and support Deep Learning.

    On top of that, GDPR compliance is a given with personalization engines running Deep Learning algorithms, because these systems don’t need individual user data in order to generate recommendations. Anyone serving European customers or planning to be compliant with the many privacy laws on the horizon can actually increase revenue while adding gdpr compliance.

    The biggest business reason for Deep Learning is that it will uplift KPIs like CTR, GMV & conversion rates by up to 20 percent. This means adding 8 to 9 digits in additional revenue to an already well-tuned ecommerce & recommendation engine stack.

  • What’s the difference between the SaaS Deep Learning Module and the on-premise Deep Learning Module offered by Yusp?

    The Deep Learning Module comprises the best practice and proprietary Deep Learning algorithm, GRU4Rec. GRU4Rec is the default algorithm portfolio used as the benchmark (1200 citations) for some of the world's biggest tech leaders, it’s been available commercially as a SaaS offering to traditional industry since 2019.

    Beginning in 2021, we created an 'on-premise' package that allows any business to purchase as a perpetual license the Deep Learning algorithms along with their source code. There is no other vendor in the world providing a SaaS and on-premise Deep Learning module, and capable of leveling the playing field between the market leaders and traditional industry. 

  • What’s deep learning, really?

    Deep Learning is a subset of machine learning methods. It’s a multi-layered, complex calculation process involving artificial neural networks.

    The functioning and structure of artificial neural networks resembles biological systems, like the nervous system.

    The past 10 years have not seen true adoptions of Deep Learning architectures in ecommerce sites and online retail, except in the top 1% of industry by the global market leaders in retail, marketplaces, video streaming and news sites.

    While a conventional collaborative filtering algorithm takes into account a limited array of variables, a deep learning algorithm calculates with multiple “layers” of data from a wide range of sources (including images, sound, and video), producing much more nuanced results

    Personalization is only one of many areas where Deep Learning has fueled revolutionary progress. Others include natural language processing and computer vision. Chatbots, virtual assistants, and other applications generally referred to as “AI” operate with deep learning methods.

    Our goal is to make Deep Learning accessible to traditional industry and SMEs, re-level the playing field and allow creativity and lifetime value for the customer again be the decisive factor in business.

  • We’re already using Yusp. How can we benefit from your deep learning know-how?

    Simple. Just order Yusp’s Deep Learning Module for an adjusted monthly fee, and we’ll plug it in for you. Then sit back and watch your revenue grow as the Deep Learning algorithms boost the efficiency of your recommendations by as much as 20 percent.

  • We’re building our own deep learning solution, and it’s not related to personalization. Can Yusp’s deep learning assets still be relevant for us?

    Yes. Our know-how is a resource we’ll be happy to share. We have a team of deep learning experts for hire, well versed in all main DL frameworks (TensorFlow, PyTorch, Theano), with years of experience conducting their own research. They can assist you with almost any deep learning-related project, well beyond the field of personalization. Should you need any guidance or hands-on support with your deep learning endeavor, you can get them on board as external or on-premise consultants.

Helpful resources