Our technology

AI technologies powering our recommendation platform.

Deep Collaborative Filtering

Our learnings and research.

Deep Collaborative filtering allows to understand and present complex patterns of human behavior. Whether it is through grouping data or referencing patterns, it is imperative to gather as much user information as possible from all available sources to acquire usable results.

Simple models fail to simultaneously learn global trends and rare individual preferences from the long tail. Deep Collaborative Filtering scales to settings surpassing traditional boundaries, such as cross-domain or unbalanced data sets.

Common approaches

Most other solutions use “Matrix Factorization” which reduces an individual's preferences to a simplistic linear model.

Using Deep Collaborative Filtering, our solution instead focuses on pinpointing complex patterns that our algorithms then use to determine user tastes.

Deep Content Extraction

Our learnings and research.

Recommending new items is a common challenge for businesses. More generally, any item without enough history of interactions from users won't be recommended by traditional methods. A novel solution to this cold-start problem is to extract deep information from text or image data using advanced natural language processing and creating a neural network of associations. Deep Content Extraction allows our software to recommend items no one has interacted with yet because the algorithm is able to understand the genre of the content by automatically extracting information from items such as a movie poster, synopsis, or reviews.

Common approaches

Other solutions commonly patch their recommender system with hand-crafted rules to diminish the cold-start problem. However this requires expensive iterations and does not scale.

Using Deep Content Extraction, our solution can generate accurate recommendations of items as soon as they are available in the data set.

Semantic Graph Embedding

Our learnings and research.

Semantic Graphing makes the goal of finding correlations between seemingly unrelated data simpler by using variables such as metadata, labels, tags, genre, actors and more. By doing this we can make sense of semantic data in the same way a human mind would. In a nutshell, if we know user likes a certain item, and this item is connected to a second item, for instance if two movies share certain key actors, then our algorithm can consider the second item as a promising candidate to recommend to our user.

Common approaches.

Most other solutions do not include any graph databases for their recommendations.

Our algorithm leverages all the available information to increase the accuracy of the recommendations, enabling users to discover hidden gems.

The importance of embeddings for recommendations.

An embedding is a low-dimensional vector of continuous numbers learned to represent relevant information in a compact and convenient form. Embeddings are designed to reduce the dimensionality of both categorical and continuous variables and meaningfully represent categories or features in the transformed space. 

When Machine learning and embeddings are combined it permits us the ability to more accurately determine which data qualifies as features.

More concretely, embedding serve as input to a machine learning model for a supervised task, such as classification or recommendation. These can be used to personalize an online visitor experience based on their interests.




Machine Learning Embedding

Our work.

Transforming raw data into features or embeddings is often part of the process that most heavily involves human time because it is driven by intuition. Recent developments in deep learning and automated processing of images, text, and signals have enabled significant automation in feature engineering for those data types. Yet, feature engineering for relational and human behavioral data remains iterative, human-intuition driven, and challenging, hence, time-consuming. Our goal is to do so in a domain or brand agnostic manner making our approach adaptable to any business and scalable to any market. 

And in simpler words.

Crossing Minds is the first and only company in the world to provide a full encoding of consumer’s behavior and taste, without ever putting their privacy at risk.  

We are able to build a “digital DNA” through the embedding of information gathered through cross-platforms by creating an anonymous mathematical representation of a consumer that contains all their tastes and behaviors.

Our learnings and research.

An embedding is a low-dimensional vector of continuous numbers learned to represent relevant information in a compact and convenient form. Embeddings are designed to reduce the dimensionality of both categorical and continuous variables and meaningfully represent categories or features in the transformed space. 

Embedding is the future when it comes to reducing the dimensionality of categorical and continuous variables. The challenge has always been deciding which clusters, classifications, and actions are significant and turning this data into features. 

Book a demo

Let's get started. Just one step and you'll be able to:

  • Get a full demo tailored to your industry and business needs

  • Find out how to get started with personalization, or how to take your program to the next level

  • Get a 30 days Free trial with our team working directly with you to build and deploy the very best models for your business

Tell us a bit about yourself

Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form
Interested in a career at Crossing Minds?