• Corpus ID: 7723585

Two Decades of Recommender Systems at Amazon

@inproceedings{Blake2017TwoDO,
  title={Two Decades of Recommender Systems at Amazon},
  author={M. Brian Blake},
  year={2017}
}
F or two decades now, 1 Amazon. com has been building a store for every customer. Each person who comes to Amazon.com sees it differently, because it’s individually personalized based on their interests. It’s as if you walked into a store and the shelves started rearranging themselves, with what you might want moving to the front, and what you’re unlikely to be interested in shuffling further away. From a catalog of hundreds of millions of items, Amazon.com’s recommendations pick a small number… 

Figures from this paper

Movie Recommender Systems Made Through Tag Interpolation
TLDR
This paper exploits the movie recommender system based on movies' genres and actors/actresses themselves as the input tags, or tag interpolation, for 100K variations of the MovieLens datasets and applies tag-based filtering and collaborative filtering that can effectively predict a list of movies that is similar to the movie that a user has been watched.
Accelerated learning from recommender systems using multi-armed bandit
TLDR
It is argued that multi armed bandit (MAB) testing as a solution to issues with A/B testing, including long lead time, and high cost of exploration, should be considered.
Improving collaborative recommendation based on item weight link prediction
TLDR
This work focuses only on binary bipartite graphs, whereas admiration can be expressed on a discrete scale where link weighting is involved, and addresses the problem of link prediction in unweighted bipartites networks.
E-Commerce Recommender System Using Product Data
TLDR
A recommender system is a system that helps to expect and recommend similar products to the given input of the product.
A Semi-Personalized System for User Cold Start Recommendation on Music Streaming Apps
TLDR
This paper presents a semi-personalized recommendation strategy, based on a deep neural network architecture and on a clustering of users from heterogeneous sources of information, recently deployed on the music streaming service Deezer to address the user cold start problem.
Interest Sustainability-Aware Recommender System
TLDR
This work proposes a novel recommender system, called CRIS, that considers the change of users' interest in each item over time, and predicts the interest sustainability of each item, that is, how likely each item will be consumed in the future.
Knowledge aggregation in people recommender systems: matching skills to tasks
TLDR
This thesis explores how people’s preferences can be elicited for single-event matchmaking under uncertainty and how to align them with appropriate tasks.
Personalizing Item Recommendation via Price Understanding
TLDR
This paper segregate items into price bands indicating how expensive they are, infer user affinity to price bands based on historical behavior and use features derived from this knowledge to re-rank items in a real-world recommendation scenario.
On the Generalizability and Predictability of Recommender Systems
TLDR
RecZilla is created, a meta-learning approach to recommender systems that uses a model to predict the best algorithm and hyperparameters for new, unseen datasets, and is able to substantially reduce the level of human involvement when faced with a new recommender system application.
DESIGNING A HOLISTIC ADAPTIVE RECOMMENDER SYSTEM (HARS) FOR CUSTOMER RELATIONSHIP DEVELOPMENT: A CONCEPTUAL FRAMEWORK
TLDR
A novel Reinforcement Learning-based recommender system that has an integrative view over data and recommendation landscape, as well as it is highly adaptive to changes in customer behavior is proposed, the Holistic Adaptive Recommender System (HARS).
...
...

References

SHOWING 1-9 OF 9 REFERENCES
Estimating the Causal Impact of Recommendation Systems from Observational Data
TLDR
This paper shows that causal identification through an instrumental variable is possible when a product experiences an instantaneous shock in direct traffic and the products recommended next to it do not, and applies a method for estimating causal effects from purely observational data to browsing logs containing anonymized activity on Amazon.com.
Evaluating collaborative filtering recommender systems
TLDR
The key decisions in evaluating collaborative filtering recommender systems are reviewed: the user tasks being evaluated, the types of analysis and datasets being used, the ways in which prediction quality is measured, the evaluation of prediction attributes other than quality, and the user-based evaluation of the system as a whole.
Using Collaborative Filtering to Automate Worker-Job Recommendations for Crowdsourcing Services
  • Julian JarrettM. Blake
  • Computer Science, Business
    2016 IEEE International Conference on Web Services (ICWS)
  • 2016
TLDR
A worker-job matching model for crowdsourcing supported by a service-oriented architecture that can effectively infer the levels of expertise of potential crowdworkers based on their profile and past performance history is introduced.
Self-Generating a Labor Force for Crowdsourcing: Is Worker Confidence a Predictor of Quality?
TLDR
This paper discusses the state-of-the-art in recommending crowd workers based on assessment information and presents a study exploring the validity of skills input by workers in light of their actual performance and other metrics captured by employers.
The YouTube video recommendation system
TLDR
The video recommendation system in use at YouTube, the world's most popular online video community, is discussed, with details on the experimentation and evaluation framework used to test and tune new algorithms.
The Netflix Recommender System
TLDR
The motivations behind and approach that Netflix uses to improve the recommendation algorithms are explained, combining A/B testing focused on improving member retention and medium term engagement, as well as offline experimentation using historical member engagement data.
Method and System for Associating Feedback with Recommendation Rules, US Patent 8,090,621
  • Patent and Trademark Office,
  • 2012
System for Detecting Probabilistic Associations between Items, US Patent 8,239,287, to Amazon.com
  • Patent and Trademark Office,
  • 2012
Benson, Collaborative Recommendations Using Item-to-Item Similarity Mappings, US Patent 6,266,649, to Amazon.com
  • Patent and Trademark Office,
  • 2001