This repository includes some papers that I have read or which I think may be very interesting.

H1st AI solves the critical “cold-start” problem of Industrial AI: encoding human expertise to augment the lack of data, while building a smooth transition toward a machine-learning future. This problem has caused most industrial-AI projects to fail.

Parameter-Efficient Transfer from Sequential Behaviors for User Modeling and Recommendation

(Python) Collective matrix factorization with cold-start functionality (recommender systems)

Scaffolding for Cloud Functions deployed with the Firebase CLI that minimize cold starts by using TypeScript async imports

Accompanying code for reproducing experiments from the HybridSVD paper. Preprint is available at https://arxiv.org/abs/1802.06398.

Python implementation of "Content-based recommendations with poisson factorization", with some extensions

Source code for KDD 2020 paper "Meta-learning on Heterogeneous Information Networks for Cold-start Recommendation"

keep lambdas warm and monitor cold starts with a simple decorator

[ACL2018] Cold-Start Aware User and Product Attention for Sentiment Classification

Our solution for Recsys Challenge 2017.

A recommender engine built for a Bay Area online dating website to maximize the successful matches by introducing hybrid recommender system and reverse match technique.

Data generation, model training and evaluation pipelines for the cold-start setting.

Example lightweight DynamoDB integration that limits AWS Lambda cold starts for Java

A service which creates a user preferences profile in order to solve the cold start issue for the recommending systems

Case study project for recommending movies to users.

Keep your Lambdas warm!

Facebook Friend Recommendation is a peer recommender system using Graph-Mining and Supervised Machine Learning techniques to find similarity between multiple users and predict users on the basis of extensive feature engineering

Personalised and popularity-based movie recommendations.

Collaborating Filtering Project

AWS Lambda custom runtime with GraalVM native-image binary

Apps used in serverless experiments

Java library for building native AWS Lambda runtimes

AWS low-level API client

Deep-learning-based service recommendation in a cold-start scenario of mashup development