Deep Recommendation

Deep recommendation systems leverage deep learning to personalize user experiences by predicting preferences based on vast datasets of user interactions and item features. Current research focuses on improving model efficiency and robustness, addressing challenges like covariate shift (changes in data distribution over time), hardware limitations (e.g., memory bottlenecks and error tolerance), and optimizing training pipelines through techniques such as reinforcement learning and efficient embedding table sharding. These advancements aim to enhance the accuracy, scalability, and cost-effectiveness of recommendation systems across various applications, impacting both the scientific understanding of large-scale machine learning and the practical deployment of personalized services.

Papers

November 12, 2024

AdaS&S: a One-Shot Supernet Approach for Automatic Embedding Size Search in Deep Recommender System
He Wei, Yuekui Yang, Yang Zhang, Haiyang Wu, Meixi Liu, Shaoping Ma
Deep Learning Jina Embeddings Supernet Training Well Trained Deep Recommendation

October 26, 2024

DQRM: Deep Quantized Recommendation Models
Yang Zhou, Zhen Dong, Ellick Chan, Dhiraj Kalamkar, Diana Marculescu, Kurt Keutzer
Recommendation Model Novel Recommendation Large Scale Recommendation System Deep Recommendation

August 28, 2023

Ad-Rec: Advanced Feature Interactions to Address Covariate-Shifts in Recommendation Networks
Muhammad Adnan, Yassaman Ebrahimzadeh Maboud, Divya Mahajan, Prashant J. Nair
Recommendation Model Covariate Shift Recommendation Task Feature Interaction Deep Recommendation

August 13, 2023

InTune: Reinforcement Learning-based Data Pipeline Optimization for Deep Recommendation Models
Kabir Nagrecha, Lingyi Liu, Pablo Delgado, Prasanna Padmanabhan
Efficient Pipeline Training Bottleneck DL Model Deep Recommendation

July 17, 2023

Evaluating and Enhancing Robustness of Deep Recommendation Systems Against Hardware Errors
Dongning Ma, Xun Jiao, Fred Lin, Mengshi Zhang, Alban Desmaison, Thomas Sellinger, Daniel Moore, Sriram Sankar
Native Robustness Pytorch Model Error Mitigation Hardware Error Robust Recommendation Deep Recommendation

August 12, 2022

AutoShard: Automated Embedding Table Sharding for Recommender Systems
Daochen Zha, Louis Feng, Bhargav Bhushanam, Dhruv Choudhary, Jade Nie, Yuandong Tian, Jay Chae, Yinbin Ma, Arun Kejariwal, Xia Hu
Recommender System Cost Model Table Sharding Deep Recommendation

July 21, 2022

The trade-offs of model size in large recommendation models : A 10000 $\times$ compressed criteo-tb DLRM model (100 GB parameters to mere 10MB)
Aditya Desai, Anshumali Shrivastava
Model Size Machine Learning Benchmark Large Scale Recommender System Deep Learning Recommendation Model Memory Deep Learning Deep Recommendation

March 28, 2022

Learning to Collide: Recommendation System Model Compression with Learned Hash Functions
Benjamin Ghaemmaghami, Mustafa Ozdal, Rakesh Komuravelli, Dmitriy Korchev, Dheevatsa Mudigere, Krishnakumar Nair, Maxim Naumov
LeArning Abstract Model Compression Recommendation Accuracy Hash Function ID Problem Hash Learning Deep Recommendation

February 28, 2022

Filter-enhanced MLP is All You Need for Sequential Recommendation
Kun Zhou, Hui Yu, Wayne Xin Zhao, Ji-Rong Wen
Sequential Recommendation Sequential Recommender System Single Scene Specific MLP Learnable Filter Deep Recommendation

February 24, 2022

BagPipe: Accelerating Deep Recommendation Model Training
Saurabh Agarwal, Chengpo Yan, Ziyi Zhang, Shivaram Venkataraman
Jina Embeddings Recommendation Model Asynchronous Training Deep Learning Recommendation Model Deep Recommendation Non Embedding

Deep Recommendation

Papers

AdaS&S: a One-Shot Supernet Approach for Automatic Embedding Size Search in Deep Recommender System

DQRM: Deep Quantized Recommendation Models

Ad-Rec: Advanced Feature Interactions to Address Covariate-Shifts in Recommendation Networks

InTune: Reinforcement Learning-based Data Pipeline Optimization for Deep Recommendation Models

Evaluating and Enhancing Robustness of Deep Recommendation Systems Against Hardware Errors

AutoShard: Automated Embedding Table Sharding for Recommender Systems

The trade-offs of model size in large recommendation models : A 10000 $\times$ compressed criteo-tb DLRM model (100 GB parameters to mere 10MB)

Learning to Collide: Recommendation System Model Compression with Learned Hash Functions

Filter-enhanced MLP is All You Need for Sequential Recommendation

BagPipe: Accelerating Deep Recommendation Model Training