Ce contenu n’est disponible qu’en anglais.
Fueling AI innovation
Behind the greatest discoveries lie collaboration and community. That’s why we believe in sharing our latest research on AI, NLP, and information retrieval.
Our research areas
Our mission is to develop AI applied to key business problems while pursuing bold ideas in key areas, from natural language (NLP/NLU) to machine learning operations (MLOps).
NLP/NLU
Personalization
Recommendations
Search
MLOps
Publications
Stay up to date with the latest findings from our research team.
-
Jacopo Tagliabue, Ville Tuulos, Ciro Greco, Valay Dave | 2021 | DCAI @ Neurips
-
You Do Not Need a Bigger Boat: Recommendations at Reasonable Scale in a (Mostly) Serverless and Open Stack
Jacopo Tagliabue | 2021 | RecSys
-
Jacopo Tagliabue, Federico Bianchi, Bingqing Yu | 2021 | Association for Computational Linguistics
-
Patrick John Chia, Bingqing Yu, Jacopo Tagliabue | 2021 | SIGIR eCom
-
Jacopo Tagliabue, Ciro Greco, Jean-Francis Roy, Bingqing Yu, Patrick John Chia, Federico Bianchi, Giovanni Cassani | 2021 | SIGIR eCom
MLOps at Reasonable Scale
Venue: Stanford MLSys | Topic: MLOps
In-Session Personalization: Ranking, Type-Ahead and Query Scoping With Deep Learning
Venue: MICES 2021 | Topics: Personalization, NLP, Search
"Are You Sure?": Blending Product Comparisons and Recommendations With AI
Venue: Berlin Buzzwords 2021 | Topic: Recommendations
From Model Cards to DAG Cards
Venue: Northeastern University | Topic: MLOps
AI From the Trenches: When Technology Meets the Real World
Venue: Northeastern University | Topic: MLOps
When One Model Must Fit All: Challenges and Opportunities in Multi-Tenant Machine Learning
Venue: WSDM 2021 | Topic: MLOps
AI in The Multiverse: Measuring ROI When AB Tests Are Not Possible
Venue: Toronto AI Summit 2021 | Topic: MLOps
The Embeddings That Came in From the Cold
Venue: Data Science Milan | Topic: Recommendations
Few-Shot and Zero-Shot Personalization for Digital Commerce Using Deep Architectures
Venue: Haystack 2020 | Topic: Personalization
The Future Does Not Need Us
Venue: TEDx Como Lake | Topic: NLP
Datasets
Download curated and anonymized datasets we release to help advance AI and democratize research.
-
SIGIR eCOM 2021 Data Challenge Dataset We hosted the 2021 SIGIR eCom Data Challenge. This is the dataset containing utility scripts and information about data preparation and testing for the Challenge.DATASET
-
Cross-shop sessions and pre-trained product embeddings Dataset released with our SIGIR eCom 2020 paper titled “Fantastic Embeddings and How to Align Them: Zero-Shot Inference in a Multi-Shop Scenario”.DATASET
-
Anonymized Clickstream Ecommerce Data Dataset released in conjunction with our paper titled “Shopper Intent Prediction From Clickstream Ecommerce Data with Minimal Browsing Information”.DATASET
