🚨 Deadline for Workshop on Challenges and Opportunities of Large Language Models in Real-World Machine Learning Applications hosted at ECML PKDD is extended to June, 19th🚨 Submit your paper here: collm2023.github.io/Calls/ #research #LLM #NLP #ecmlpkdd2023

Delighted to have two papers accepted at #KDD2023; the 1st one, on optimizing recommendations for long-term user satisfaction, while avoiding long cold-start for new content, still a challenge in recommender systems. Work as part of Tom B @ internship at Spotify Research.

And our second #KDD2023 paper presents our journey on prototyping reinforcement learning based playlist personalization at Spotify.

🚨 Deadline for Workshop on Challenges and Opportunities of Large Language Models in Real-World Machine Learning Applications hosted at ECML PKDD is extended to June, 29th✍️ Submit your paper here: collm2023.github.io/Calls/ #research #LLM #NLP #ecmlpkdd2023

Our first paper proposes an optimisation-based sequencing approach that captures two types of sequential consumption 1) position-aware preferences & 2) local-sequential preferences. Extensive experiments confirm that we can effectively incorporate sequential preferences of users.

Our second paper models the task of query suggestion as a query shortcut link prediction task on a heterogeneous graph including queries & documents, enabling Graph Learning methods to effectively generate query suggestions encompassing both semantic & collaborative information.

That's a wrap! Paper 👉 aclanthology.org/2023.acl-indus… Code 👉 github.com/amazon-science… Part of the work done in my previous internship at Amazon Alexa @ Turin🤗 Joint work w/ Andrea Caciolai Alessandro Pedrani Kay Rottmann Alessandro Manzotti Enrico Palumbo @dawydeb

Come and join us at the #COLLM2023 workshop tomorrow afternoon at ECML PKDD (Polito Aula 2i) to learn and discuss #LLMs with top-notch industry and academia experts! The program features invited talks, paper presentations and a panel session: collm2023.github.io/Program/ 😎

#COLLM2023 it’s a wrap ECML PKDD ! 🎉 It’s been an honor hosting an event with such insightful talks, paper presentations, discussions and questions from the audience 💡 A big thank you to all the speakers, to the conference organizers and to everybody who joined to discuss #LLMS

Graph Learning for Exploratory Query Suggestions in Spotify Search System, by Francesco Fabbri at #CIKM2023

Spotify Research at #CIKM2023, two papers presentations and being in person, feels so good.

Can't find the right words to search the catalog? Our #CIKM2023 describes a graph-based machine learning model for query suggestions that boosts exploratory queries by +9.37%. Find out more on our research blog. research.atspotify.com/2023/10/graph-…

👋 My stint as a #ResearchScientist at #Spotify has ended with recent #layoffs. I was lucky to work with highly talented folks on hard problems in #UX, #algorithms, and #AI. I am back in the research job market with an interest in human-centered AI and #ux research. Please RT.

We are happy to share that our paper "Long-term Off-Policy Evaluation and Learning" has been accepted at #TheWebConference2024. This work stems from the successful internship of @usait0en from Cornell during summer 2023. Congratulations to all authors.

Check out the blog post of our #TheWebConference papers on "Personalizing Audiobooks and Podcasts with graph-based models", describing a big step towards a graph-based, foundation model tailored to the personalization domain, from research to roll-out 🚀 research.atspotify.com/2024/05/person…

ROEGEN at ACM RecSys will features 4 senior researchers including from #meta and #deepmind as keynote speakers. We invite paper on #Risks, #Opportunities, and #Evaluation of #Generative Models in #RecSys. Link: roegen-recsys2024.github.io 📑 Submission Deadline: Aug 30, 2024

Happy to share that "Bridging Search and Recommendation in Generative Retrieval: Does One Task Help the Other?" has been accepted at #RecSys2024. We show that a generative retrieval model leveraging the synergy between search and recommendation outperforms task-specific ones.

Also happy to share that "Encouraging Exploration in Spotify Search through Query Recommendations" has been accepted at #RecSys2024 (Industry Track). We explain how query recommendations were rolled in Spotify's Search, resulting in a 9% increase in number of exploratory queries.

😎 Deadline extended to September, 5th AOE

🚨Hiring alert! My team Spotify Research is looking to hire a research scientist with a machine learning background! Come help us improve personalization experiences by integrating innovative generative AI technologies into our recommender systems. lifeatspotify.com/jobs/research-…