The pre-peer review version of our #ECIR2025 reproducibility work "Are Representation Disentanglement and Interpretability Linked in Recommender Systems? " with Tuukka Ruotsalo, Maria Maistro and Christina Lioma is now on arXiv ( https://t.co/eYCiusGZOG).

Just a thought: why not place top ML conferences (#ICLR, #ICML and #NeurIPS) on the same submission/notification timeline with no multiple submission allowed? Could help reduce the size of these conferences.

I am pleased to share that I will be returning to @AmazonScience this fall as an Applied Scientist intern, in Aachen Germany, focusing on #LLMs and #GenAI!

Discrete sum vs. continuous integral

The use of AI in reviewing is a growing problem. Several of my ICCV papers have AI reviews -- one reviewer was so lazy that they left in the prompts! A common refrain that I hear is that people have difficulty writing in English and need to use AI to clean up their review.

Heading to #ECIR2025! I will be presenting our reproducibility paper “Are Representation Disentanglement and Interpretability Linked in Recommendation Models?”. If you are in Lucca, happy to discuss representation learning and #recsys!

Every academic needs three types of friend: 1) Friend who critiques your work with brutal honesty 2) Friend who drinks with you to recover from brutally honest critiques of your work 3) Friend with no knowledge of, or interest in, your work

The Fourier Transform, explained in one sentence by Stuart Riffle. [ https://t.co/qxTOx44EfS]

Code, data and results are available on Github ( https://t.co/PZeX1CwRxu).

Existing literature claim improved interpretability when learning disentangled representations. In this work, we investigate this link in #recsys by using existing disentanglement metrics and propose adaptions of LIME and SHAP as global representation interpretability measures.

Our work "Are Representation Disentanglement and Interpretability Linked in Recommendation Models? A Critical Review and Reproducibility Study" has been accepted as a reproducibility paper at #ECIR2025! Preprint and code coming soon! #recsys

And with today’s keynote and sessions, #recsys2024 is over. Learned and ate a ton! Bari did not disappoint as did not the #recsys2024 organization! See you at #recsys2025 in Prague!

Let’s stick to Bari for future #recsys conferences! #recsys2024 #day2

Il buongiorno si vede dal caffè! @ACMRecSys

I’m in Bari, Italy for this year’s installment of the #RecSys conference. Looking forward to Italian food and coffee…ah, right and to the conference sessions! Hit me up if you want to talk representation learning in #recsys!

As of this past week I am officially an Applied Scientist intern @AmazonScience in Cambridge, UK! I will continue investigating #recsys and #MachineLearning, this time through the industry perspective. If you are around in Cambridge/London, I’m happy to catch up!

I am pleased to share that I will join @AmazonScience this summer as an Applied Scientist Intern in Cambridge, UK. I look forward to exciting research in #recsys and #machinelearning!

Looking forward to #ICLR2025 in Singapore!

#ICLR2024 6000+ attendees!

With Mr. CNN himself, part deux