Danilo Comminiello
dacom.bsky.social
Danilo Comminiello
@dacom.bsky.social
Odd time math and signatures | Associate Professor of Machine Learning for Signal Processing at Sapienza University of Rome
Pinned
𝗜𝗝𝗖𝗡𝗡 𝟮𝟬𝟮𝟱 in Rome, June 30 to July 5, 2025 🧠
👉https://2025.ijcnn.org 👈

We are very glad to invite you all to contribute to the International Joint Conference on Neural Networks #IJCNN2025

Looking forward to welcoming you to IJCNN 2025 in Rome!

#INNS #NeuralNetworks #AI #MachineLearning
#AIConf
Yet another pre-Christmas release!! 🎅🎄
Here is 𝐒𝐭𝐚𝐛𝐥𝐞-𝐕𝟐𝐀 which generates sound effects from silent video frames showing semantic and temporal alignment.
🎶🥁🎛️

Huge thanks to @riccardofosco.bsky.social Christian Marinoni and all co-authors 🤟
December 23, 2024 at 2:01 PM
We released 𝐆𝐑𝐀𝐌, a method for learning and aligning any 𝘯 modalities simultaneously! 🎬🎤🎼📽️
GRAM may pave the way for several ML tasks in different fields of application.

Project page👉https://ispamm.github.io/GRAM/

Immensely proud of Giordano Cicchetti @eleonoragrassucci.bsky.social Luigi Sigillo
December 19, 2024 at 12:21 PM
Reposted by Danilo Comminiello
The link between diffusion models and optimal transport is still a bit of an enigma to me.

One thing that's clear: different diffusion models trained on similar datasets tend to recover similar mappings. If these are generally not OT, in what sense are they optimal instead?
I wrote a summary of the main ingredients of the neat proof by Hugo Lavenant that diffusion models do not generally define optimal transport. github.com/mathematical...
November 30, 2024 at 12:56 PM
𝗜𝗝𝗖𝗡𝗡 𝟮𝟬𝟮𝟱 in Rome, June 30 to July 5, 2025 🧠
👉https://2025.ijcnn.org 👈

We are very glad to invite you all to contribute to the International Joint Conference on Neural Networks #IJCNN2025

Looking forward to welcoming you to IJCNN 2025 in Rome!

#INNS #NeuralNetworks #AI #MachineLearning
#AIConf
November 18, 2024 at 7:29 PM