@FraRonchini

PhD student @IsplPolimi - @polimi Former AI Research Intern @BoschResearch Exploring text-based generative models for audio and music applications

 Milan, Italy
Joined December 2020
  • Tweets136
  • Following388
  • Followers170
9Media
Francesca Ronchini retweeted
arXiv Sound@ArxivSound19 Sep 2025
Francisco Messina, Francesca Ronchini, Luca Comanducci, Paolo Bestagini, Fabio Antonacci, "Mitigating data replication in text-to-audio generative diffusion models through anti-memorization guidance," arxiv.org/abs/2509.14934

Mitigating data replication in text-to-audio generative diffusion...

A persistent challenge in generative audio models is data replication, where the model unintentionally generates parts of its training data during inference. In this work, we address this issue in...

arxiv.org
1
390
Francesca Ronchini retweeted
Jannik Sinner
@janniksin24 May 2025
Happy to be back 👋🏻 #RolandGarros
240
738
61
10,378
183,978
Energy consumption of 7 text-to-audio diffusion models was analyzed, identifying the trade-off between audio quality and energy usage at inference time by finding Pareto-optimal solutions.
arXiv Sound@ArxivSound13 May 2025
``Diffused Responsibility: Analyzing the Energy Consumption of Generative Text-to-Audio Diffusion Models,'' Riccardo Passoni, Francesca Ronchini, Luca Comanducci, Romain Serizel, Fabio Antonacci, ift.tt/DNEBwqO
1
2
158
Francesca Ronchini retweeted
arXiv Sound@ArxivSound13 May 2025
``TACOS: Temporally-aligned Audio CaptiOnS for Language-Audio Pretraining,'' Paul Primus, Florian Schmid, Gerhard Widmer, ift.tt/NiCDlUQ

TACOS: Temporally-aligned Audio CaptiOnS for Language-Audio Pretraining

Learning to associate audio with textual descriptions is valuable for a range of tasks, including pretraining, zero-shot classification, audio retrieval, audio captioning, and text-conditioned...

arxiv.org
2
5
625
Francesca Ronchini retweeted
DCASE Challenge@DCASE_Challenge10 Dec 2024
The tasks for DCASE challenge 2025 have been announced. dcase.community/articles/cha… Stay tuned for more details.
1
8
3
20
2,811
Francesca Ronchini retweeted
Yixiao Zhang@Yixiao_Zhang_2 Dec 2024
I find the ismir.net lab information is not always up-to-date, so I decide to run a new COMPLETE list of MIR labs. Features (goals): ⭐️ Research topics ⭐️ PI, active students & alnumi ⭐️ Distinguish active & inactive labs ⭐️ Map view github.com/ldzhangyx/awesome…
3
1
23
1,824
I am so grateful and honored to receive this grant! Thanks for the trust!
IEEE Signal Processing Society@IEEEsps22 Nov 2024
🌟The IEEE Signal Processing Society (SPS) is proud to announce the 2024 class of SPS Scholarship recipients! 🌟Read about their expressed interest and commitment to pursuing signal processing education and real-world career experiences at hubs.la/Q02Zcs7W0.
6
317
It’s time for @milanomusicweek and @linecheck_it! I’ll be hosting a hands-on workshop on gender-friendly music recommendation systems! 🎶Join us to help shape a more #inclusive #music #industry 👉 shorturl.at/D6wfr #DiversityInMusic #InclusiveTech
1
106
I have shared our latest research, "Synthetic Training Set Generation Using #TextToAudio Models for Environmental #Sound Classification" at #DCASE2024 in Tokyo last month! 📝 Article: shorturl.at/0KW55 🔸 Slides: shorturl.at/oK0tP
1
3
22
1,065
In this study, we explore how text-to-audio models can generate #synthetic #datasets to train data-driven sound recognition models. Check out the study, results, and conclusions in the article! If you have questions, feedback, or want to discuss this further, reach out!
122
Francesca Ronchini retweeted
roser batlle roca@roserbatlleroca6 Nov 2024
📢 Call to all #MIRCommunity! Have you ever wondered what an open model means? Help us shape the definition of open models in generative AI for music by taking our survey — just 10 minutes! 👉 forms.gle/Z48t6HPBXwWC3rud9 thank you 💫

Open Generative Models in MIR: What does open model mean?

This survey is part of a research initiative on open models lead by Roser Batlle-Roca, Xavier Serra and Martín Rocamora from the Music Technology Group (MTG) at Universitat Pompeu Fabra (Barcelona).

docs.google.com
1
6
2
12
3,663
Francesca Ronchini retweeted
arXiv Sound@ArxivSound17 Sep 2024
``MambaFoley: Foley Sound Generation using Selective State-Space Models,'' Marco Furio Colombo, Francesca Ronchini, Luca Comanducci, Fabio Antonacci, ift.tt/dh1eEBF

MambaFoley: Foley Sound Generation using Selective State-Space Models

Recent advancements in deep learning have led to widespread use of techniques for audio content generation, notably employing Denoising Diffusion Probabilistic Models (DDPM) across various tasks....

arxiv.org
4
1
9
1,641
Francesca Ronchini retweeted
arXiv Sound@ArxivSound6 Sep 2024
``PAGURI: a user experience study of creative interaction with text-to-music models,'' Francesca Ronchini, Luca Comanducci, Gabriele Perego, Fabio Antonacci, ift.tt/oLctA24

PAGURI: a user experience study of creative interaction with...

In recent years, text-to-music models have been the biggest breakthrough in automatic music generation. While they are unquestionably a showcase of technological progress, it is not clear yet how...

arxiv.org
2
9
1,087
Francesca Ronchini retweeted
Repertorium@Repertorium_AI29 Aug 2024
We’re excited to share that Repertorium project members @FraRonchini, @LucaComan, Fabio Antonacci and @M_Pexx from @polimi are presenting their latest research at the European Signal Processing Conference (EUSIPCO), currently taking place in Lyon! repertorium.eu/dissemination…
3
3
632
Our paper “Room Transfer Function Reconstruction Using #Complex-#Valued #Neural #Networks and Irregularly Distributed Microphones” will be presented at #EUSIPCO2024! I won't be able to attend but DM me if you want to know more! Preprint 👉 shorturl.at/MW8JZv
2
6
318
Francesca Ronchini retweeted
Emmanouil Benetos@emmanouilb7 Aug 2024
📢Job opening: Research Assistant / Postdoctoral Research Assistant in music information research @c4dm @QMEECS on a project funded by @innovateuk Closing date: 31 August 2024 - more info: qmul-jobs.tal.net/vx/mobile-…
13
1
23
2,101
Today I will present part of my research at #Deep #Learning for #MIR workshop series at CCRMA @Stanford 👉 shorturl.at/ssBJY The talk will focus on #opportunities and #challenges of incorporating #text-#to-#music #generative #models into music production practices. (1/n)
1
2
144
Additionally, I will present the results of a #human-#centered #Music #Information #Retrieval (MIR) study aimed at understanding how text-to-music generative models can be practically integrated into the artistic practices of musicians and music professionals. (2/n)
1
1
96
Always happy and open to chat more, so DM me in case you have any questions or like to discuss it further!
1
62
happy to announce that MiRA is out!💫 MiRA (Music Replication Assessment) is an open model-independent tool to assess audio data replication of the training set 🎶🔧🤖👀 installable as a PyPI package 🎁 github.com/roserbatlleroca/m…

GitHub - roserbatlleroca/mira: MiRA (Music Replication Assessment) tool is a model-independent open...

MiRA (Music Replication Assessment) tool is a model-independent open evaluation method based on four diverse audio music similarity metrics to assess exact data replication of the training set. - r...

github.com
1
9
1
42
5,023
Load more