Elies Gil-Fuster
@eliesgf.bsky.social
Researcher in the making. Loves music and boardgames. 'we didn't come here to pet spiders'.
eliesgilfuster.eu.
eliesgilfuster.eu.
The German saying goes "wer schön sein will muss leiden". My German is not so good, so I understood "wer schön sein will muss *nach* Leiden". Off to spend a couple months @unileiden.bsky.social, to start new projects on #QuantumMachineLearning
October 7, 2025 at 8:13 AM
The German saying goes "wer schön sein will muss leiden". My German is not so good, so I understood "wer schön sein will muss *nach* Leiden". Off to spend a couple months @unileiden.bsky.social, to start new projects on #QuantumMachineLearning
When I woke up this morning I didn't think I would see my name next to that of a #physics giant.
In case you're curious about the paper I presented: arxiv.org/abs/2406.07072
In case you're curious about the paper I presented: arxiv.org/abs/2406.07072
July 24, 2025 at 2:31 PM
When I woke up this morning I didn't think I would see my name next to that of a #physics giant.
In case you're curious about the paper I presented: arxiv.org/abs/2406.07072
In case you're curious about the paper I presented: arxiv.org/abs/2406.07072
We first prove formally that there will be a peak in test error (key signature of double descent) using random matrix theory. We also run several numerical experiments on real-life data (Fashion MNIST and California Housing prices), all confirming the presence of double descent in QML.
January 20, 2025 at 10:56 AM
We first prove formally that there will be a peak in test error (key signature of double descent) using random matrix theory. We also run several numerical experiments on real-life data (Fashion MNIST and California Housing prices), all confirming the presence of double descent in QML.
The question we tackle is then: will QML models show double descent? Truth is that QML hasn't been performing at the level of classical ML models, BUT: if QML models can display double descent, then we might think that great performance is possible in the future.
January 20, 2025 at 10:56 AM
The question we tackle is then: will QML models show double descent? Truth is that QML hasn't been performing at the level of classical ML models, BUT: if QML models can display double descent, then we might think that great performance is possible in the future.
📝 New work out today!
Our next installment on the quest for modern learning theory in Quantum Machine Learning #QML:
"Double Descent in QML"
scirate.com/arxiv/2501.1...
Our next installment on the quest for modern learning theory in Quantum Machine Learning #QML:
"Double Descent in QML"
scirate.com/arxiv/2501.1...
January 20, 2025 at 10:56 AM
📝 New work out today!
Our next installment on the quest for modern learning theory in Quantum Machine Learning #QML:
"Double Descent in QML"
scirate.com/arxiv/2501.1...
Our next installment on the quest for modern learning theory in Quantum Machine Learning #QML:
"Double Descent in QML"
scirate.com/arxiv/2501.1...
New work out today 📝
Very insightful collaboration with colleagues from Fraunhofer HHI and the great @jenseisert.bsky.social. We offer a preview of explainable AI #xAI for #Quantum learning models #QML ⚛️🧠. Check it out and let us know your thoughts!
Very insightful collaboration with colleagues from Fraunhofer HHI and the great @jenseisert.bsky.social. We offer a preview of explainable AI #xAI for #Quantum learning models #QML ⚛️🧠. Check it out and let us know your thoughts!
December 20, 2024 at 10:40 PM
New work out today 📝
Very insightful collaboration with colleagues from Fraunhofer HHI and the great @jenseisert.bsky.social. We offer a preview of explainable AI #xAI for #Quantum learning models #QML ⚛️🧠. Check it out and let us know your thoughts!
Very insightful collaboration with colleagues from Fraunhofer HHI and the great @jenseisert.bsky.social. We offer a preview of explainable AI #xAI for #Quantum learning models #QML ⚛️🧠. Check it out and let us know your thoughts!
Roguelike-themed street art in Melbourne? ✅!
December 5, 2024 at 2:06 AM
Roguelike-themed street art in Melbourne? ✅!
Downtown Sydney is home to a minion watching all of us
December 4, 2024 at 7:54 AM
Downtown Sydney is home to a minion watching all of us
Also in this session was Adrián Pérez-Salinas talking about our recent work about trainability and dequantization in variational #QML.
#QTML2024
arxiv.org/abs/2406.07072
#QTML2024
arxiv.org/abs/2406.07072
November 24, 2024 at 11:49 PM
Also in this session was Adrián Pérez-Salinas talking about our recent work about trainability and dequantization in variational #QML.
#QTML2024
arxiv.org/abs/2406.07072
#QTML2024
arxiv.org/abs/2406.07072
Disclaimers appreciated! It's @mvscerezo.bsky.social 's opening talk at #QTML2024
November 24, 2024 at 9:46 PM
Disclaimers appreciated! It's @mvscerezo.bsky.social 's opening talk at #QTML2024
