FO
Francesco Orabona
cs.LGstat.MLmath.OCcs.AIcs.CLcs.ITmath.ITmath.PRmath.STstat.TH
On Valency
published · living versionsW_x7h4jshv·v1 · currentpublished
Conic-Lifting for Learning-rate-free Convex Optimization
with Julien Grand-Clément, Joon Kwon
1 version
W_nxvpg5ra·v2published
W_xbx4f9am·v1published
Preprints & journals
81 papers in the corpus · 2006–2026Optimal Priors for Betting Tests with Bounded Payoffs: An Exact Expected-Time Characterization2610.25821v2 · Francesco Orabona2026 · 0 citationsarXivon Valency
Monitoring for Shifts at Unknown Times: Parameter-Free Conditional Conformal Martingales2610.17343v1 · Tuo Liu, Drew Prinster, Francesco Orabona2026 · 0 citationsarXiv
\k{appa}-LoRA: Condition Numbers Reveal Which LoRA Matrices Worth Updating2607.22489v1 · Jianghui Wang, Silong Yong, Francesco Orabona et al.2026 · 0 citationsarXiv
A Single Stepsize Suffices for Unprojected Linear TD(0): Simultaneous Robust and Fast Rates via Polyak--Ruppert Averaging2606.24981v1 · Wei-Cheng Lee, Francesco Orabona2026 · 0 citationsarXiv
Last-Iterate Convergence of Optimistic Multiplicative Weight Update2606.11773v1 · Francesco Orabona2026 · 0 citationsarXiv
From Betting to Empirical Bernstein LIL2605.22124v1 · Francesco Orabona2026 · 0 citationsarXiv
A Note on How to Remove the Term from the Squint Bound2604.26926v1 · Francesco Orabona2026 · 0 citationsarXiv
Nonparametric Teaching of Attention Learners2602.20461v1 · Chen Zhang, Jianghui Wang, Bingyang Cheng et al.2026 · 0 citationsarXiv
Online Conformal Prediction via Universal Portfolio Algorithms2602.03168v1 · Tuo Liu, Edgar Dobriban, Francesco Orabona2026 · 0 citationsarXiv
Mix, MinHash, and Match: Cross-Source Agreement for Multilingual Pretraining Datasets2512.18834v2 · Sultan Alrashed, Francesco Orabona2025 · 0 citationsarXiv
SmolKalam: Ensemble Quality-Filtered Translation at Scale for High Quality Arabic Post-Training Data2511.18411v1 · Sultan Alrashed, Chadi Helwe, Francesco Orabona2025 · 0 citationsarXiv
A Best-of-Both-Worlds Proof for Tsallis-INF without Fenchel Conjugates2511.11211v1 · Wei-Cheng Lee, Francesco Orabona2025 · 0 citationsarXiv
Global PIQA: Evaluating Commonsense Reasoning Across 100+ Languages and Cultures2510.24081v2 · Tyler A. Chang, Catherine Arnett, Abdelrahman Sadallah et al.2025 · 0 citationsarXiv
Beyond the Ideal: Analyzing the Inexact Muon Update2510.19933v1 · Egor Shulgin, Sultan AlRashed, Francesco Orabona et al.2025 · 0 citationsarXiv
Offline and Online KL-Regularized RLHF under Differential Privacy2510.13512v1 · Yulian Wu, Rushil Thareja, Praneeth Vepakomma et al.2025 · 0 citationsarXiv
Dynamic Regret Reduces to Kernelized Static Regret2507.05478v2 · Andrew Jacobsen, Alessandro Rudi, Francesco Orabona et al.2025 · 0 citationsarXiv
Position: Machine Learning Conferences Should Establish a "Refutations and Critiques" Track2506.19882v3 · Rylan Schaeffer, Joshua Kazdan, Yegor Denisov-Blanch et al.2025 · 0 citationsarXiv
A Robust Rate for Unprojected TD Learning with Linear Function Approximation2506.01052v3 · Wei-Cheng Lee, Francesco Orabona2025 · 0 citationsarXiv
STaR-Bets: Sequential Target-Recalculating Bets for Tighter Confidence Intervals2505.22422v2 · V'aclav Vor'acek, Francesco Orabona2025 · 0 citationsarXiv
Dual Averaging Converges for Nonconvex Smooth Stochastic Optimization2505.21394v1 · Tuo Liu, El Mehdi Saad, Wojciech Kot\lowski et al.2025 · 0 citationsarXiv
SquarePO: Differentially Private and Robust -Preference Optimization in Offline Direct Alignment2505.21395v1 · Xingyu Zhou, Yulian Wu, Wenqian Weng et al.2025 · 0 citationsarXiv
New Perspectives on the Polyak Stepsize: Surrogate Functions and Negative Results2505.20219v2 · Francesco Orabona, Ryan D'Orazio2025 · 0 citationsarXiv
A Unified Theoretical Analysis of Private and Robust Offline Alignment: from RLHF to DPO2505.15694v1 · Xingyu Zhou, Yulian Wu, Francesco Orabona2025 · 0 citationsarXiv
Optimal Regret of Bernoulli Bandits under Global Differential Privacy2505.05613v1 · Achraf Azize, Yulian Wu, Junya Honda et al.2025 · 0 citationsarXiv
New Lower Bounds for Stochastic Non-Convex Optimization through Divergence Decomposition2502.14060v2 · El Mehdi Saad, Wei-Cheng Lee, Francesco Orabona2025 · 0 citationsarXiv
ATA: Adaptive Task Allocation for Efficient Resource Management in Distributed Machine Learning2502.00775v2 · Artavazd Maranjyan, El Mehdi Saad, Peter Richt'arik et al.2025 · 0 citationsarXiv
Self-Directed Learning of Convex Labelings on Graphs2409.01428v2 · Georgy Sokolov, Maximilian Thiessen, Margarita Akhmejanova et al.2024 · 0 citationsarXiv
An Equivalence Between Static and Dynamic Regret Minimization2406.01577v2 · Andrew Jacobsen, Francesco Orabona2024 · 0 citationsarXiv
Better-than-KL PAC-Bayes Bounds2402.09201v2 · Ilja Kuzborskij, Kwang-Sung Jun, Yulian Wu et al.2024 · 0 citationsarXiv
Towards Training Without Depth Limits: Batch Normalization Without Gradient Explosion2310.02012v1 · Alexandru Meterez, Amir Joudaki, Francesco Orabona et al.2023 · 0 citationsarXiv
Normalized Gradients for All2308.05621v1 · Francesco Orabona2023 · 0 citationsarXiv
Implicit Interpretation of Importance Weight Aware Updates2307.11955v1 · Keyi Chen, Francesco Orabona2023 · 0 citationsarXiv
Generalized Implicit Follow-The-Regularized-Leader2306.00201v1 · Keyi Chen, Francesco Orabona2023 · 1 citationarXiv
Tighter PAC-Bayes Bounds Through Coin-Betting2302.05829v1 · Kyoungseok Jang, Kwang-Sung Jun, Ilja Kuzborskij et al.2023 · 2 citationsarXiv
Optimal Stochastic Non-smooth Non-convex Optimization through Online-to-Non-convex Conversion2302.03775v3 · Ashok Cutkosky, Harsh Mehta, Francesco Orabona2023 · 4 citationsarXiv
Robustness to Unbounded Smoothness of Generalized SignSGD2208.11195v1 · Michael Crawshaw, Mingrui Liu, Francesco Orabona et al.2022 · 11 citationsarXiv
Implicit Parameter-free Online Learning with Truncated Linear Models2203.10327v1 · Keyi Chen, Ashok Cutkosky, Francesco Orabona2022 · 1 citationarXiv
Understanding AdamW through Proximal Methods and Scale-Freeness2202.00089v1 · Zhenxun Zhuang, Mingrui Liu, Ashok Cutkosky et al.2022 · 49 citationsarXiv
Tight Concentrations and Confidence Sequences from the Regret of Universal Portfolio2110.14099v3 · Francesco Orabona, Kwang-Sung Jun2021 · 15 citationsarXiv
Minimax Optimal Quantile and Semi-Adversarial Regret via Root-Logarithmic Regularizers2110.14804v2 · Jeffrey Negrea, Blair Bilodeau, Nicolo Campolongo et al.2021 · 0 citationsNeurIPS 2021
Online Learning with Optimism and Delay2106.06885v4 · Genevieve Flaspohler, Francesco Orabona, Judah Cohen et al.2021 · 6 citationsarXiv
On the Initialization for Convex-Concave Min-max Problems2103.00284v2 · Mingrui Liu, Francesco Orabona2021 · 0 citationsarXiv
A closer look at temporal variability in dynamic online learning2102.07666v1 · Nicolo Campolongo, Francesco Orabona2021 · 3 citationsarXiv
On the Last Iterate Convergence of Momentum Methods2102.07002v3 · Xiaoyu Li, Mingrui Liu, Francesco Orabona2021 · 2 citationsarXiv
Parameter-free Stochastic Optimization of Variationally Coherent Functions2102.00236v1 · Francesco Orabona, D'avid P'al2021 · 1 citationarXiv
Adam: A Stochastic Method with Adaptive Variance Reduction2011.11985v1 · Mingrui Liu, Wei Zhang, Francesco Orabona et al.2020 · 12 citationsarXiv
A High Probability Analysis of Adaptive SGD with Momentum2007.14294v1 · Xiaoyu Li, Francesco Orabona2020 · 13 citationsarXiv
Temporal Variability in Implicit Online Learning2006.07503v2 · Nicolo Campolongo, Francesco Orabona2020 · 10 citationsarXiv
Better Parameter-free Stochastic Optimization with ODE Updates for Coin-Betting2006.07507v3 · Keyi Chen, John Langford, Francesco Orabona2020 · 3 citationsarXiv
A Second look at Exponential and Cosine Step Sizes: Simplicity, Adaptivity, and Performance2002.05273v4 · Xiaoyu Li, Zhenxun Zhuang, Francesco Orabona2020 · 17 citationsarXiv
Career total: 191 works. 81 are in this corpus.Showing the 50 most recent.
Profile built from the corpus for this byline.
Author records are still filling in while the Hub is in alpha. If this is your page, you'll be able to claim it soon. Spot a mistake? Tell us.