To be held in Milan from 20^th May 2024 to 24^th May 2024

Welcome! The purpose of this workshop is to exchange ideas between AI, cryptography, average-case complexity, and statistical physics. We aim to bring together experts in these areas in order to share recent research and develop new connections between these important fields. We have scheduled a diverse set of talks ranging from learning dynamics of neural networks, to evidence of hardness for random optimization problems, to cryptographic applications of hard computational tasks. Check out the schedule below.

All talks with be streamed and can be accessed using Zoom on the link below:

• https://unibocconi-it.zoom.us/j/97858207240

Please note that the stream will be recorded.

All times are CEST (Central European Summer Time). Program subject to change.
Click on ▶ Abstract to see the abstract for the corresponding talk (if available).
To see all days of the workshop, you may have to scroll left-to-right ↔.

Program

	Monday 20th May	Tuesday 21th May	Wednesday 22th May	Thursday 23th May	Friday 24th May
9:30 - 11:00	Giulio Biroli ENS Paris Title: Dynamical Regimes of Diffusion Models arXiv ▶  Abstract	Giulio Malavolta Bocconi University Title: Permutation Superposition Oracles for Quantum Query Lower Bounds ▶  Abstract	Vincent Cohen-Addad Google Research Title: Sensitivity Sampling for Coreset-Based Data Selection ▶  Abstract	Simona Samardjiska Radboud University Title: Theoretical and Practical Hardness of Some Equivalence Problems ▶  Abstract	Damiano Abram Aarhus University Title: Cryptography from Planted Graphs: Security with Logarithmic-Size Messages eprint ▶  Abstract
				Shuichi Hirahara National Institute of Informatics Title: One-Way Functions and Zero Knowledge eccc ▶  Abstract	Ran Canetti Boston University Title: Towards General-Purpose Program Obfuscation via Local Mixing eprint ▶  Abstract
	Cengiz Pehlevan Harvard University Title: A Dynamical Model of Neural Scaling Laws arXiv ▶  Abstract	Keisuke Fujii Osaka University Title: Quantum Machine Learning: Interplay between Implicit and Explicit Quantum Models pdf ▶  Abstract	Guilhem Semerjian LPT-ENS Title: Matrix Denoising via Low-Degree Polynomials arXiv ▶  Abstract
				Rajendra Kumar IIT Delhi Title: Fine-Grained Hardness of Lattice Problems ▶  Abstract	Miranda Christ Columbia University Title: Pseudorandom Error-Correcting Codes eprint ▶  Abstract
11:00 - 11:30	☕ Coffee break
11:30 - 13:00	Matthieu Wyart EPFL Title: Learning and Generating Compositional Data ▶  Abstract	Giuseppe Carleo EPFL Title: Neural Simulation of Quantum Many-Body Physics ▶  Abstract	Marc Lelarge INRIA and ENS Title: Chaining Graph Neural Networks for Combinatorial Optimization: Learning Graph Matching ▶  Abstract	Yuval Ishai Technion - Israel Institute of Technology Title: Low-Complexity Cryptography ▶  Abstract	Andrej Bogdanov University of Ottawa Title: Hardness Versus Randomness for Statistical Inference ▶  Abstract
				Pravesh K. Kothari Princeton University Title: Clustering of Large Independent Sets in Expanders and Algorithmic Applications ▶  Abstract	Krzysztof Pietrzak ISTA Title: Longest-Chain Blockchains from Work, Space and Speed: An Exact Classification ▶  Abstract
	Florent Krzakala EPFL Title: How Do Two-Layer Neural Nets Learn Multi-Index Functions? ▶  Abstract	Michael Walter Ruhr University Bochum Title: (Hidden) Symmetries of Computational Problems arXiv ▶  Abstract	Omri Shmueli Tel Aviv University, Weizmann University Title: Quantum Algorithms in a Superposition of Spacetimes eccc ▶  Abstract
				Ilias Zadik Yale University Title: Counting Stars is Constant-Degree Optimal For Detecting Any Planted Subgraph arXiv ▶  Abstract
13:00 - 14:30	🍝 Lunch
14:30 - 16:00	Zohar Ringel The Hebrew University of Jerusalem Title: Some Advancements in GP Theory with Applications to Language Models ▶  Abstract	Stephane Mallat ENS Paris Title: Approximation and Generalisation by Score Diffusion and Renormalisation Group ▶  Abstract	Nobutaka Shimizu Tokyo Institute of Technology Title: Hardness Amplification for Planted Clique Problem eccc ▶  Abstract	Tim Kunisky Yale University Title: Low Coordinate Degree Algorithms: A Flexible Framework for Computational Thresholds arXiv ▶  Abstract
			Vinod Vaikuntanathan Massachusetts Institute of Technology Title: Theoretical Foundations of Real-World Cryptography ▶  Abstract	Afonso S. Bandeira ETH Zürich Title: Spike Models, Free Probability, and the Kikuchi method for Tensor PCA ▶  Abstract
	Song Mei UC Berkeley Title: Revisiting Neural Network Approximation Theory in the Age of Generative AI ▶  Abstract	Johanni Brea EPFL Title: Are Modern Hopfield Networks Good Memory Models? ▶  Abstract
			Federico Ricci-Tersenghi Sapienza Università di Roma Title: Dynamical Phase Transitions, Algorithmic Thresholds, and Possible One-Way Functions ▶  Abstract	Prashant Nalini Vasudevan National University of Singapore Title: Average-Case Complexity of Problems from Fine-Grained Complexity ▶  Abstract
16:00 - 16:30	☕ Coffee break
16:30 - 18:00	Yue M. Lu Harvard University Title: Nonlinear Random Matrices in Machine Learning: Random Features, Kernels, and Self-Attentions ▶  Abstract	Moritz Helias Jülich Research Centre, RWTH Aachen Title: Critical Feature Learning in Deep Neural Networks ▶  Abstract	Chris Jones Bocconi University Title: Diagram Analysis of Iterative Algorithms arXiv ▶  Abstract	🍲 Collaboration Time
			Eren C. Kızıldağ Columbia University Title: Statistical-Computational Tradeoffs in Random Optimization Problems ▶  Abstract
	Marc Mézard Bocconi University Title: Discussion on the New Journal of Artificial Intelligence Theory ▶  Abstract	SueYeon Chung NYU, Flatiron Institute Title: Computing with Neural Manifolds: Towards a Multi-Scale Understanding of Biological and Artificial Neural Networks ▶  Abstract
18:00 - 20:00	🍲 Discussion and Aperitivo
20:00 - 22:00			🍲 Dinner for invited speakers

Attending

The workshop is located in the Roentgen building at Bocconi University located at:
📍 Via Guglielmo Röntgen 1, 20136 Milano, MI.
Tram line 9 and busses 59, 71, and 79 stop on the street right to it.

The workshop will take place in Aula AS01 which can be found down the large stairs right next to the GūD coffee shop.

Registration for the workshop is free but mandatory. Please use the link below.
📝 Click here to register for the workshop.

Recommended hotels:

• Hotel d'Este (special price available)
• UNAHOTELS Mediterraneo Milano
• Uptown Palace
• Grand Visconti Palace

The workshop is being co-organized under the Physics and ML research pillar of the European Laboratory for Learning and Intelligent Systems (ELLIS) and the CIFRA Institute. The workshop organizers are Nicolas Brunel, Vedran Dunjko, Chris Jones, Bert Kappen, Carlo Lucibello, Enrico Malatesta, Marc Mézard, Tamer Mour, Gabriele Perugini, Alon Rosen, Luca Saglietti, Nikolaj I. Schwartzbach, Sofia Vallecorsa, Roberta Zambrini, Riccardo Zecchina.