Apple is sponsoring the International Conference on Machine Learning (ICML) 2024, taking place in-person from July 21–27 at the Messe Wien Congress & Exhibition Center in Vienna, Austria. ICML is recognized worldwide for presenting and publishing cutting-edge research on all aspects of machine learning that is used in closely related areas such as artificial intelligence, statistics, and data science, as well as in important application areas such as computer vision, computational biology, speech recognition, and robotics. Below is a schedule of sponsored workshops and events at ICML 2024.
Schedule
Visit the Apple booth in Halle/Hall B, booth #110, from 11:30 a.m. to 6:45 p.m. CEST on July 22; from 10:00 a.m. to 6:00 p.m. CEST on July 23 and 24.
Sunday, July 21
Monday, July 22
- ai – Opens in a new window” class=”icon icon-after icon-external” rel=”noopener nofollow”>Queer in ai
- 9:00 am – 4:00 pm CEST, Pride 2
- Isha Garg and Adam Golinski will represent Apple.
- ai – Opens in a new window” class=”icon icon-after icon-external” rel=”noopener nofollow”>LatinX in ai
- 9:00 am – 4:00 pm CEST, Pride 1
- Pablo Rodriguez Lopez, Eduardo Martinez Montes will be representing Apple.
- ai – Opens in a new window” class=”icon icon-after icon-external” rel=”noopener nofollow”>LatinX in ai
- 5:30 p.m. – 7:30 p.m. CEST, The View Café-Bar Restaurant
- Pablo Rodriguez Lopez, Eduardo Martinez Montes will be representing Apple.
Tuesday, July 23
Wednesday, July 24
- Women in Machine Learning (WiML)
- 9:00 – 16:00 CEST, Schubert 1-6
- Catherine Vilhauer, Pau Rodríguez López and Tillie Hands will represent Apple.
- How fluid is attention?
- 11:30 – 13:00 CEST, Hall/Pavilion C 4-9
- Valérie Castin (ENS), Pierre Ablin, Gabriel Peyré (ENS)
- Bootstrap distillation without data
- 13:30 – 15:00 CEST, Hall/Pavilion C 4-9
- Jiatao Gu, Shuangfei Zhai, Yizhe Zhang, Lingjie Liu (University of Pennsylvania), Josh Susskind
Thursday, July 25
- Executable code actions generate better LLM agents
- 13:30 – 15:00 CEST, Hall/Pavilion C 4-9
- Xingyao Wang (University of Illinois at Urbana-Champaign), Yangyi Chen (University of Illinois at Urbana-Champaign), Lifan Yuan (University of Illinois at Urbana-Champaign), Yizhe Zhang, Hao Peng (University of Illinois at Urbana-Champaign), Ji Heng (University of Illinois at Urbana-Champaign)
Friday, July 26th
- Workshop on foundation models in nature 2024
- 09:00 CEST, Straus 1
- Text-to-Image Diffusion Alignment like GFlowNets
- Dinghuai Zhang (University of Montreal), Yizhe Zhang, Jiatao Gu, Ruixiang Zhang, Josh Susskind, Navdeep Jaitly, Shuangfei Zhai
- Projected language models: a large model presegmented into smaller models
- David Grangier, Angelos Katharopoulos, Pierre Ablin, Awni Hannun
Accepted articles
Contrasting multiple representations and the multimarginal correspondence gap
Zoe Piran, Michal Klein, James Thornton, Marco Cuturi Cameto
Bootstrap distillation without data
Jiatao Gu, Shuangfei Zhai, Yizhe Zhang, Lingjie Liu (University of Pennsylvania), Josh Susskind
On a practical implementation of Brenier's polar factorization theorem and its applications to optimization and sampling
Marco Cuturi Cameto, Nina Vesseron (ENSAE)
On the minimum degree bias in the generalization of the invisible to non-boolean functions
Denys Pushkin (EPFL), Raphael Berthier (EPFL), Emmanuel Abbe (Apple/EPFL)
Scalable pretraining of large autoregressive image models
Alaaeldin Mohamed Elnouby Ali, Michal Klein, Shuangfei Zhai, Miguel Angel Bautista Martin, Josh Susskind, Armand Joulin (Google Deepmind (work done while working at Apple))
Revealing the utilized range of learning subspaces in neural networks
Isha Garg, Eshan Verma, Daniel Ulbricht, Christian Koguchi
Overlay Incitement: Enhancement and Acceleration of Recovery: Increased Generation
Thomas Merth, Qichen Fu, Mohammad Rastegari (Meta (work done at Apple)), Mahyar Najibi
Text-to-Image Diffusion Alignment like GFlowNets
Dinghuai Zhang (University of Montreal), Yizhe Zhang, Jiatao Gu, Ruixiang Zhang, Josh Susskind, Navdeep Jaitly, Shuangfei Zhai
Beware the scalpel: how to improve gradient surgery with an EMA
Pierre Ablin, James Thornton, Eugene Ndiaye, Yu-Guan Hsieh, Michal Klein, Marco Cuturi Cameto
Executable code actions generate better LLM agents
Xingyao Wang (University of Illinois at Urbana-Champaign), Yangyi Chen (University of Illinois at Urbana-Champaign), Lifan Yuan (University of Illinois at Urbana-Champaign), Yizhe Zhang, Hao Peng (University of Illinois at Urbana-Champaign), Ji Heng (University of Illinois at Urbana-Champaign)
How fluid is attention?
Valérie Castin (ENS), Pierre Ablin, Gabriel Peyré (ENS)
Improved modeling of federated datasets using Dirichlet multinomial mixtures
Jonny Scott (Institute of Science and technology Austria), Aine Cahill
Transferring knowledge from Vision Foundation models to efficiently train small, task-specific models
Raviteja Vemulapalli, Hadi Pour Ansari, Fartash Faghri, Sachin Mehta, Mehrdad Farajtabar, Mohammad Rastegari (Meta (Work done while at Apple)), Oncel Tuzel
KV-Runahead: Scalable causal LLM inference via parallel key-value cache generation
Minsik Cho, Mohammad Rastegari (Meta (work done while at Apple)), Devang Naik
Backlash-free optimization on the Stiefel generalized random manifold
Simon Vary (University of Leuven), Pierre Ablin, Bin Gao (Chinese Academy of Sciences), Pierre-Antoine Absil (University of Leuven)
Estimating the mean of private vectors in the Shuffle model: optimal rates require many messages
Hilal Asi, Vitaly Feldman, Jelani Nelson (University of California, Berkeley), Kunal Talwar, Huy Nguyen (Northeastern University), Samson Zhou (Texas A&M University)
Projected language models: a large model presegmented into smaller models
David Grangier, Angelos Katharopoulos, Pierre Ablin, Awni Hannun
Swallowing the bitter pill: simplified and scalable generation of conformers
Yuyang Wang, Ahmed Elhag (University of Oxford), Navdeep Jaitly, Josh Susskind, Miguel Ángel Bautista Martin
Whispering experts: neural interventions to mitigate toxicity in linguistic models
Xavier Suau Cuadros, Pieter Delobelle (KU Leuven), Rin Metcalf Susa, Armand Joulin (Google Deepmind (work done at Apple)), Nick Apostoloff, Luca Zapella, Pau Rodríguez López
Population
mlx
We are demonstrating inference and training of large models on devices with MLX. MLX is a flexible matrix framework that is optimized for Apple silicon and made available by Apple Machine Learning Research. It enables training and inference of models of arbitrary complexity on Apple silicon devices with great brevity and flexibility.
In this demo we demonstrate fine-tuning of a 7B parameter LLM on an iPhone, image generation using a large diffusion model on an iPad, and text generation using multiple large language models on an M2 Ultra Mac Studio and an M3 Macbook Pro.
Private Federated Learning (PFL)
This demo showcases Apple’s Private Federated Learning (PFL) technology. PFL-research is the open-source framework that enables this technology for research simulations, which was released to the market in March 2024 at Apple’s PPML workshop. Here we can show how Siri can play music and podcasts on iPhones, which leverages several technologies such as Siri Signals and Siri Inference, Private Federated Learning (PFL), and Differential Privacy (DP). This highlights a user-facing feature in iOS that was launched thanks to this framework.
Thanks
Ozan Sener and Pau Rodríguez López are the ICML 2024 Area Presidents.
Aadirupa Saha is a co-organizer of the workshop Human Feedback Models for ai Alignment.
Rin Metcalf Susa is a panelist at the Human Feedback Models for ai Alignment workshop.
Marco Cuturi, Samy Bengio and Vladlen Kotlun are senior meta reviewers for ICML 2024.
Arno Blaas, Bailin Wang, Gustaf Ahdritz, Junpei Zhou, Miguel Angel Bautista Martin, Miguel Sarabia del Castillo, Qichen Fu and Ray Zhang are reviewers for ICML 2024.
Natalie Schluter is Senior Chair of the ICML 2024 Workshop.