Apple Advances the Field of Machine Learning

Apple researchers are at the forefront of machine learning, conducting fundamental studies that not only enhance Apple's ecosystem but also contribute to the broader scientific community.

Joining the NeurIPS Conference

The 38th annual Neural Information Processing Systems (NeurIPS) conference, a major yearly event in ML and AI research, will be held in Vancouver, Canada. Apple is excited to participate and sponsor this key gathering for the community.

Featured Research Works

Apple researchers will showcase a range of ML topics including enhancing privacy, boosting multimodal model capabilities, and exploring large-scale language model (LLM) reasoning abilities.

Advancing Privacy
Apple will present two papers on federated learning, a privacy-preserving technique that enables machine learning without sharing raw data.

Enhancing Multimodal Models
Apple’s work on multimodal models focuses on improving performance despite training data limitations, introducing a new method to refine the learning of CLIP, a vision-language model, with limited annotation data.

Exploring Self-Supervised Learning
Self-supervised learning (SSL) is a method that allows machines to learn without direct human supervision. Apple researchers will discuss how representations are learned differently under two major SSL paradigms: Masked Autoencoders (MAE) and Joint Embedding Predictive Architectures (JEPA).

Supporting the ML Research Community

Apple is committed to supporting underrepresented groups within the ML community, sponsoring several affinity groups organizing events at NeurIPS, including Black in AI, Women in Machine Learning (WiML), LatinX in AI, and Queer in AI. Ultimately, Apple is proud to share its innovative research and connect with the community at this conference.