Research
International Conference on Learning Representations (ICLR) 2026
ICLR 2026 is a major international conference on machine learning and deep learning where Apple's research team will present their latest work and innovations in representation learning. The conferenc
ICLR 2026 is a major international conference on machine learning and deep learning where Apple's research team will present their latest work and innovations in representation learning. The conference brings together researchers and practitioners from academia and industry to share advances in learning algorithms, neural networks, and related AI techniques. Apple's participation highlights the company's ongoing contributions to fundamental machine learning research and its engagement with the broader scientific community.
Related
- Dual Mamba for Node-Specific Representation Learning: Tackling Over-Smoothing with Selective State Space Modeling
- MixAtlas: Uncertainty-aware Data Mixture Optimization for Multimodal LLM Midtraining
- A Minimal Model of Representation Collapse: Frustration, Stop-Gradient, and Dynamics
- Random Matrix Theory for Deep Learning: Beyond Eigenvalues of Linear Models
- Silhouette Loss: Differentiable Global Structure Learning for Deep Representations
Source: Apple ML Research | 2026-04-17