We're excited to be at ICLR2024 this week. With more than 7,400 submissions and an acceptance rate of just over 30%, we're thrilled to announce that Sony AI has not one, but 10 papers accepted. Here’s a closer look at what makes each one stand out in the fiercely competitive field of AI research: 1. Consistency Trajectory Models: This paper introduces a novel approach to understanding probability flow in diffusion models, marking a significant step forward in generative model research. 2. Manifold Preserving Guided Diffusion: A testament to our team's innovation, this research showcases state-of-the-art (SOTA) achievements, enhancing the robustness and quality of generated content. 3. SAN: Inducing Metrizability of GAN with Discriminative Normalized Linear Layer: By redefining the structural foundations of GANs, this paper opens new avenues for generative models, ensuring greater stability and performance. 4. Detecting, Explaining, and Mitigating Memorization in Diffusion Models: Selected for an oral presentation, this work addresses the critical issue of data memorization in AI models, proposing novel solutions for enhancing data privacy. 5. FedWon: Triumphing Multi-domain Federated Learning Without Normalization: Explores advancements in federated learning, paving the way for more inclusive and diversified AI training environments. 6. FedP3: Federated Personalized and Privacy-friendly Network Pruning under Model Heterogeneity: This paper focuses on making federated learning more accessible and efficient, demonstrating Sony AI's leadership in privacy-aware AI development. 7. Understanding multimodal contrastive learning through pointwise mutual information: Integrating different modalities for real-time applications. 8. DIAGNOSIS: Detecting unauthorized data usages in text-to-image diffusion models. 9. Views Can Be Deceiving: Improved SSL Through Feature Space Augmentation: This paper and workshop delve into the intricacies of Self-Supervised Learning (SSL), demonstrating how common augmentation techniques can unintentionally emphasize spurious features. 10. Towards Principled Representation Learning from Videos for Reinforcement Learning: Sets new benchmarks for model performance in dynamic environments. Workshops Led by the AI Ethics Team: * ICLR 2024 Workshop on Data-centric Machine Learning Research (DMLR): Spearheaded by Jerone Andrews, focusing on the importance of high-quality training data, robust data management, and ethical considerations. * Navigating and Addressing Data Problems for Foundation Models (DPFM): Explores how curated training data can address critical issues such as ethics, privacy, and security. We're excited to contribute to the global AI community and look forward to engaging discussions on these topics. Let us know if you’ll be at ICLR 2024 in the comments and what you are most looking forward to! https://bit.ly/44z2k74
Moncef Aguejdad
2mo
Congratulations Jerone Andrews!
Senior Manager at DOCOMO/NTT-COM.
2mocongratulation! Jerone Andrews, Lingjuan LYU