(Include but not limited to)
Track 1: Large Models
Architecture and training methods of large language models (LLMs)
Efficient training and parallel training systems for large models (data parallelism, pipeline parallelism, tensor parallelism)
Distributed training and optimization of heterogeneous cluster scheduling
Fine-tuning of large models (full-scale fine-tuning, efficient parameter fine-tuning with LoRA/Adapter)
Accelerating inference and compressing models (pruning, quantization, distillation)
Long context modeling and optimization of attention mechanisms
Security, alignment, and interpretability of large models
Multimodal large models and cross-modal learning
Generative AI and diffusion models
Continuous learning and domain adaptation of large models
Evaluation, probe analysis, and behavior understanding of large models
Optimization of model parallelism and mixed parallelism strategies
Communication optimization and load balancing in large model training
Integration of reinforcement learning and large models
Track 2: Distributed Computing
Architecture and design of distributed machine learning systems
Heterogeneous computing and hardware accelerators (GPU, TPU, NPU, etc.)
High-performance computing and parallel computing frameworks
Distributed training communication optimization (parameter synchronization, gradient compression)
Federated learning and privacy-preserving distributed training
Collaborative architecture of cloud computing and edge computing
Distributed storage systems and memory pooling technology
Resource scheduling, load balancing, and fault-tolerant mechanisms
Distributed data processing and big data platforms
Computing within networks and intelligent network card acceleration
Service oblivious computing and cloud-native platforms
Blockchain and distributed trust mechanisms
Performance evaluation and bottleneck analysis of distributed systems
Cyber-physical systems and IoT computing infrastructure
Distributed databases and data management systems
Track 3: Machine Learning
Deep learning theory and algorithm innovation
Graph neural networks and complex network analysis
Self-supervised learning, contrastive learning, and meta-learning
Reinforcement learning and deep reinforcement learning
Generative models (VAE, GAN, diffusion models)
Multimodal fusion and cross-modal representation learning
Interpretable artificial intelligence and trustworthy machine learning
Small sample learning and transfer learning
Neural architecture search (NAS) and AutoML
Adversarial learning and robust machine learning
Applications of machine learning in scientific research and engineering
Data science, knowledge discovery, and data mining
Multi-agent learning and collaborative intelligence
Edge intelligence and lightweight model deployment at the edge
Data quality, data governance, and data engineering
Cross-disciplinary intersection of machine learning and quantum computing ......