за фундаментальные открытия и изобретения, обеспечивающие возможность машинного обучения с использованием искусственных нейронных сетей

...
DATASET: Papers that introduce new datasets or make significant modifications to existing ones
DATA: Papers focusing on data processing, cleaning, collection, or curation methodologies
BENCHMARK: Papers proposing or analyzing model evaluation frameworks and benchmarks
AGENTS: Papers exploring autonomous agents, web agents, or agent-based architectures
NLP: Papers advancing natural language processing techniques or applications
CV: Papers developing computer vision methods or visual processing systems
RL: Papers investigating reinforcement learning theory or applications
RLHF: Papers specifically about human feedback in RL (PPO, DPO, etc.)
RAG: Papers advancing retrieval-augmented generation techniques
CODE: Papers about code-related models or programming benchmarks
INFERENCE: Papers optimizing model deployment (quantization, pruning, etc.)
3D: Papers on 3D content generation, processing, or understanding
AUDIO: Papers advancing speech/audio processing or generation
...