2024/12/16~2024/12/20 面白そうな論文リスト
Byte Latent Transformer: Patches Scale Better Than Tokens
https://arxiv.org/abs/2412.09871
いい感じのバイトにエンコード・デコードすることで言語モデルの負担を減らして性能を上げる感じ？
MaxInfoRL: Boosting exploration in reinforcement learning through information gain maximization
https://arxiv.org/abs/2412.12098
最大情報量に基づく探索の強化学習
AlphaZero Neural Scaling and Zipf's Law: a Tale of Board Games and Power Laws
https://arxiv.org/abs/2412.11979
AlphaZeroにおけるZipf's Law
Policy Decorator: Model-Agnostic Online Refinement for Large Policy Model
https://arxiv.org/abs/2412.13630
Offline模倣学習からOnlineRLへ
Autoregressive Video Generation without Vector Quantization
https://arxiv.org/abs/2412.14169
離散化しないで自己回帰的に生成（VAEは使っている）
マルチトークンで一個分みたいな？
高速らしいけどなんで？
Latent Reward: LLM-Empowered Credit Assignment in Episodic Reinforcement Learning
https://arxiv.org/abs/2412.11120
LLMでRLを強化する系
Scaling 4D Representations
https://arxiv.org/abs/2412.15212
自動運転系
GaussianWorld: Gaussian World Model for Streaming 3D Occupancy Prediction
https://arxiv.org/abs/2412.10373
3DGSベースの世界モデル
GaussianAD: Gaussian-Centric End-to-End Autonomous Driving
https://arxiv.org/abs/2412.10371
3DGSベースの表現
WiseAD: Knowledge Augmented End-to-End Autonomous Driving with Vision-Language Model
https://arxiv.org/abs/2412.09951
VLMを用いたAD
Bench2Drive-R: Turning Real World Data into Reactive Closed-Loop Autonomous Driving Benchmark by Generative Model
https://arxiv.org/abs/2412.09647
拡散モデルを組み合わせてシミュレータ作り
GEM: A Generalizable Ego-Vision Multimodal World Model for Fine-Grained Ego-Motion, Object Dynamics, and Scene Composition Control
https://arxiv.org/abs/2412.11198
OpenEMMA: Open-Source Multimodal Model for End-to-End Autonomous Driving
https://arxiv.org/abs/2412.15208
VLM-AD: End-to-End Autonomous Driving through Vision-Language Model Supervision
https://arxiv.org/abs/2412.14446
DriveGPT: Scaling Autoregressive Behavior Models for Driving
https://arxiv.org/abs/2412.14415