【2024/12/23】今日の最新AI論文をまとめて紹介

この動画は、2024年12月23日のAI論文を簡単にまとめた内容を紹介しています。

【お知らせ】
Twitter始めました。よかったらフォローしてね😊
  / jiroharapeko  

【論文リスト】
Fietje-An_open,_efficient_LLM_for_Dutch
https://arxiv.org/pdf/2412.15450
Sequence_Matters-Harnessing_Video_Models_in_3D_Super-Resolution
https://arxiv.org/pdf/2412.11525
Multi-LLM_Text_Summarization
https://arxiv.org/pdf/2412.15487
TRecViT-A_Recurrent_Video_Transformer
https://arxiv.org/pdf/2412.14294
MixLLM-LLM_Quantization_with_Global_Mixed-precision_between_Output-features_and_Highly-efficient_System_Design
https://arxiv.org/pdf/2412.14590
Taming_Multimodal_Joint_Training_for_High-Quality_Video-to-Audio_Synthesis
https://arxiv.org/pdf/2412.15322
Offline_Reinforcement_Learning_for_LLM_Multi-Step_Reasoning
https://arxiv.org/pdf/2412.16145
IDOL-Instant_Photorealistic_3D_Human_Creation_from_a_Single_Image
https://arxiv.org/pdf/2412.14963
SCOPE-Optimizing_Key-Value_Cache_Compression_in_Long-context_Generation
https://arxiv.org/pdf/2412.13649
LLMs_Lost_in_Translation-M-ALERT_uncovers_Cross-Linguistic_Safety_Gaps
https://arxiv.org/pdf/2412.15035
Parallelized_Autoregressive_Visual_Generation
https://arxiv.org/pdf/2412.15119
CLEAR-Conv-Like_Linearization_Revs_Pre-Trained_Diffusion_Transformers_Up
https://arxiv.org/pdf/2412.16112



【タイムスタンプ】
00:00 イントロ
00:39 Fietje-An_open,_efficient_LLM_for_Dutch
03:05 Sequence_Matters-Harnessing_Video_Models_in_3D_Super-Resolution
06:00 Multi-LLM_Text_Summarization
08:17 TRecViT-A_Recurrent_Video_Transformer
10:31 MixLLM-LLM_Quantization_with_Global_Mixed-precision_between_Output-features_and_Highly-efficient_System_Design
13:24 Taming_Multimodal_Joint_Training_for_High-Quality_Video-to-Audio_Synthesis
15:57 Offline_Reinforcement_Learning_for_LLM_Multi-Step_Reasoning
18:40 IDOL-Instant_Photorealistic_3D_Human_Creation_from_a_Single_Image
20:57 SCOPE-Optimizing_Key-Value_Cache_Compression_in_Long-context_Generation
23:01 LLMs_Lost_in_Translation-M-ALERT_uncovers_Cross-Linguistic_Safety_Gaps
25:01 Parallelized_Autoregressive_Visual_Generation
27:09 CLEAR-Conv-Like_Linearization_Revs_Pre-Trained_Diffusion_Transformers_Up




#AI論文
#LLM
#生成AI


【注意】
この動画は誤った情報を含む可能性があります。
問題点や改善点があれば、コメントで教えてください。

音声ソフト：VOICEVOX(ずんだもん)