现已上线！2025 年 Telegram 研究 — 年度关键洞察 
AI with Papers - Artificial Intelligence & Deep Learning
前往频道在 Telegram
All the AI with papers. Every day fresh updates about #DeepLearning #MachineLearning #LLM & #ComputerVision Curated by Alessandro Ferrari | https://www.linkedin.com/in/visionarynet/ #AI #chatGPT
显示更多📈 Telegram 频道 AI with Papers - Artificial Intelligence & Deep Learning 的分析概览
频道 AI with Papers - Artificial Intelligence & Deep Learning (@ai_deeplearning) 英语 语言赛道中的 是活跃参与者。目前社区聚集了 17 142 名订阅者，在 技术与应用 类别中位列第 7 723，并在 马来西亚 地区排名第 2 241 位。
📊 受众指标与增长动态
自 невідомо 创建以来，项目保持高速增长，吸引了 17 142 名订阅者。
根据 23 六月, 2026 的最新数据，频道保持稳定运转。过去 30 天订阅人数变化为 -190，过去 24 小时变化为 -2，整体触达仍然可观。
- 认证状态： 未认证
- 互动率 (ER)： 平均受众互动率为 25.09%。内容发布后 24 小时内通常能获得 6.86% 的反应，占订阅者总量。
- 帖子覆盖： 每篇帖子平均可获得 4 302 次浏览，首日通常累积 1 177 次浏览。
- 互动与反馈： 受众积极参与，单帖平均反应数为 26。
- 主题关注点： 内容集中在 framework, object, dataset, tba, depth 等核心主题上。
📝 描述与内容策略
作者将该频道定位为表达主观观点的平台：
“All the AI with papers. Every day fresh updates about #DeepLearning #MachineLearning #LLM & #ComputerVision
Curated by Alessandro Ferrari | https://www.linkedin.com/in/visionarynet/
#AI #chatGPT”
凭借高频更新（最新数据采集于 24 六月, 2026），频道始终保持新鲜度与高覆盖。分析显示受众积极互动，使其成为 技术与应用 类别中的关键影响点。
17 142
订阅者
-224 小时
-367 天
-19030 天
帖子存档
🔥 EfficientSAM: 20x faster Segment Anything 🔥
👉Meta AI Research unveils a novel family of SAM-like models, light-weight SAM models with SOTA quality-efficiency trade-offs. Up to 20x faster!
👉Review https://t.ly/966QS
👉Paper https://lnkd.in/duijp_Rh
👉Project https://lnkd.in/dW-p2CuH
👉Code https://lnkd.in/dAbZaB2t
👉Demo https://lnkd.in/d-tjKiUd
🩰 Magic Animating Human 🩰
👉MagicAnimate: the new SOTA in human animation. Code available: let's dance!
👉Review https://t.ly/Oq7Za
👉Paper https://lnkd.in/dSUbGgCs
👉Project https://lnkd.in/dkVFf-SV
👉Code https://lnkd.in/dj2dbzdg
👉Demo https://lnkd.in/dHEKPE9q
Hello everybody,
a lot of you asked me to re-open the sharing of the contents to involve more people. I want to follow your suggestion, hope you will enjoy this new mood!
👍 FREE TO FORWARD TO OTHER TELEGRAM CHANNELS
🔥 NO COPY OF THE POSTS
🔥 NO COMMERCIAL USAGE
🔥 NO UNRESPECTFUL USAGE
⚠️ UNDO THE FORWARDING OPTION AT THE FIRST VIOLATION ⚠️
🔎 Generative Powers of Ten 🔍
👉A text-to-image model to generate consistent content across multiple image scales, enabling extreme semantic zooms into a scene. From universe to a human cell 🤯
👉Review https://t.ly/2DG44
👉Paper https://lnkd.in/eDcSpU59
👉Project https://lnkd.in/e6NKu8n9
🍡 Animate Anyone: new SOTA! 🍡
👉Alibaba unveils Animate Anyone: novel #AI for transforming character images into animated videos controlled by desired pose sequences. Animating any character image into a video, unconstrained by specific domains 🚀
👉Review https://t.ly/qCahZ
👉Paper https://lnkd.in/d-zi8EZ6
👉Project https://lnkd.in/djwjQRvq
👉Code https://lnkd.in/dDMkjnKz
👑 HD Generative #AI With No $$$ 👑
👉DemoFusion: a novel approach for HD image generation w/ no money. Progressive Upscaling, Skip Residual, & Dilated Sampling to achieve higher-resolution ever 🔥
👉Review https://t.ly/sIqDV
👉Paper https://lnkd.in/deDt-zcK
👉Project https://lnkd.in/dFGj47Xw
👉Code https://lnkd.in/dY3UcXwp
🧱 Material Palette from Images 🧱
👉A novel problem in #AI: material extraction from a real-world image without any prior knowledge 🤯
👉Discussion https://t.ly/AIWs-
👉Paper https://lnkd.in/dBFAVWPF
👉Project https://lnkd.in/dV5jK8Sm
👉Code https://lnkd.in/dNhMnfFb
👉Dataset (coming) ...
🌳 NebulOS: (more than) Green AI 🌳
👉A novel hardware-aware Training-Free NAS approach that considers both training-free metrics & HW constraints, aiming to find the optimal balance between validation accuracy & energy consumption. 🚀
👉Review https://t.ly/Ozso1
👉Project sites.google.com/view/nebulos
👉Code https://github.com/fracapuano/NebulOS
👉Video https://lnkd.in/exN4Q2Fu
👉Hugging Face demo https://lnkd.in/eyCcPEPc
🎡 Panoptic Video Scene Graph 🎡
👉Combining video scene graph generation w/ panoptic segmentation for holistic video understanding. Novel HQ dataset with fine, temporal scene graph annotations & panoptic segmentation. Code released!🔥
👉Review https://t.ly/tckDT
👉Project jingkang50.github.io/PVSG/
👉Paper arxiv.org/pdf/2311.17058.pdf
👉Code github.com/LilyDaytoy/OpenPVSG
👉Tool github.com/lilyDaytoy/PVSGAnnotation
🔥 Stable (Stability.AI) Video Diffusion 🔥
👉 #StabilityAI released Stable Video Diffusion: latent video diffusion model for high-resolution, SOTA text-to-video and image-to-video generation
👉 Review https://t.ly/XwHys
👉 Code https://lnkd.in/dQw_yNuV
👉 Paper https://lnkd.in/dHn6f787
🦖T-Rex: Counting by Visual Prompting🦖
👉T-Rex: a novel interactive object counting model to detect and count any objects. Impressive results!
👉Review https://t.ly/4SfFX
👉Project https://lnkd.in/dVtEndHv
👉Paper https://lnkd.in/dBGQsbdP
👉Code (not announced, but an empty repo exists): https://lnkd.in/dnZnGRUn
🧿 Model-aware 3D Eye Gaze 🧿
👉 Novel hybrid approach that outputs 3D eye model, semantic segmentation, cam-intrinsic & pose. Only 2D eye semantic segmentation masks and fewer 3D gaze labels for supervision.
👉Review https://t.ly/AdKRf
👉Paper https://lnkd.in/dWb9GHPh
👉Code https://lnkd.in/dfAWFVky
🔳 SOTA Semantic Boundary 🔳
👉Mobile-Seed, a lightweight, dual-task framework tailored for simultaneous semantic segmentation and boundary detection.
👉Review https://t.ly/GsArZ
👉Project whu-usi3dv.github.io/Mobile-Seed/
👉Paper arxiv.org/pdf/2311.12651.pdf
👉Code github.com/WHU-USI3DV/Mobile-Seed
🍿 Segmenting anything in 3D 🍿
👉 OmniSeg3D: omniversal segmentation method aims for segmenting anything in 3D all at once.
👉Review https://t.ly/Q0jrK
👉Paper https://lnkd.in/d9qpxXY9
👉Code (soon)
🌦️ 100+ GPU weather training 🌦️
👉#NVIDIA just released Makani: massively parallel training of weather and climate prediction models on 100+ GPUs and to enable the development of the next generation of weather and climate models.
👉 Discussion https://lnkd.in/dMgakzWm
👉 Project & Code https://lnkd.in/d4NFZ5xi
🐓 Emu: image edit / video gen. 🐓
👉#Meta the new SOTA in text-to-video generation and instruction-based image editing.
👉 Review https://t.ly/PMTBc
👉 Paper (image edit): https://lnkd.in/eVadH-QS
👉 Project https://lnkd.in/eG8eWUJY
👉 Paper (video gen): https://lnkd.in/eVadH-QS
👉 Project https://lnkd.in/eu6Zu6gp
💥🚗 CrashCar101: Generative Damaged Cars💥🚗
👉 CrashCar101: procedural generation pipeline that damages 3D car models to obtain synthetic damaged cars paired with pixel-accurate annotations
👉 Review https://t.ly/pITHm
👉 Paper https://lnkd.in/dzp6q3T5
👉 Project https://lnkd.in/daRXg73N
🔥Florence-2: unified Computer Vision🔥
👉#Microsoft announces Florence-2: novel foundation model with unified, prompt-based, representation for a large variety of #computervision & vision-language task. One backbone -> multiple tasks!
👉Review https://t.ly/pOins
👉Paper arxiv.org/pdf/2311.06242.pdf
👉Project www.microsoft.com/en-us/research/project/projectflorence/
