✨Toward Ambulatory Vision: Learning Visually-Grounded Active View Selection

📝 Summary:
VG-AVS, a task and framework fine-tunes VLMs to select the most informative next viewpoint for visual question answering, enhancing performance and generalization. AI-generated summary Vision Language...

🔹 Publication Date: Published on Dec 15

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2512.13250
• PDF: https://arxiv.org/pdf/2512.13250
• Project Page: https://active-view-selection.github.io
• Github: https://github.com/KAIST-Visual-AI-Group/VG-AVS

==================================

For more data science resources:
✓ https://news.1rj.ru/str/DataScienceT

#AI #DataScience #MachineLearning #HuggingFace #Research

✨LitePT: Lighter Yet Stronger Point Transformer

📝 Summary:
LitePT combines early convolutions and deep attention for 3D point clouds, using PointROPE positional encoding. This new model is highly efficient, outperforming state-of-the-art while using fewer resources.

🔹 Publication Date: Published on Dec 15

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2512.13689
• PDF: https://arxiv.org/pdf/2512.13689
• Project Page: https://litept.github.io/
• Github: https://github.com/prs-eth/LitePT

🔹 Models citing this paper:
• https://huggingface.co/yuanwenyue/LitePT

==================================

For more data science resources:
✓ https://news.1rj.ru/str/DataScienceT

#AI #DataScience #MachineLearning #HuggingFace #Research

✨VLSA: Vision-Language-Action Models with Plug-and-Play Safety Constraint Layer

📝 Summary:
AEGIS, a Vision-Language-Safe Action architecture with a plug-and-play safety constraint layer using control barrier functions, enhances safety and performance in robotic manipulation tasks. AI-genera...

🔹 Publication Date: Published on Dec 9

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2512.11891
• PDF: https://arxiv.org/pdf/2512.11891
• Github: https://vlsa-aegis.github.io

==================================

For more data science resources:
✓ https://news.1rj.ru/str/DataScienceT

#AI #DataScience #MachineLearning #HuggingFace #Research

✨Towards Interactive Intelligence for Digital Humans

📝 Summary:
Interactive Intelligence, realized through Mio framework, enables advanced digital humans with personality, adaptive interactions, and self-evolution, surpassing current benchmarks. AI-generated summa...

🔹 Publication Date: Published on Dec 15

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2512.13674
• PDF: https://arxiv.org/pdf/2512.13674
• Project Page: https://shandaai.github.io/project_mio_page/

==================================

For more data science resources:
✓ https://news.1rj.ru/str/DataScienceT

#DigitalHumans #InteractiveAI #ArtificialIntelligence #AIResearch #VirtualAgents

✨DiffusionBrowser: Interactive Diffusion Previews via Multi-Branch Decoders

📝 Summary:
DiffusionBrowser is a lightweight decoder for interactive video previews during diffusion model denoising. It enables fast multi-modal previews, enhancing user control and revealing how video details are composed internally.

🔹 Publication Date: Published on Dec 15

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2512.13690
• PDF: https://arxiv.org/pdf/2512.13690
• Github: https://susunghong.github.io/DiffusionBrowser

==================================

For more data science resources:
✓ https://news.1rj.ru/str/DataScienceT

#AI #DataScience #MachineLearning #HuggingFace #Research

✨RecTok: Reconstruction Distillation along Rectified Flow

📝 Summary:
RecTok improves diffusion models by enriching forward flow semantics and enhancing reconstruction, achieving state-of-the-art results with high-dimensional visual tokenizers. AI-generated summary Visu...

🔹 Publication Date: Published on Dec 15

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2512.13421
• PDF: https://arxiv.org/pdf/2512.13421
• Project Page: https://shi-qingyu.github.io/rectok.github.io/
• Github: https://github.com/Shi-qingyu/RecTok

==================================

For more data science resources:
✓ https://news.1rj.ru/str/DataScienceT

#AI #DataScience #MachineLearning #HuggingFace #Research

✨Self-Supervised Prompt Optimization

📝 Summary:
A self-supervised framework optimizes prompts for both closed and open-ended tasks by evaluating LLM outputs without external references, reducing costs and required data. AI-generated summary Well-de...

🔹 Publication Date: Published on Feb 7

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2502.06855
• PDF: https://arxiv.org/pdf/2502.06855
• Github: https://github.com/geekan/metagpt

✨ Spaces citing this paper:
• https://huggingface.co/spaces/XiangJinYu/SPO
• https://huggingface.co/spaces/tang-x/SPO
• https://huggingface.co/spaces/ositamiles/SPO

==================================

For more data science resources:
✓ https://news.1rj.ru/str/DataScienceT

#AI #DataScience #MachineLearning #HuggingFace #Research

✨DeepSeek-V3 Technical Report

📝 Summary:
DeepSeek-V3 is a parameter-efficient Mixture-of-Experts language model using MLA and DeepSeekMoE architectures, achieving high performance with efficient training and minimal computational cost. AI-ge...

🔹 Publication Date: Published on Dec 27, 2024

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2412.19437
• PDF: https://arxiv.org/pdf/2412.19437
• Github: https://github.com/deepseek-ai/deepseek-v3

🔹 Models citing this paper:
• https://huggingface.co/deepseek-ai/DeepSeek-V3
• https://huggingface.co/deepseek-ai/DeepSeek-V3-0324
• https://huggingface.co/deepseek-ai/DeepSeek-V3-Base

✨ Spaces citing this paper:
• https://huggingface.co/spaces/nanotron/ultrascale-playbook
• https://huggingface.co/spaces/weege007/ultrascale-playbook
• https://huggingface.co/spaces/Ki-Seki/ultrascale-playbook-zh-cn

==================================

For more data science resources:
✓ https://news.1rj.ru/str/DataScienceT

#AI #DataScience #MachineLearning #HuggingFace #Research

✨DeepSeek-OCR: Contexts Optical Compression

📝 Summary:
DeepSeek-OCR uses optical 2D mapping to compress long contexts, achieving high OCR precision with reduced vision tokens and demonstrating practical value in document processing. AI-generated summary W...

🔹 Publication Date: Published on Oct 21

🔹 Paper Links:
• arXiv Page: https://arxivexplained.com/papers/deepseek-ocr-contexts-optical-compression
• PDF: https://arxiv.org/pdf/2510.18234
• Github: https://github.com/deepseek-ai/DeepSeek-OCR

🔹 Models citing this paper:
• https://huggingface.co/deepseek-ai/DeepSeek-OCR
• https://huggingface.co/unsloth/DeepSeek-OCR
• https://huggingface.co/Jalea96/DeepSeek-OCR-bnb-4bit-NF4

✨ Spaces citing this paper:
• https://huggingface.co/spaces/merterbak/DeepSeek-OCR-Demo
• https://huggingface.co/spaces/khang119966/DeepSeek-OCR-DEMO
• https://huggingface.co/spaces/prithivMLmods/Super-OCRs-Demo

==================================

For more data science resources:
✓ https://news.1rj.ru/str/DataScienceT

#AI #DataScience #MachineLearning #HuggingFace #Research

✨Multi-module GRPO: Composing Policy Gradients and Prompt Optimization for Language Model Programs

📝 Summary:
mmGRPO, a multi-module extension of GRPO, enhances accuracy in modular AI systems by optimizing LM calls and prompts across various tasks. AI-generated summary Group Relative Policy Optimization ( GRP...

🔹 Publication Date: Published on Aug 6

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2508.04660
• PDF: https://arxiv.org/pdf/2508.04660
• Project Page: https://dspy.ai
• Github: https://github.com/stanfordnlp/dspy

==================================

For more data science resources:
✓ https://news.1rj.ru/str/DataScienceT

#AI #DataScience #MachineLearning #HuggingFace #Research

✨Directional Textual Inversion for Personalized Text-to-Image Generation

📝 Summary:
Directional Textual Inversion DTI enhances text-to-image personalization by fixing learned token magnitudes and optimizing only their direction. This prevents norm inflation issues of standard Textual Inversion, improving prompt conditioning and enabling smooth interpolation. DTI offers better te...

🔹 Publication Date: Published on Dec 15

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2512.13672
• PDF: https://arxiv.org/pdf/2512.13672
• Project Page: https://kunheek.github.io/dti
• Github: https://github.com/kunheek/dti

==================================

For more data science resources:
✓ https://news.1rj.ru/str/DataScienceT

#TextualInversion #TextToImage #GenerativeAI #DeepLearning #AI

✨One-to-All Animation: Alignment-Free Character Animation and Image Pose Transfer

📝 Summary:
One-to-All Animation is a unified framework for high-fidelity character animation and image pose transfer. It tackles misaligned and partially visible references using self-supervised outpainting, a robust reference extractor, and identity-robust pose control to outperform existing methods.

🔹 Publication Date: Published on Nov 28

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2511.22940
• PDF: https://arxiv.org/pdf/2511.22940
• Project Page: https://ssj9596.github.io/one-to-all-animation-project/
• Github: https://github.com/ssj9596/One-to-All-Animation

🔹 Models citing this paper:
• https://huggingface.co/MochunniaN1/One-to-All-14b
• https://huggingface.co/MochunniaN1/One-to-All-1.3b_2
• https://huggingface.co/MochunniaN1/One-to-All-1.3b_1

✨ Datasets citing this paper:
• https://huggingface.co/datasets/MochunniaN1/One-to-All-sub

==================================

For more data science resources:
✓ https://news.1rj.ru/str/DataScienceT

#CharacterAnimation #PoseTransfer #ComputerVision #AI #DeepLearning

✨What matters for Representation Alignment: Global Information or Spatial Structure?

📝 Summary:
Representation alignment enhances generative training by transferring spatial structure from pretrained vision encoders to diffusion models, surpassing the importance of global semantic performance. A...

🔹 Publication Date: Published on Dec 11

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2512.10794
• PDF: https://arxiv.org/pdf/2512.10794
• Project Page: https://end2end-diffusion.github.io/irepa
• Github: https://github.com/end2end-diffusion/irepa

==================================

For more data science resources:
✓ https://news.1rj.ru/str/DataScienceT

#AI #DataScience #MachineLearning #HuggingFace #Research

✨DrivePI: Spatial-aware 4D MLLM for Unified Autonomous Driving Understanding, Perception, Prediction and Planning

📝 Summary:
DrivePI is a new spatial-aware 4D MLLM for autonomous driving, unifying understanding, 3D perception, prediction, and planning. It integrates point clouds, images, and language instructions, achieving state-of-the-art performance by outperforming existing VLA and specialized VA models.

🔹 Publication Date: Published on Dec 14

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2512.12799
• PDF: https://arxiv.org/pdf/2512.12799
• Github: https://github.com/happinesslz/DrivePI

==================================

For more data science resources:
✓ https://news.1rj.ru/str/DataScienceT

#AutonomousDriving #MLLM #ComputerVision #DeepLearning #AI

✨Towards Scalable Pre-training of Visual Tokenizers for Generation

📝 Summary:
Traditional visual tokenizer training fails to improve generation quality with more compute. VTP is a new framework that jointly optimizes image-text contrastive, self-supervised, and reconstruction losses. This enables better scaling, faster convergence, and significantly improved generative per...

🔹 Publication Date: Published on Dec 15

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2512.13687
• PDF: https://arxiv.org/pdf/2512.13687
• Github: https://github.com/hustvl

🔹 Models citing this paper:
• https://huggingface.co/MiniMaxAI/VTP-Base-f16d64
• https://huggingface.co/MiniMaxAI/VTP-Small-f16d64
• https://huggingface.co/MiniMaxAI/VTP-Large-f16d64

==================================

For more data science resources:
✓ https://news.1rj.ru/str/DataScienceT

#AI #DataScience #MachineLearning #HuggingFace #Research

✨Learning Robot Manipulation from Audio World Models

📝 Summary:
A generative latent flow matching model is proposed to predict future audio for robotic manipulation tasks, improving performance over methods without future lookahead by accurately capturing intrinsi...

🔹 Publication Date: Published on Dec 9

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2512.08405
• PDF: https://arxiv.org/pdf/2512.08405

==================================

For more data science resources:
✓ https://news.1rj.ru/str/DataScienceT

#AI #DataScience #MachineLearning #HuggingFace #Research

✨WebOperator: Action-Aware Tree Search for Autonomous Agents in Web Environment

📝 Summary:
WebOperator is a tree-search framework that enhances web agents with reliable backtracking and strategic exploration. It addresses challenges like irreversible actions and partial observability by using a safety-aware search and verifying paths. WebOperator achieves state-of-the-art results on We...

🔹 Publication Date: Published on Dec 14

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2512.12692
• PDF: https://arxiv.org/pdf/2512.12692
• Project Page: https://kagnlp.github.io/WebOperator
• Github: https://kagnlp.github.io/WebOperator

==================================

For more data science resources:
✓ https://news.1rj.ru/str/DataScienceT

#WebAgents #TreeSearch #AI #AutonomousAgents #MachineLearning

✨Towards Visual Re-Identification of Fish using Fine-Grained Classification for Electronic Monitoring in Fisheries

📝 Summary:
A deep learning pipeline was optimized for automated fish re-identification in electronic monitoring systems. Using the Swin-T architecture and AutoFish dataset, it achieved 90.43% Rank-1 accuracy, with intra-species viewpoint differences being the main challenge.

🔹 Publication Date: Published on Dec 9

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2512.08400
• PDF: https://arxiv.org/pdf/2512.08400
• Github: https://github.com/msamdk/Fish_Re_Identification.git

==================================

For more data science resources:
✓ https://news.1rj.ru/str/DataScienceT

#FishReID #DeepLearning #ComputerVision #FisheriesTech #FineGrainedClassification