✨Query-focused and Memory-aware Reranker for Long Context Processing

📝 Summary:
This reranking framework uses attention scores from selected LLM heads to estimate passage-query relevance. It's lightweight, achieves strong performance, and outperforms state-of-the-art rerankers across various domains, including long narrative datasets and the LoCoMo benchmark.

🔹 Publication Date: Published on Feb 12

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2602.12192
• PDF: https://arxiv.org/pdf/2602.12192
• Project Page: https://qdcassie-li.github.io/QRRanker/

🔹 Models citing this paper:
• https://huggingface.co/MindscapeRAG/QRRanker

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

For more data science resources:
✓ https://t.iss.one/DataScienceT

#Reranking #LLM #NLP #InformationRetrieval #LongContext

✨QuantVLA: Scale-Calibrated Post-Training Quantization for Vision-Language-Action Models

📝 Summary:
QuantVLA is a training-free post-training quantization framework for vision-language-action models. Through scale-calibrated components, it significantly reduces memory and speeds up inference while maintaining performance, enabling efficient deployment for embodied AI.

🔹 Publication Date: Published on Feb 23

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2602.20309
• PDF: https://arxiv.org/pdf/2602.20309
• Project Page: https://quantvla.github.io/

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

For more data science resources:
✓ https://t.iss.one/DataScienceT

#Quantization #VLAModels #EmbodiedAI #AIResearch #DeepLearning

✨Multi-Vector Index Compression in Any Modality

📝 Summary:
This paper introduces attention-guided clustering AGC for compressing multi-vector document representations across various modalities. AGC consistently outperforms other compression methods in text, visual-document, and video retrieval, often matching or improving upon uncompressed indexes.

🔹 Publication Date: Published on Feb 24

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2602.21202
• PDF: https://arxiv.org/pdf/2602.21202

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

For more data science resources:
✓ https://t.iss.one/DataScienceT

#IndexCompression #MultiModal #InformationRetrieval #MachineLearning #VectorDatabases

✨PETS: A Principled Framework Towards Optimal Trajectory Allocation for Efficient Test-Time Self-Consistency

📝 Summary:
PETS is a principled framework for efficient test-time self-consistency that optimizes trajectory allocation. It defines a new self-consistency rate, reducing sampling requirements while maintaining accuracy. PETS significantly cuts sampling budgets by up to 75 percent offline and 55 percent onli...

🔹 Publication Date: Published on Feb 18

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2602.16745
• PDF: https://arxiv.org/pdf/2602.16745
• Github: https://github.com/ZDCSlab/PETS

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

For more data science resources:
✓ https://t.iss.one/DataScienceT

#SelfConsistency #MachineLearning #Optimization #AI #Efficiency

✨DeepSeek-V3 Technical Report

📝 Summary:
DeepSeek-V3 is an efficient Mixture-of-Experts language model 671B parameters using MLA and DeepSeekMoE architectures. It achieves strong performance, comparable to leading models, with highly stable and cost-effective training on 14.8T tokens.

🔹 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/Ki-Seki/ultrascale-playbook-zh-cn
• https://huggingface.co/spaces/weege007/ultrascale-playbook

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

For more data science resources:
✓ https://t.iss.one/DataScienceT

#DeepSeekV3 #MoE #LLM #AI #MachineLearning

✨See and Fix the Flaws: Enabling VLMs and Diffusion Models to Comprehend Visual Artifacts via Agentic Data Synthesis

📝 Summary:
ArtiAgent automates creating artifact-annotated image datasets. It uses three agents to perceive entities, inject artifacts into real images via diffusion transformers, and curate the results. This enables training models to detect and fix visual flaws in AI-generated content.

🔹 Publication Date: Published on Feb 24

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2602.20951
• PDF: https://arxiv.org/pdf/2602.20951
• Github: https://github.com/krafton-ai/ArtiAgent

✨ Datasets citing this paper:
• https://huggingface.co/datasets/KRAFTON/ArtiBench

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

For more data science resources:
✓ https://t.iss.one/DataScienceT

#AI #DataScience #MachineLearning #HuggingFace #Research

✨TAPE: Tool-Guided Adaptive Planning and Constrained Execution in Language Model Agents

📝 Summary:
TAPE framework improves language model agent performance in complex environments through enhanced planning and constrained execution strategies. AI-generated summary Language Model (LM) agents have de...

🔹 Publication Date: Published on Feb 23

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2602.19633
• PDF: https://arxiv.org/pdf/2602.19633
• Github: https://github.com/UW-Madison-Lee-Lab/TAPE

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

For more data science resources:
✓ https://t.iss.one/DataScienceT

#AI #DataScience #MachineLearning #HuggingFace #Research

✨Benchmark Test-Time Scaling of General LLM Agents

📝 Summary:
General AgentBench evaluates large language model agents across multiple domains and scaling methods, revealing performance degradation and fundamental limitations in sequential and parallel scaling a...

🔹 Publication Date: Published on Feb 22

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2602.18998
• PDF: https://arxiv.org/pdf/2602.18998
• Project Page: https://general-agentbench.github.io/
• Github: https://github.com/cxcscmu/General-AgentBench

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

For more data science resources:
✓ https://t.iss.one/DataScienceT

#AI #DataScience #MachineLearning #HuggingFace #Research

✨Learning to Detect Language Model Training Data via Active Reconstruction

📝 Summary:
Active Data Reconstruction Attack uses reinforcement learning to identify training data by measuring the reconstructibility of text from model behavior, outperforming existing membership inference att...

🔹 Publication Date: Published on Feb 22

🔹 Paper Links:
• arXiv Page: https://arxiv.org/pdf/2602.19020
• PDF: https://arxiv.org/pdf/2602.19020
• Project Page: https://huggingface.co/ADRA-RL
• Github: https://github.com/oseyosey/MIA-RL

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

For more data science resources:
✓ https://t.iss.one/DataScienceT

#AI #DataScience #MachineLearning #HuggingFace #Research

✨SIMSPINE: A Biomechanics-Aware Simulation Framework for 3D Spine Motion Annotation and Benchmarking

📝 Summary:
The SIMSPINE framework and dataset provide anatomically consistent 3D spinal annotations for natural human movements. This enables data-driven learning of vertebral kinematics and improves spine motion estimation accuracy, offering a benchmark for research.

🔹 Publication Date: Published on Feb 24

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2602.20792
• PDF: https://arxiv.org/pdf/2602.20792
• Project Page: https://saifkhichi.com/research/simspine
• Github: https://github.com/dfki-av/simspine

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

For more data science resources:
✓ https://t.iss.one/DataScienceT

#AI #DataScience #MachineLearning #HuggingFace #Research

✨RankEvolve: Automating the Discovery of Retrieval Algorithms via LLM-Driven Evolution

📝 Summary:
Large language models guided by evaluators and evolutionary search can automatically discover improved lexical retrieval algorithms through program evolution techniques. AI-generated summary Retrieval...

🔹 Publication Date: Published on Feb 18

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2602.16932
• PDF: https://arxiv.org/pdf/2602.16932
• Github: https://github.com/fangchenli/ranking-evolved

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

For more data science resources:
✓ https://t.iss.one/DataScienceT

#AI #DataScience #MachineLearning #HuggingFace #Research

✨JavisDiT++: Unified Modeling and Optimization for Joint Audio-Video Generation

📝 Summary:
JavisDiT++ presents a unified framework for high-quality, synchronized joint audio-video generation. It uses modality-specific Mixture-of-Experts, temporal-aligned RoPE for frame-level sync, and audio-video direct preference optimization. This achieves state-of-the-art performance with limited tr...

🔹 Publication Date: Published on Feb 22

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2602.19163
• PDF: https://arxiv.org/pdf/2602.19163
• Project Page: https://javisverse.github.io/JavisDiT2-page/
• Github: https://javisverse.github.io/JavisDiT2-page/

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

For more data science resources:
✓ https://t.iss.one/DataScienceT

#AI #DataScience #MachineLearning #HuggingFace #Research

✨HyTRec: A Hybrid Temporal-Aware Attention Architecture for Long Behavior Sequential Recommendation

📝 Summary:
HyTRec addresses the challenge of modeling long user behavior sequences by combining linear and softmax attention mechanisms with a temporal-aware delta network to balance efficiency and retrieval pre...

🔹 Publication Date: Published on Feb 20

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2602.18283
• PDF: https://arxiv.org/pdf/2602.18283

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

For more data science resources:
✓ https://t.iss.one/DataScienceT

#AI #DataScience #MachineLearning #HuggingFace #Research

✨ARLArena: A Unified Framework for Stable Agentic Reinforcement Learning

📝 Summary:
ARLArena framework analyzes training stability in agentic reinforcement learning and proposes SAMPO method for stable policy optimization across diverse tasks. AI-generated summary Agentic reinforceme...

🔹 Publication Date: Published on Feb 25

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2602.21534
• PDF: https://arxiv.org/pdf/2602.21534
• Github: https://github.com/WillDreamer/ARL-Arena

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

For more data science resources:
✓ https://t.iss.one/DataScienceT

#AI #DataScience #MachineLearning #HuggingFace #Research

✨SkyReels-V4: Multi-modal Video-Audio Generation, Inpainting and Editing model

📝 Summary:
SkyReels V4 is a unified multimodal video foundation model that generates, edits, and inpaints video and audio simultaneously using a dual-stream architecture with shared text encoding and efficient h...

🔹 Publication Date: Published on Feb 25

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2602.21818
• PDF: https://arxiv.org/pdf/2602.21818

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

For more data science resources:
✓ https://t.iss.one/DataScienceT

#AI #DataScience #MachineLearning #HuggingFace #Research

✨World Guidance: World Modeling in Condition Space for Action Generation

📝 Summary:
World Guidance framework enhances Vision-Language-Action models by mapping future observations into compact conditions for improved action generation and generalization. AI-generated summary Leveragin...

🔹 Publication Date: Published on Feb 25

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2602.22010
• PDF: https://arxiv.org/pdf/2602.22010
• Project Page: https://selen-suyue.github.io/WoGNet/

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

For more data science resources:
✓ https://t.iss.one/DataScienceT

#AI #DataScience #MachineLearning #HuggingFace #Research

✨JAEGER: Joint 3D Audio-Visual Grounding and Reasoning in Simulated Physical Environments

📝 Summary:
JAEGER extends audio-visual large language models to 3D space by integrating RGB-D observations and multi-channel audio to improve spatial reasoning and source localization. AI-generated summary Curre...

🔹 Publication Date: Published on Feb 20

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2602.18527
• PDF: https://arxiv.org/pdf/2602.18527

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

For more data science resources:
✓ https://t.iss.one/DataScienceT

#AI #DataScience #MachineLearning #HuggingFace #Research

✨Model Context Protocol (MCP) Tool Descriptions Are Smelly! Towards Improving AI Agent Efficiency with Augmented MCP Tool Descriptions

📝 Summary:
Foundation model agents rely on natural language tool descriptions for effective interaction with external systems, but poor description quality significantly impacts performance and efficiency. AI-ge...

🔹 Publication Date: Published on Feb 16

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2602.14878
• PDF: https://arxiv.org/pdf/2602.14878

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

For more data science resources:
✓ https://t.iss.one/DataScienceT

#AI #DataScience #MachineLearning #HuggingFace #Research

✨UniVBench: Towards Unified Evaluation for Video Foundation Models

📝 Summary:
UniVBench introduces a comprehensive benchmark for evaluating video foundation models across multiple capabilities including understanding, generation, editing, and reconstruction using high-quality, ...

🔹 Publication Date: Published on Feb 25

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2602.21835
• PDF: https://arxiv.org/pdf/2602.21835
• Github: https://github.com/JianhuiWei7/UniVBench

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

For more data science resources:
✓ https://t.iss.one/DataScienceT

#AI #DataScience #MachineLearning #HuggingFace #Research

✨The Design Space of Tri-Modal Masked Diffusion Models

📝 Summary:
A large-scale study of tri-modal discrete diffusion models demonstrates improved performance across text, image, and speech generation tasks through systematic analysis of scaling laws and optimized i...

🔹 Publication Date: Published on Feb 25

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2602.21472
• PDF: https://arxiv.org/pdf/2602.21472

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

For more data science resources:
✓ https://t.iss.one/DataScienceT

#AI #DataScience #MachineLearning #HuggingFace #Research