✨ATTN-FIQA: Interpretable Attention-based Face Image Quality Assessment with Vision Transformers

📝 Summary:
ATTN-FIQA uses pre-softmax attention scores from Vision Transformers to assess face image quality without additional training or architectural changes. AI-generated summary Face Image Quality Assessme...

🔹 Publication Date: Published on Apr 21

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2604.22841
• PDF: https://arxiv.org/pdf/2604.22841
• Github: https://github.com/gurayozgur/ATTN-FIQA

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

For more data science resources:
✓ https://xn--r1a.website/DataScienceT

#AI #DataScience #MachineLearning #HuggingFace #Research

✨EX-FIQA: Leveraging Intermediate Early eXit Representations from Vision Transformers for Face Image Quality Assessment

📝 Summary:
ViT-based face quality assessment method utilizes intermediate representations through early exit mechanisms and score fusion strategies, demonstrating that different transformer block depths capture ...

🔹 Publication Date: Published on Apr 21

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2604.22842
• PDF: https://arxiv.org/pdf/2604.22842
• Github: https://github.com/gurayozgur/EX-FIQA

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

For more data science resources:
✓ https://xn--r1a.website/DataScienceT

#AI #DataScience #MachineLearning #HuggingFace #Research

✨Improving Vision-language Models with Perception-centric Process Reward Models

📝 Summary:
A process reward model called Perceval enables token-level error detection and correction in vision-language models through perception-intensive training and fine-grained supervision during reinforcem...

🔹 Publication Date: Published on Apr 27

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2604.24583
• PDF: https://arxiv.org/pdf/2604.24583
• Github: https://github.com/RUCAIBox/Perceval

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

For more data science resources:
✓ https://xn--r1a.website/DataScienceT

#AI #DataScience #MachineLearning #HuggingFace #Research

✨TexOCR: Advancing Document OCR Models for Compilable Page-to-LaTeX Reconstruction

📝 Summary:
This research presents TexOCR for reconstructing scientific PDFs into compilable LaTeX, addressing limitations of current OCR. It introduces a new benchmark and trains TexOCR using reinforcement learning with verifiable rewards. This approach significantly improves structural accuracy and compila...

🔹 Publication Date: Published on Apr 24

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2604.22880
• PDF: https://arxiv.org/pdf/2604.22880
• Github: https://github.com/QDRhhhh/TexOCR

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

For more data science resources:
✓ https://xn--r1a.website/DataScienceT

#AI #DataScience #MachineLearning #HuggingFace #Research

✨How Much Is One Recurrence Worth? Iso-Depth Scaling Laws for Looped Language Models

📝 Summary:
Research quantifies the computational value of recurrent connections in language models through a scaling law that establishes a recurrence-equivalence exponent of 0.46, indicating that additional rec...

🔹 Publication Date: Published on Apr 27

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2604.21106
• PDF: https://arxiv.org/pdf/2604.21106
• Project Page: https://kschwethelm.github.io/looped-lm-scaling
• Github: https://github.com/kschwethelm/looped-lm-scaling

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

For more data science resources:
✓ https://xn--r1a.website/DataScienceT

#AI #DataScience #MachineLearning #HuggingFace #Research

✨UniGeo: Unifying Geometric Guidance for Camera-Controllable Image Editing via Video Models

📝 Summary:
UniGeo unifies geometric guidance across representation, architecture, and loss function levels in camera-controllable image editing. This novel framework addresses geometric drift and structural degradation, achieving superior visual quality and geometric consistency.

🔹 Publication Date: Published on Apr 19

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2604.17565
• PDF: https://arxiv.org/pdf/2604.17565
• Project Page: https://mo230761.github.io/UniGeo.github.io/
• Github: https://github.com/mo230761/UniGeo

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

For more data science resources:
✓ https://xn--r1a.website/DataScienceT

#AI #DataScience #MachineLearning #HuggingFace #Research

✨Quantum Kernel Advantage over Classical Collapse in Medical Foundation Model Embeddings

📝 Summary:
Quantum support vector machines demonstrate superior performance over classical linear and RBF kernels in binary insurance classification tasks using medical imaging data, with quantum kernels showing...

🔹 Publication Date: Published on Apr 27

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2604.24597
• PDF: https://arxiv.org/pdf/2604.24597
• Project Page: https://sebasmos.github.io/qml-medimage/
• Github: https://github.com/sebasmos/qml-medimage

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

For more data science resources:
✓ https://xn--r1a.website/DataScienceT

#QuantumMachineLearning #MedicalAI #QuantumComputing #MachineLearning #DataScience

✨Learning to Identify Out-of-Distribution Objects for 3D LiDAR Anomaly Segmentation

📝 Summary:
A novel 3D LiDAR anomaly segmentation method operates directly in feature space to distinguish known from unknown objects, addressing limitations of existing datasets through mixed real-synthetic data...

🔹 Publication Date: Published on Apr 26

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2604.23604
• PDF: https://arxiv.org/pdf/2604.23604
• Project Page: https://simom0.github.io/lido-page/
• Github: https://github.com/SiMoM0/LIDO

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

For more data science resources:
✓ https://xn--r1a.website/DataScienceT

#AI #DataScience #MachineLearning #HuggingFace #Research

✨PageGuide: Browser extension to assist users in navigating a webpage and locating information

📝 Summary:
PageGuide is a browser extension that enhances AI assistant interactions by providing visual grounding of responses in web page elements, improving verification, guidance, and focus during web browsin...

🔹 Publication Date: Published on Apr 26

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2604.23772
• PDF: https://arxiv.org/pdf/2604.23772
• Project Page: https://pageguide.github.io/
• Github: https://github.com/tin-xai/pageguide

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

For more data science resources:
✓ https://xn--r1a.website/DataScienceT

#AI #DataScience #MachineLearning #HuggingFace #Research

✨RaV-IDP: A Reconstruction-as-Validation Framework for Faithful Intelligent Document Processing

📝 Summary:
Reconstruction as Validation (RaV-IDP) introduces a document processing pipeline that uses reconstruction and comparison against original sources to validate extraction quality, triggering fallback me...

🔹 Publication Date: Published on Apr 26

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2604.23644
• PDF: https://arxiv.org/pdf/2604.23644
• Project Page: https://github.com/pritesh-2711/RaV-IDP/releases/tag/v1.0.0
• Github: https://github.com/pritesh-2711/RaV-IDP

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

For more data science resources:
✓ https://xn--r1a.website/DataScienceT

#AI #DataScience #MachineLearning #HuggingFace #Research

✨Disentangled Robot Learning via Separate Forward and Inverse Dynamics Pretraining

📝 Summary:
DeFI addresses vision-language-action model limitations by decoupling visual forward and inverse dynamics pretraining to improve 3D action prediction and enable learning from large-scale action-free v...

🔹 Publication Date: Published on Mar 27

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2604.16391
• PDF: https://arxiv.org/pdf/2604.16391
• Project Page: https://github.com/LogosRoboticsGroup/DeFi
• Github: https://github.com/LogosRoboticsGroup/DeFi

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

For more data science resources:
✓ https://xn--r1a.website/DataScienceT

#AI #DataScience #MachineLearning #HuggingFace #Research

✨Credal Concept Bottleneck Models for Epistemic-Aleatoric Uncertainty Decomposition

📝 Summary:
CREDENCE is a CBM framework that decomposes concept uncertainty into epistemic and aleatoric components. It uses credal predictions and ensemble methods to provide distinct uncertainty signals, enabling more informed decisions based on whether uncertainty is due to model underspecification or inp...

🔹 Publication Date: Published on Apr 27

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2604.24170
• PDF: https://arxiv.org/pdf/2604.24170
• Github: https://github.com/Tankiit/Credal_Sets/tree/ensemble-credal-cbm

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

For more data science resources:
✓ https://xn--r1a.website/DataScienceT

#AI #DataScience #MachineLearning #HuggingFace #Research

✨Sapiens2

📝 Summary:
Sapiens2 is a high-resolution transformer model for human-centric vision. It achieves state-of-the-art performance by combining unified pretraining objectives, a large 1-billion image dataset, and architectural improvements, excelling in tasks like pose and segmentation.

🔹 Publication Date: Published on Apr 23

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2604.21681
• PDF: https://arxiv.org/pdf/2604.21681
• Github: https://github.com/facebookresearch/sapiens2

🔹 Models citing this paper:
• https://huggingface.co/facebook/sapiens2
• https://huggingface.co/facebook/sapiens2-seg-5b
• https://huggingface.co/facebook/sapiens2-seg-1b

✨ Spaces citing this paper:
• https://huggingface.co/spaces/facebook/sapiens2-seg
• https://huggingface.co/spaces/facebook/sapiens2-pointmap
• https://huggingface.co/spaces/facebook/sapiens2-normal

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

For more data science resources:
✓ https://xn--r1a.website/DataScienceT

#Sapiens2 #ComputerVision #TransformerModels #HumanCentricAI #DeepLearning

✨Personality Shapes Gender Bias in Persona-Conditioned LLM Narratives Across English and Hindi: An Empirical Investigation

📝 Summary:
This study found that gender bias in persona-conditioned LLM narratives is context-dependent, varying with personality traits and language. Dark Triad traits consistently led to more gender-stereotypical representations, highlighting uneven representational harms.

🔹 Publication Date: Published on Apr 26

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2604.23600
• PDF: https://arxiv.org/pdf/2604.23600

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

For more data science resources:
✓ https://xn--r1a.website/DataScienceT

#LLM #GenderBias #AIethics #Personality #NLP

✨Improving Robustness of Tabular Retrieval via Representational Stability

📝 Summary:
T r a n s f o r m e r - b a s e d t a b l e r e t r i e v a l s y s t e m s f l a t t e n s t r u c t u r e d t a b l e s i n t o t o k e n s e q u e n c e s , m a k i n g r e t r i e v a l s e n s i ...

🔹 Publication Date: Published on Apr 27

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2604.24040
• PDF: https://arxiv.org/pdf/2604.24040
• Github: https://github.com/KBhandari11/Centroid-Aligned-Table-Retrieval

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

For more data science resources:
✓ https://xn--r1a.website/DataScienceT

#AI #DataScience #MachineLearning #HuggingFace #Research

✨Why Fine-Tuning Encourages Hallucinations and How to Fix It

📝 Summary:
Supervised fine-tuning in large language models can cause factual hallucinations due to knowledge degradation, which can be reduced through self-distillation regularization and parameter freezing tech...

🔹 Publication Date: Published on Apr 16

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2604.15574
• PDF: https://arxiv.org/pdf/2604.15574

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

For more data science resources:
✓ https://xn--r1a.website/DataScienceT

#AI #DataScience #MachineLearning #HuggingFace #Research

✨IndustryAssetEQA: A Neurosymbolic Operational Intelligence System for Embodied Question Answering in Industrial Asset Maintenance

📝 Summary:
IndustryAssetEQA is a neurosymbolic system that improves industrial maintenance AI. It combines telemetry data with a knowledge graph for embodied question answering, offering more reliable and explainable insights. This reduces overclaims and enhances explanation quality over LLM-only baselines.

🔹 Publication Date: Published on Apr 25

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2604.23446
• PDF: https://arxiv.org/pdf/2604.23446
• Project Page: https://github.com/IBM/AssetOpsBench/tree/IndustryAssetEQA/IndustryAssetEQA

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

For more data science resources:
✓ https://xn--r1a.website/DataScienceT

#AI #DataScience #MachineLearning #HuggingFace #Research

✨Recursive Multi-Agent Systems

📝 Summary:
RecursiveMAS scales multi-agent collaboration via recursive latent-space computation, connecting agents with RecursiveLink. It achieves 8.3% higher accuracy, 1.2-2.4x faster inference, and 34.6-75.6% less token usage on various benchmarks.

🔹 Publication Date: Published on Apr 28

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2604.25917
• PDF: https://arxiv.org/pdf/2604.25917
• Project Page: https://recursivemas.github.io
• Github: https://recursivemas.github.io

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

For more data science resources:
✓ https://xn--r1a.website/DataScienceT

#AI #DataScience #MachineLearning #HuggingFace #Research

✨Programming with Data: Test-Driven Data Engineering for Self-Improving LLMs from Raw Corpora

📝 Summary:
Training data can be structured as source code and evaluated through unit testing, enabling systematic debugging and improvement of language models' domain-specific capabilities. AI-generated summary ...

🔹 Publication Date: Published on Apr 27

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2604.24819
• PDF: https://arxiv.org/pdf/2604.24819
• Project Page: https://openraiser.github.io/ProDa-webpage
• Github: https://github.com/OpenRaiser/ProDa

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

For more data science resources:
✓ https://xn--r1a.website/DataScienceT

#AI #DataScience #MachineLearning #HuggingFace #Research

✨Meta-CoT: Enhancing Granularity and Generalization in Image Editing

📝 Summary:
Meta-CoT enhances image editing by decomposing editing operations into task-target-understanding triplets and fundamental meta-tasks, improving both granularity and generalization through CoT-editing ...

🔹 Publication Date: Published on Apr 27

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2604.24625
• PDF: https://arxiv.org/pdf/2604.24625
• Project Page: https://shiyi-zh0408.github.io/projectpages/Meta-CoT/
• Github: https://github.com/shiyi-zh0408/Meta-CoT

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

For more data science resources:
✓ https://xn--r1a.website/DataScienceT

#AI #DataScience #MachineLearning #HuggingFace #Research