✨Beyond Multiple Choice: Verifiable OpenQA for Robust Vision-Language RFT

📝 Summary:
ReVeL converts multiple-choice questions to verifiable open-form questions to address unreliable MCQA metrics and answer guessing. This framework improves data efficiency and robustness for multimodal language models, revealing significant score inflation in MCQA benchmarks.

🔹 Publication Date: Published on Nov 21

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2511.17405
• PDF: https://arxiv.org/pdf/2511.17405
• Github: https://flageval-baai.github.io/ReVeL/

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

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

#OpenQA #VisionLanguage #LanguageModels #AIEvaluation #MachineLearning

✨Agent0-VL: Exploring Self-Evolving Agent for Tool-Integrated Vision-Language Reasoning

📝 Summary:
Agent0-VL is a self-evolving vision-language agent that integrates tool usage into both reasoning and self-evaluation. It uses a Solver and Verifier in a self-evolving cycle for continuous improvement without human annotation or external rewards, achieving a 12.5% performance gain.

🔹 Publication Date: Published on Nov 25

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2511.19900
• PDF: https://arxiv.org/pdf/2511.19900

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

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

#AIAgents #VisionLanguage #SelfEvolvingAI #ToolAugmentedAI #AIResearch

✨Decouple to Generalize: Context-First Self-Evolving Learning for Data-Scarce Vision-Language Reasoning

📝 Summary:
DoGe is a framework that addresses data scarcity in vision-language models. It decouples context learning from problem solving, using a curriculum to improve reward signals and data diversity. This enhances generalization and performance.

🔹 Publication Date: Published on Dec 7

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2512.06835
• PDF: https://arxiv.org/pdf/2512.06835

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

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

#VisionLanguage #DataScarcity #MachineLearning #AIResearch #DeepLearning

✨An Anatomy of Vision-Language-Action Models: From Modules to Milestones and Challenges

📝 Summary:
This survey offers a structured guide to Vision-Language-Action VLA models in robotics. It breaks down five key challenges: representation, execution, generalization, safety, and datasets, serving as a roadmap for researchers.

🔹 Publication Date: Published on Dec 12

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2512.11362
• PDF: https://arxiv.org/pdf/2512.11362
• Project Page: https://suyuz1.github.io/Survery/
• Github: https://suyuz1.github.io/VLA-Survey-Anatomy/

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

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

#VLAModels #Robotics #ArtificialIntelligence #VisionLanguage #AIResearch

✨CASA: Cross-Attention via Self-Attention for Efficient Vision-Language Fusion

📝 Summary:
CASA enhances cross-attention for vision-language models by adding local text-to-text interaction. This approach substantially reduces the performance gap with costly token insertion methods on detailed visual tasks. CASA maintains efficiency and scalability for long-context multimodal applicatio...

🔹 Publication Date: Published on Dec 22

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2512.19535
• PDF: https://arxiv.org/pdf/2512.19535
• Project Page: https://kyutai.org/casa
• Github: https://github.com/kyutai-labs/casa

🔹 Models citing this paper:
• https://huggingface.co/kyutai/CASA-Helium1-VL-2B

✨ Spaces citing this paper:
• https://huggingface.co/spaces/kyutai/casa-samples

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

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

#VisionLanguage #MultimodalAI #AttentionMechanisms #EfficientAI #DeepLearning

✨Youtu-VL: Unleashing Visual Potential via Unified Vision-Language Supervision

📝 Summary:
Youtu-VL introduces a Vision-Language Unified Autoregressive Supervision paradigm. It shifts from vision-as-input to vision-as-target, integrating visual tokens into the prediction stream. This improves multimodal comprehension and vision-centric task performance, fostering generalist visual agents.

🔹 Publication Date: Published on Jan 27

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2601.19798
• PDF: https://arxiv.org/pdf/2601.19798
• Project Page: https://youtu-tip.com/#llm
• Github: https://github.com/TencentCloudADP/youtu-vl

🔹 Models citing this paper:
• https://huggingface.co/tencent/Youtu-VL-4B-Instruct
• https://huggingface.co/tencent/Youtu-VL-4B-Instruct-GGUF
• https://huggingface.co/tencent/Youtu-Parsing

✨ Spaces citing this paper:
• https://huggingface.co/spaces/tencent/Youtu-Parsing

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

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

#YoutuVL #VisionLanguage #MultimodalAI #ComputerVision #DeepLearning

✨ExpAlign: Expectation-Guided Vision-Language Alignment for Open-Vocabulary Grounding

📝 Summary:
ExpAlign proposes an expectation-guided vision-language alignment framework using multiple instance learning and attention pooling. It implicitly selects tokens and instances without extra annotations, significantly boosting open-vocabulary detection and zero-shot instance segmentation.

🔹 Publication Date: Published on Jan 30

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2601.22666
• PDF: https://arxiv.org/pdf/2601.22666

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

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

#ComputerVision #DeepLearning #AI #VisionLanguage #OpenVocabulary

✨MedXIAOHE: A Comprehensive Recipe for Building Medical MLLMs

📝 Summary:
MedXIAOHE is a medical vision-language foundation model achieving state-of-the-art performance. It uses entity-aware pretraining, reinforcement learning, and tool-augmented training for reliable, expert-level diagnostic reasoning with low hallucination.

🔹 Publication Date: Published on Feb 13

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2602.12705
• PDF: https://arxiv.org/pdf/2602.12705

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

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

#MedicalAI #MLLMs #VisionLanguage #DiagnosticAI #FoundationModels

✨CoME-VL: Scaling Complementary Multi-Encoder Vision-Language Learning

📝 Summary:
CoME-VL fuses contrastive and self-supervised vision encoders to improve vision-language models. It uses entropy-guided aggregation and RoPE-enhanced attention for better visual understanding and grounding, outperforming single-encoder baselines.

🔹 Publication Date: Published on Apr 3

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2604.03231
• PDF: https://arxiv.org/pdf/2604.03231
• Project Page: https://mbzuai-oryx.github.io/CoME-VL/
• Github: https://github.com/mbzuai-oryx/CoME-VL

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

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

#VisionLanguage #MultimodalAI #ComputerVision #MachineLearning #DeepLearning

✨Concrete Jungle: Towards Concreteness Paved Contrastive Negative Mining for Compositional Understanding

📝 Summary:
This paper improves vision-language models for compositional reasoning by using concreteness-based negative sample selection and a novel margin-based loss. Their framework, Slipform, achieves state-of-the-art accuracy on compositional benchmarks and cross-modal retrieval.

🔹 Publication Date: Published on Apr 14

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2604.13313
• PDF: https://arxiv.org/pdf/2604.13313

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

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

#VisionLanguage #DeepLearning #AIResearch #ComputerVision #NLP