✨Investigating Autonomous Agent Contributions in the Wild: Activity Patterns and Code Change over Time

📝 Summary:
Researchers analyzed AI coding agent contributions to open source projects. They found increasing agent activity but higher code churn over time compared to human-authored code.

🔹 Publication Date: Published on Apr 1

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2604.00917
• PDF: https://arxiv.org/pdf/2604.00917
• Project Page: https://arxiv.org/html/2604.00917v1

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

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

#AIAgents #SoftwareEngineering #OpenSource #CodeQuality #AIResearch

✨AutoMIA: Improved Baselines for Membership Inference Attack via Agentic Self-Exploration

📝 Summary:
AutoMIA is an agentic framework that automates membership inference attacks. It dynamically generates and refines attack strategies via self-exploration and closed-loop evaluation. This approach consistently outperforms static methods by eliminating manual feature engineering and improving adapta...

🔹 Publication Date: Published on Apr 1

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2604.01014
• PDF: https://arxiv.org/pdf/2604.01014
• Github: https://github.com/amiya-special/AutoMIA

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

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

#MembershipInference #MLSecurity #Cybersecurity #AI #DataPrivacy

✨Memory-Augmented Vision-Language Agents for Persistent and Semantically Consistent Object Captioning

📝 Summary:
A memory-augmented VLM agent resolves inconsistent object descriptions across viewpoints. It unifies data association, captioning, and exploration within a single framework, leveraging object-level memory for persistent semantic consistency and improved scores.

🔹 Publication Date: Published on Mar 30

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2603.24257
• PDF: https://arxiv.org/pdf/2603.24257
• Project Page: https://hsp-iit.github.io/epos-vlm/

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

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

#VLM #ObjectCaptioning #AI #ComputerVision #DeepLearning

✨Tex3D: Objects as Attack Surfaces via Adversarial 3D Textures for Vision-Language-Action Models

📝 Summary:
Tex3D is the first framework optimizing 3D adversarial textures to attack vision-language-action models. It significantly degrades robotic manipulation performance in real-world settings, revealing critical vulnerabilities.

🔹 Publication Date: Published on Apr 2

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2604.01618
• PDF: https://arxiv.org/pdf/2604.01618
• Project Page: https://vla-attack.github.io/tex3d/
• Github: https://github.com/vla-attack/tex3d

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

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

#AdversarialAI #Robotics #VLAmodels #Cybersecurity #ComputerVision

✨Efficient Universal Perception Encoder

📝 Summary:
EUPE enhances edge device performance through a novel two-stage knowledge distillation approach. It scales up to a large proxy teacher then down to an efficient encoder. This method provides superior, versatile representations for diverse tasks, outperforming prior techniques.

🔹 Publication Date: Published on Mar 23

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2603.22387
• PDF: https://arxiv.org/pdf/2603.22387
• Github: https://github.com/facebookresearch/eupe

🔹 Models citing this paper:
• https://huggingface.co/facebook/EUPE-ConvNeXt-S
• https://huggingface.co/facebook/EUPE-ViT-S
• https://huggingface.co/facebook/EUPE-ViT-B

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

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

#KnowledgeDistillation #EdgeAI #ComputerVision #DeepLearning #RepresentationLearning

✨Steerable Visual Representations

📝 Summary:
Steerable Visual Representations allow language-guided focus on specific image elements while maintaining high representation quality. This is achieved through early fusion of text directly into the visual encoder. Our method outperforms dedicated approaches and generalizes well to new tasks.

🔹 Publication Date: Published on Apr 2

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2604.02327
• PDF: https://arxiv.org/pdf/2604.02327
• Project Page: https://jonaruthardt.github.io/project/SteerViT/
• Github: https://github.com/JonaRuthardt/SteerViT

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

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

#ComputerVision #DeepLearning #MultimodalAI #ImageRecognition #AI

✨ASI-Evolve: AI Accelerates AI

📝 Summary:
ASI-Evolve is an AI framework demonstrating AI-driven discovery across key AI development components. It achieved superior performance in neural architecture design, data curation, and reinforcement learning algorithm design, showing AI can accelerate AI itself.

🔹 Publication Date: Published on Mar 31

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2603.29640
• PDF: https://arxiv.org/pdf/2603.29640
• Github: https://github.com/GAIR-NLP/ASI-Evolve

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

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

#AI #AIAcceleration #MachineLearning #DeepLearning #AIResearch

✨AIBench: Evaluating Visual-Logical Consistency in Academic Illustration Generation

📝 Summary:
AIBench evaluates academic illustration quality through logic correctness and aesthetics using VQA and VLM assessments, revealing significant performance gaps and the challenge of optimizing both aspe...

🔹 Publication Date: Published on Mar 31

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2603.28068
• PDF: https://arxiv.org/pdf/2603.28068
• Project Page: https://deep-kaixun.github.io/aibench-page/
• Github: https://deep-kaixun.github.io/aibench-page/

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

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

#AI #DataScience #MachineLearning #HuggingFace #Research

✨Efficient and Principled Scientific Discovery through Bayesian Optimization: A Tutorial

📝 Summary:
Bayesian optimisation provides a principled probabilistic framework for automating scientific discovery by iteratively refining hypotheses and selecting experiments to balance exploration and exploita...

🔹 Publication Date: Published on Apr 1

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2604.01328
• PDF: https://arxiv.org/pdf/2604.01328

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

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

#AI #DataScience #MachineLearning #HuggingFace #Research

✨MDPBench: A Benchmark for Multilingual Document Parsing in Real-World Scenarios

📝 Summary:
A new multilingual document parsing benchmark reveals significant performance gaps between closed-source and open-source models, especially on non-Latin scripts and photographed documents. AI-generate...

🔹 Publication Date: Published on Mar 30

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2603.28130
• PDF: https://arxiv.org/pdf/2603.28130
• Github: https://github.com/Yuliang-Liu/MultimodalOCR

✨ Datasets citing this paper:
• https://huggingface.co/datasets/Delores-Lin/MDPBench

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

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

#MultilingualNLP #DocumentAI #OCR #AIbenchmark #MachineLearning

✨Ask or Assume? Uncertainty-Aware Clarification-Seeking in Coding Agents

📝 Summary:
A multi-agent system using uncertainty-aware design improves LLM agent performance on underspecified software development tasks by detecting ambiguity and proactively seeking clarification. AI-generat...

🔹 Publication Date: Published on Mar 27

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2603.26233
• PDF: https://arxiv.org/pdf/2603.26233
• Github: https://github.com/nedwards99/ask-or-assume

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

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

#AI #DataScience #MachineLearning #HuggingFace #Research

✨Brainstacks: Cross-Domain Cognitive Capabilities via Frozen MoE-LoRA Stacks for Continual LLM Learning

📝 Summary:
Brainstacks enables continual multi-domain fine-tuning of large language models through modular adapter stacks with MoE-LoRA, residual boosting, and outcome-based routing that discovers transferable c...

🔹 Publication Date: Published on Apr 1

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2604.01152
• PDF: https://arxiv.org/pdf/2604.01152
• Project Page: https://huggingface.co/papers?q=null-space%20projection
• Github: https://github.com/achelousace/brainstacks

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

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

#LLM #ContinualLearning #MoELoRA #DeepLearning #AIResearch

✨ActionParty: Multi-Subject Action Binding in Generative Video Games

📝 Summary:
ActionParty solves the multi-agent control problem in video diffusion models. It introduces subject state tokens to disentangle global video rendering from individual action control. This allows simultaneous control of up to seven players across diverse environments, improving action following an...

🔹 Publication Date: Published on Apr 2

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2604.02330
• PDF: https://arxiv.org/pdf/2604.02330
• Project Page: https://action-party.github.io/
• Github: https://action-party.github.io/

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

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

#AI #DataScience #MachineLearning #HuggingFace #Research

✨Working Notes on Late Interaction Dynamics: Analyzing Targeted Behaviors of Late Interaction Models

📝 Summary:
This work analyzes Late Interaction models, showing a length bias in multi-vector scoring for causal and sometimes bi-directional models. It also confirms MaxSim efficiently exploits token-level similarity, with no significant trends beyond the top token.

🔹 Publication Date: Published on Mar 27

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2603.26259
• PDF: https://arxiv.org/pdf/2603.26259

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

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

#LateInteraction #InformationRetrieval #ModelAnalysis #AlgorithmBias #SimilaritySearch

✨NearID: Identity Representation Learning via Near-identity Distractors

📝 Summary:
Researchers developed a novel framework using Near-identity distractors to improve identity-focused vision tasks by creating a dataset and evaluation protocol that better isolates identity from backgr...

🔹 Publication Date: Published on Apr 2

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2604.01973
• PDF: https://arxiv.org/pdf/2604.01973
• Project Page: https://gorluxor.github.io/NearID/
• Github: https://github.com/Gorluxor/NearID

✨ Datasets citing this paper:
• https://huggingface.co/datasets/Aleksandar/NearID
• https://huggingface.co/datasets/Aleksandar/NearID-Flux
• https://huggingface.co/datasets/Aleksandar/NearID-FluxC

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

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

#AI #DataScience #MachineLearning #HuggingFace #Research

✨Therefore I am. I Think

📝 Summary:
Large language models often make action choices before generating any reasoning text. Evidence shows early decision signals can be decoded and causally steered, with the subsequent 'thinking' rationalizing the pre-made choice. This suggests decisions precede explicit deliberation.

🔹 Publication Date: Published on Apr 2

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2604.01202
• PDF: https://arxiv.org/pdf/2604.01202

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

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

#AI #DataScience #MachineLearning #HuggingFace #Research

✨MultiGen: Level-Design for Editable Multiplayer Worlds in Diffusion Game Engines

📝 Summary:
Video world models with explicit external memory enable user-controlled environment editing and real-time multiplayer interactions by decomposing generation into memory, observation, and dynamics modu...

🔹 Publication Date: Published on Mar 30

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2603.06679
• PDF: https://arxiv.org/pdf/2603.06679

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

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

#AI #DataScience #MachineLearning #HuggingFace #Research

✨Forecasting Supply Chain Disruptions with Foresight Learning

📝 Summary:
Large language models can be trained to produce calibrated probabilistic forecasts for supply chain disruptions, outperforming existing baselines and enabling decision-ready predictions through domain...

🔹 Publication Date: Published on Apr 1

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2604.01298
• PDF: https://arxiv.org/pdf/2604.01298
• Project Page: https://lightningrod.ai

✨ Datasets citing this paper:
• https://huggingface.co/datasets/LightningRodLabs/supply-chain-predictions

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

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

#AI #DataScience #MachineLearning #HuggingFace #Research

✨CORAL: Towards Autonomous Multi-Agent Evolution for Open-Ended Discovery

📝 Summary:
Autonomous multi-agent evolution framework enables open-ended discovery through persistent memory, asynchronous execution, and collaborative problem-solving, achieving superior performance on mathemat...

🔹 Publication Date: Published on Apr 2

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2604.01658
• PDF: https://arxiv.org/pdf/2604.01658
• Project Page: https://human-agent-society.github.io/CORAL
• Github: https://github.com/Human-Agent-Society/CORAL

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

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

#AI #DataScience #MachineLearning #HuggingFace #Research

✨Video Models Reason Early: Exploiting Plan Commitment for Maze Solving

📝 Summary:
Video diffusion models demonstrate emergent reasoning abilities in maze solving through early plan commitment and path length prediction, with improved performance achieved via Chaining with Early Pla...

🔹 Publication Date: Published on Mar 31

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2603.30043
• PDF: https://arxiv.org/pdf/2603.30043
• Project Page: https://video-maze-reasoning.github.io/

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

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

#AI #DataScience #MachineLearning #HuggingFace #Research