✨How2Everything: Mining the Web for How-To Procedures to Evaluate and Improve LLMs

📝 Summary:
A scalable framework for evaluating and improving goal-conditioned procedure generation using large-scale web mining, automated scoring, and reinforcement learning to enhance step-by-step instruction ...

🔹 Publication Date: Published on Feb 9

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2602.08808
• PDF: https://arxiv.org/pdf/2602.08808
• Github: https://github.com/lilakk/how2everything

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

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

#AI #DataScience #MachineLearning #HuggingFace #Research

✨GISA: A Benchmark for General Information-Seeking Assistant

📝 Summary:
A new benchmark called GISA is introduced for evaluating information-seeking assistants, featuring human-crafted queries with structured answer formats and live updates to prevent memorization. AI-gen...

🔹 Publication Date: Published on Feb 9

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2602.08543
• PDF: https://arxiv.org/pdf/2602.08543

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

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

#AI #DataScience #MachineLearning #HuggingFace #Research

✨Rolling Sink: Bridging Limited-Horizon Training and Open-Ended Testing in Autoregressive Video Diffusion

📝 Summary:
Autoregressive video diffusion models suffer from train-test gaps when generating long videos, but a training-free approach called Rolling Sink addresses this by maintaining AR cache and enabling ultr...

🔹 Publication Date: Published on Feb 8

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2602.07775
• PDF: https://arxiv.org/pdf/2602.07775
• Project Page: https://rolling-sink.github.io/

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

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

#AI #DataScience #MachineLearning #HuggingFace #Research

✨Concept-Aware Privacy Mechanisms for Defending Embedding Inversion Attacks

📝 Summary:
SPARSE is a user-centric framework that protects text embeddings from privacy leaks by selectively perturbing sensitive dimensions using differentiable masking and Mahalanobis noise calibration. AI-ge...

🔹 Publication Date: Published on Feb 6

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2602.07090
• PDF: https://arxiv.org/pdf/2602.07090

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

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

#AI #DataScience #MachineLearning #HuggingFace #Research

✨Aster: Autonomous Scientific Discovery over 20x Faster Than Existing Methods

📝 Summary:
Aster is an AI agent that accelerates scientific discovery by iteratively improving programs, achieving state-of-the-art results across multiple domains including mathematics, biology, and machine lea...

🔹 Publication Date: Published on Feb 3

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2602.07040
• PDF: https://arxiv.org/pdf/2602.07040
• Project Page: https://www.asterlab.ai/

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

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

#AI #DataScience #MachineLearning #HuggingFace #Research

✨Weak-Driven Learning: How Weak Agents make Strong Agents Stronger

📝 Summary:
WMSS is a post-training paradigm that uses weak model checkpoints to identify and fill learning gaps, enabling continued improvement beyond conventional saturation points in large language models. AI-...

🔹 Publication Date: Published on Feb 9

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2602.08222
• PDF: https://arxiv.org/pdf/2602.08222

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

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

#AI #DataScience #MachineLearning #HuggingFace #Research

✨Theory of Space: Can Foundation Models Construct Spatial Beliefs through Active Exploration?

📝 Summary:
Current multimodal foundation models show limitations in maintaining coherent spatial beliefs during active exploration, exhibiting gaps between active and passive performance, inefficient exploration...

🔹 Publication Date: Published on Feb 4

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2602.07055
• PDF: https://arxiv.org/pdf/2602.07055
• Project Page: https://theory-of-space.github.io/
• Github: https://github.com/mll-lab-nu/Theory-of-Space

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

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

#AI #DataScience #MachineLearning #HuggingFace #Research

✨Learning-guided Kansa collocation for forward and inverse PDEs beyond linearity

📝 Summary:
Research explores PDE solvers including neural frameworks for scientific simulations, examining forward solutions, inverse problems, and equation discovery across multi-variable and non-linear systems...

🔹 Publication Date: Published on Feb 8

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2602.07970
• PDF: https://arxiv.org/pdf/2602.07970

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

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

#AI #DataScience #MachineLearning #HuggingFace #Research

✨MotionCrafter: Dense Geometry and Motion Reconstruction with a 4D VAE

📝 Summary:
MotionCrafter is a video diffusion framework that jointly reconstructs 4D geometry and estimates dense motion using a novel joint representation and 4D VAE architecture. AI-generated summary We introd...

🔹 Publication Date: Published on Feb 9

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2602.08961
• PDF: https://arxiv.org/pdf/2602.08961
• Project Page: https://ruijiezhu94.github.io/MotionCrafter_Page
• Github: https://github.com/TencentARC/MotionCrafter

🔹 Models citing this paper:
• https://huggingface.co/TencentARC/MotionCrafter

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

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

#AI #DataScience #MachineLearning #HuggingFace #Research

✨SoulX-Singer: Towards High-Quality Zero-Shot Singing Voice Synthesis

📝 Summary:
A high-quality open-source singing voice synthesis system is presented with support for multiple languages and controllable generation, along with a dedicated benchmark for evaluating zero-shot perfor...

🔹 Publication Date: Published on Feb 8

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2602.07803
• PDF: https://arxiv.org/pdf/2602.07803

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

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

#AI #DataScience #MachineLearning #HuggingFace #Research

✨AVERE: Improving Audiovisual Emotion Reasoning with Preference Optimization

📝 Summary:
A benchmark and optimization technique are presented to improve multimodal large language models' emotion understanding by addressing spurious associations and hallucinations in audiovisual cues. AI-g...

🔹 Publication Date: Published on Feb 4

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2602.07054
• PDF: https://arxiv.org/pdf/2602.07054
• Project Page: https://avere-iclr.github.io/
• Github: https://avere-iclr.github.io/

✨ Datasets citing this paper:
• https://huggingface.co/datasets/chaubeyG/EmoReAlM

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

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

#AI #DataScience #MachineLearning #HuggingFace #Research

✨Learning Query-Aware Budget-Tier Routing for Runtime Agent Memory

📝 Summary:
BudgetMem is a runtime memory framework for LLM agents. It uses modular components with budget tiers and a neural router to optimize memory performance-cost trade-offs, outperforming baselines and achieving better accuracy-cost frontiers.

🔹 Publication Date: Published on Feb 5

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2602.06025
• PDF: https://arxiv.org/pdf/2602.06025
• Project Page: https://viktoraxelsen.github.io/BudgetMem/
• Github: https://github.com/ViktorAxelsen/BudgetMem

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

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

#LLMAgents #MemoryManagement #AI #MachineLearning #Optimization

✨GEBench: Benchmarking Image Generation Models as GUI Environments

📝 Summary:
This paper introduces GEBench, a new benchmark and GE-Score metric for evaluating temporal coherence and dynamic interaction in GUI generation models. Evaluations show current models struggle significantly with consistency and grounding over longer interaction sequences.

🔹 Publication Date: Published on Feb 9

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2602.09007
• PDF: https://arxiv.org/pdf/2602.09007
• Github: https://github.com/stepfun-ai/GEBench

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

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

#ImageGeneration #GUIGeneration #AIResearch #Benchmarking #MachineLearning

✨Thinking Makes LLM Agents Introverted: How Mandatory Thinking Can Backfire in User-Engaged Agents

📝 Summary:
Mandatory explicit thinking in user-engaged LLM agents often degrades performance. This occurs because thinking makes agents introverted, shortening responses and reducing information disclosure. Prompting for transparency significantly improves agent performance by enhancing communication.

🔹 Publication Date: Published on Feb 8

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2602.07796
• PDF: https://arxiv.org/pdf/2602.07796

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

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

#LLMAgents #AIResearch #PromptEngineering #HumanAIInteraction #AIBehavior

✨FlexMoRE: A Flexible Mixture of Rank-heterogeneous Experts for Efficient Federatedly-trained Large Language Models

📝 Summary:
FlexMoRE proposes replacing full-sized experts with low-rank adapters in Mixture-of-Experts for federated LLMs. This flexible approach improves performance using significantly fewer parameters, with optimal expert rank depending on task complexity.

🔹 Publication Date: Published on Feb 9

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2602.08818
• PDF: https://arxiv.org/pdf/2602.08818

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

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

#LLM #FederatedLearning #MixtureOfExperts #AI #DeepLearning

✨GraphAgents: Knowledge Graph-Guided Agentic AI for Cross-Domain Materials Design

📝 Summary:
GraphAgents is a multi-agent AI framework using knowledge graphs to solve complex materials design problems. It deploys specialized agents for tasks like evidence retrieval and graph traversal, outperforming single-shot LLMs. This approach effectively identifies sustainable PFAS alternatives, exp...

🔹 Publication Date: Published on Feb 7

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2602.07491
• PDF: https://arxiv.org/pdf/2602.07491

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

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

#AI #KnowledgeGraphs #AgenticAI #MaterialsDesign #MultiAgentSystems

✨On Randomness in Agentic Evals

📝 Summary:
Agentic system evaluations using single-run pass@1 scores are highly unreliable due to significant variance, often masking genuine progress. Small reported improvements may reflect evaluation noise. Reliable assessment requires multiple runs, statistical analysis, and metrics like pass@k.

🔹 Publication Date: Published on Feb 6

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2602.07150
• PDF: https://arxiv.org/pdf/2602.07150

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

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

#AIEvaluation #AgenticAI #MachineLearning #StatisticalMethods #AIResearch

✨Echoes as Anchors: Probabilistic Costs and Attention Refocusing in LLM Reasoning

📝 Summary:
This paper formalizes the Echo of Prompt EOP, spontaneous question repetition by LLMs, as a compute-shaping mechanism. It introduces Echo-Distilled SFT and Echoic Prompting to leverage EOP, improving reasoning accuracy and efficiency by refocusing attention.

🔹 Publication Date: Published on Feb 6

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2602.06600
• PDF: https://arxiv.org/pdf/2602.06600
• Github: https://github.com/hhh2210/echoes-as-anchors

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

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

#LLM #PromptEngineering #AIResearch #DeepLearning #AIAttention

✨AIRS-Bench: a Suite of Tasks for Frontier AI Research Science Agents

📝 Summary:
AIRS-Bench is a new benchmark of 20 scientific tasks evaluating AI agents across the full research lifecycle. Agents exceed human state-of-the-art in 4 tasks but largely fall short, highlighting significant room for improvement in autonomous scientific research. The suite is open-sourced to accel...

🔹 Publication Date: Published on Feb 6

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2602.06855
• PDF: https://arxiv.org/pdf/2602.06855
• Github: https://github.com/facebookresearch/airs-bench

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

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

#AIagents #ScientificResearch #AIBenchmark #FrontierAI #AutonomousResearch

✨Fundamental Reasoning Paradigms Induce Out-of-Domain Generalization in Language Models

📝 Summary:
This study explores how fundamental reasoning paradigms deduction induction and abduction influence LLM generalization. By training LLMs on a new dataset of symbolic reasoning trajectories, the research shows substantial performance gains and strong generalizability on realistic out-of-domain tasks.

🔹 Publication Date: Published on Feb 9

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2602.08658
• PDF: https://arxiv.org/pdf/2602.08658
• Github: https://github.com/voalmciaf/FR-OOD

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

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

#LLM #AI #MachineLearning #Reasoning #Generalization

✨Data Science and Technology Towards AGI Part I: Tiered Data Management

📝 Summary:
This paper proposes an LLM-guided, tiered data management framework L0-L4 to optimize data quality, acquisition cost, and training efficiency. This systematic approach, used across LLM development stages, significantly improves model performance and sustainability.

🔹 Publication Date: Published on Feb 9

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2602.09003
• PDF: https://arxiv.org/pdf/2602.09003
• Project Page: https://ultradata.openbmb.cn/
• Github: https://github.com/UltraData-OpenBMB/UltraData-Math

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

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

#DataScience #LLM #AGI #DataManagement #AIResearch