✨RouteMoA: Dynamic Routing without Pre-Inference Boosts Efficient Mixture-of-Agents

📝 Summary:
RouteMoA reduces computational costs and latency in mixture-of-agents frameworks by using dynamic routing with lightweight scoring and judgment mechanisms. AI-generated summary Mixture-of-Agents (MoA)...

🔹 Publication Date: Published on Jan 26

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2601.18130
• PDF: https://arxiv.org/pdf/2601.18130

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

For more data science resources:
✓ https://news.1rj.ru/str/DataScienceT

#AI #DataScience #MachineLearning #HuggingFace #Research

✨SkyReels-V3 Technique Report

📝 Summary:
SkyReels-V3 is a unified multimodal video generation model that supports reference image-to-video, video-to-video extension, and audio-guided video generation through diffusion Transformers and in-con...

🔹 Publication Date: Published on Jan 24

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2601.17323
• PDF: https://arxiv.org/pdf/2601.17323

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

For more data science resources:
✓ https://news.1rj.ru/str/DataScienceT

#AI #DataScience #MachineLearning #HuggingFace #Research

✨UI Remix: Supporting UI Design Through Interactive Example Retrieval and Remixing

📝 Summary:
UI Remix is an interactive system that supports mobile UI design through example-driven workflows using a multimodal retrieval-augmented generation model, enabling iterative design adaptation with sou...

🔹 Publication Date: Published on Jan 26

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2601.18759
• PDF: https://arxiv.org/pdf/2601.18759

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

For more data science resources:
✓ https://news.1rj.ru/str/DataScienceT

#AI #DataScience #MachineLearning #HuggingFace #Research

✨SAGE: Steerable Agentic Data Generation for Deep Search with Execution Feedback

📝 Summary:
Deep search agents trained on synthetic question-answer pairs generated through an iterative agent-based pipeline demonstrate improved performance and adaptability across different search environments...

🔹 Publication Date: Published on Jan 26

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2601.18202
• PDF: https://arxiv.org/pdf/2601.18202

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

For more data science resources:
✓ https://news.1rj.ru/str/DataScienceT

#AI #DataScience #MachineLearning #HuggingFace #Research

✨Agentic Very Long Video Understanding

📝 Summary:
An agentic framework using entity scene graphs enables long-horizon video understanding with structured search, temporal reasoning, and cross-modal capabilities for extended visual and audio interpret...

🔹 Publication Date: Published on Jan 26

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2601.18157
• PDF: https://arxiv.org/pdf/2601.18157

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

For more data science resources:
✓ https://news.1rj.ru/str/DataScienceT

#AI #DataScience #MachineLearning #HuggingFace #Research

✨One Adapts to Any: Meta Reward Modeling for Personalized LLM Alignment

📝 Summary:
Meta Reward Modeling reformulates personalized reward modeling as a meta-learning problem to enable efficient adaptation to individual users with limited feedback. AI-generated summary Alignment of La...

🔹 Publication Date: Published on Jan 26

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2601.18731
• PDF: https://arxiv.org/pdf/2601.18731
• Github: https://github.com/ModalityDance/MRM

🔹 Models citing this paper:
• https://huggingface.co/ModalityDance/MRM-Reddit150-V2
• https://huggingface.co/ModalityDance/MRM-Reddit100-V2
• https://huggingface.co/ModalityDance/MRM-Reddit150-V1

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

For more data science resources:
✓ https://news.1rj.ru/str/DataScienceT

#AI #DataScience #MachineLearning #HuggingFace #Research

✨daVinci-Dev: Agent-native Mid-training for Software Engineering

📝 Summary:
This paper introduces agentic mid-training for LLMs, bridging static data and dynamic development environments. Using agent-native data with contextually and environmentally native trajectories, it outperforms prior work on SWE-Bench Verified with fewer tokens.

🔹 Publication Date: Published on Jan 26

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2601.18418
• PDF: https://arxiv.org/pdf/2601.18418
• Github: https://github.com/GAIR-NLP/daVinci-Dev

🔹 Models citing this paper:
• https://huggingface.co/GAIR/daVinci-Dev-72B
• https://huggingface.co/GAIR/daVinci-Dev-32B-MT
• https://huggingface.co/GAIR/daVinci-Dev-72B-MT

✨ Datasets citing this paper:
• https://huggingface.co/datasets/GAIR/daVinci-Dev

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

For more data science resources:
✓ https://news.1rj.ru/str/DataScienceT

#AI #DataScience #MachineLearning #HuggingFace #Research

✨Paying Less Generalization Tax: A Cross-Domain Generalization Study of RL Training for LLM Agents

📝 Summary:
Research examines factors influencing out-of-domain performance in reinforcement learning agents, identifying state information richness and planning complexity as key determinants, while proposing a ...

🔹 Publication Date: Published on Jan 26

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2601.18217
• PDF: https://arxiv.org/pdf/2601.18217

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

For more data science resources:
✓ https://news.1rj.ru/str/DataScienceT

#AI #DataScience #MachineLearning #HuggingFace #Research

✨Teaching Models to Teach Themselves: Reasoning at the Edge of Learnability

📝 Summary:
A self-improvement framework enables pretrained language models to generate automated curricula for solving previously unsolvable problems by leveraging latent knowledge and meta-reinforcement learnin...

🔹 Publication Date: Published on Jan 26

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2601.18778
• PDF: https://arxiv.org/pdf/2601.18778
• Github: https://ssundaram21.github.io/soar/

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

For more data science resources:
✓ https://news.1rj.ru/str/DataScienceT

#AI #DataScience #MachineLearning #HuggingFace #Research

✨Diffusion In Diffusion: Reclaiming Global Coherence in Semi-Autoregressive Diffusion

📝 Summary:
A 'draft-then-refine' framework for discrete diffusion language models that recovers global contextual understanding while maintaining semi-autoregressive efficiency through block diffusion and global...

🔹 Publication Date: Published on Jan 20

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2601.13599
• PDF: https://arxiv.org/pdf/2601.13599
• Project Page: https://noah-dllm.github.io/
• Github: https://noah-dllm.github.io/

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

For more data science resources:
✓ https://news.1rj.ru/str/DataScienceT

#AI #DataScience #MachineLearning #HuggingFace #Research

✨The Side Effects of Being Smart: Safety Risks in MLLMs' Multi-Image Reasoning

📝 Summary:
A new benchmark, MIR-SafetyBench, reveals advanced MLLMs are more vulnerable to safety risks in multi-image reasoning. These models often give superficial safe responses, with unsafe generations showing lower attention entropy, indicating an over-focus on task completion.

🔹 Publication Date: Published on Jan 20

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2601.14127
• PDF: https://arxiv.org/pdf/2601.14127
• Github: https://github.com/thu-coai/MIR-SafetyBench

✨ Datasets citing this paper:
• https://huggingface.co/datasets/thu-coai/MIR-SafetyBench

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

For more data science resources:
✓ https://news.1rj.ru/str/DataScienceT

#AIsafety #MLLMs #MultiModalAI #ResponsibleAI #AIresearch

✨End-to-End Joint ASR and Speaker Role Diarization with Child-Adult Interactions

📝 Summary:
This paper presents a unified end-to-end framework extending Whisper for joint ASR and child-adult speaker role diarization. It significantly improves trannoscription accuracy and scalability by preventing error propagation, achieving lower word error rates and competitive diarization performance.

🔹 Publication Date: Published on Jan 25

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2601.17640
• PDF: https://arxiv.org/pdf/2601.17640
• Github: https://github.com/usc-sail/joint-asr-diarization-child-adult

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

For more data science resources:
✓ https://news.1rj.ru/str/DataScienceT

#ASR #SpeakerDiarization #SpeechProcessing #DeepLearning #ChildAdultInteraction

✨PingPong: A Natural Benchmark for Multi-Turn Code-Switching Dialogues

📝 Summary:
PingPong is a new human-authored benchmark for natural, multi-party code-switching dialogues, including trilingual conversations. It offers greater structural diversity than machine-generated data. Evaluations show current language models struggle with code-switched inputs, emphasizing the need f...

🔹 Publication Date: Published on Jan 24

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2601.17277
• PDF: https://arxiv.org/pdf/2601.17277

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

For more data science resources:
✓ https://news.1rj.ru/str/DataScienceT

#CodeSwitching #NLP #DialogueSystems #MultilingualAI #LLMs

✨Less Is More -- Until It Breaks: Security Pitfalls of Vision Token Compression in Large Vision-Language Models

📝 Summary:
Visual token compression degrades LVLM robustness via unstable token importance ranking. This causes critical information loss, creating vulnerabilities only under compression. An attack exploits this, revealing an efficiency-security trade-off.

🔹 Publication Date: Published on Jan 17

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2601.12042
• PDF: https://arxiv.org/pdf/2601.12042

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

For more data science resources:
✓ https://news.1rj.ru/str/DataScienceT

#LVLM #AIsecurity #VisionAI #ModelRobustness #DeepLearning

✨Plug-and-Play Benchmarking of Reinforcement Learning Algorithms for Large-Scale Flow Control

📝 Summary:
FluidGym presents a standalone, fully differentiable reinforcement learning benchmark for active flow control that operates without external CFD solvers and supports standardized evaluation protocols....

🔹 Publication Date: Published on Jan 21

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2601.15015v1
• PDF: https://arxiv.org/pdf/2601.15015
• Github: https://github.com/safe-autonomous-systems/fluidgym

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

For more data science resources:
✓ https://news.1rj.ru/str/DataScienceT

#AI #DataScience #MachineLearning #HuggingFace #Research

✨STAR: Semantic Table Representation with Header-Aware Clustering and Adaptive Weighted Fusion

📝 Summary:
STAR improves table representation for table retrieval tasks. It uses header-aware clustering to create diverse partial tables and generate cluster-specific queries. STAR then employs weighted fusion for fine-grained alignment, outperforming previous methods on benchmarks.

🔹 Publication Date: Published on Jan 22

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2601.15860
• PDF: https://arxiv.org/pdf/2601.15860
• Github: https://github.com/adsl135789/STAR

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

For more data science resources:
✓ https://news.1rj.ru/str/DataScienceT

#TableRepresentation #InformationRetrieval #Clustering #DataScience #MachineLearning

✨DeepPlanning: Benchmarking Long-Horizon Agentic Planning with Verifiable Constraints

📝 Summary:
DeepPlanning is a new benchmark for long-horizon agent planning, addressing the lack of global optimization and fine-grained local constraints in current LLM assessments. It features complex real-world tasks where even frontier LLMs struggle, highlighting the need for explicit reasoning and paral...

🔹 Publication Date: Published on Jan 26

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2601.18137
• PDF: https://arxiv.org/pdf/2601.18137

✨ Datasets citing this paper:
• https://huggingface.co/datasets/Qwen/DeepPlanning

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

For more data science resources:
✓ https://news.1rj.ru/str/DataScienceT

#AIPlanning #LLMs #AgentAI #Benchmarking #DeepLearning

✨A Mechanistic View on Video Generation as World Models: State and Dynamics

📝 Summary:
Video generation models are categorized based on state construction and dynamics modeling approaches, with emphasis on transitioning evaluation metrics from visual quality to functional capabilities l...

🔹 Publication Date: Published on Jan 22

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2601.17067
• PDF: https://arxiv.org/pdf/2601.17067

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

For more data science resources:
✓ https://news.1rj.ru/str/DataScienceT

#AI #DataScience #MachineLearning #HuggingFace #Research

✨TensorLens: End-to-End Transformer Analysis via High-Order Attention Tensors

📝 Summary:
TensorLens presents a novel mathematical framework that represents the complete transformer architecture as a single input-dependent linear operator using high-order tensors, enabling comprehensive an...

🔹 Publication Date: Published on Jan 25

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2601.17958
• PDF: https://arxiv.org/pdf/2601.17958

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

For more data science resources:
✓ https://news.1rj.ru/str/DataScienceT

#AI #DataScience #MachineLearning #HuggingFace #Research

✨HalluGuard: Demystifying Data-Driven and Reasoning-Driven Hallucinations in LLMs

📝 Summary:
HalluGuard presents a theoretical framework that decomposes LLM hallucination risk into data-driven and reasoning-driven components. It introduces an NTK-based score to jointly detect both types of hallucinations, achieving state-of-the-art performance across various benchmarks and LLMs.

🔹 Publication Date: Published on Jan 26

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2601.18753
• PDF: https://arxiv.org/pdf/2601.18753

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

For more data science resources:
✓ https://news.1rj.ru/str/DataScienceT

#LLMs #AI #MachineLearning #Hallucination #NLP