✨mHC-lite: You Don't Need 20 Sinkhorn-Knopp Iterations

📝 Summary:
mHC-lite proposes a novel reparameterization for Hyper-Connections, explicitly constructing exactly doubly stochastic matrices via convex combinations of permutations. This approach guarantees stability, improves training throughput with native operations, and outperforms prior methods.

🔹 Publication Date: Published on Jan 9

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2601.05732
• PDF: https://arxiv.org/pdf/2601.05732
• Github: https://github.com/FFTYYY/mhc-lite

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

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

#DeepLearning #MachineLearning #Optimization #Algorithm #AI

✨Benchmarking Small Language Models and Small Reasoning Language Models on System Log Severity Classification

📝 Summary:
Severity classification benchmarks small language models for log understanding and deployability. RAG significantly boosts many models, even tiny ones, but efficiency and RAG integration vary widely, crucial for real-time systems.

🔹 Publication Date: Published on Jan 12

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2601.07790
• PDF: https://arxiv.org/pdf/2601.07790
• Github: https://github.com/stccenter/Benchmarking-SLMs-and-SRLMs-on-System-Log-Severity-Classification

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

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

#AI #DataScience #MachineLearning #HuggingFace #Research

✨RealMem: Benchmarking LLMs in Real-World Memory-Driven Interaction

📝 Summary:
RealMem benchmark evaluates memory systems for long-term project-oriented interactions in large language models, revealing challenges in managing dynamic context dependencies. AI-generated summary As ...

🔹 Publication Date: Published on Jan 11

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2601.06966
• PDF: https://arxiv.org/pdf/2601.06966

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

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

#AI #DataScience #MachineLearning #HuggingFace #Research

✨Sci-Reasoning: A Dataset Decoding AI Innovation Patterns

📝 Summary:
Sci-Reasoning is a new dataset that maps intellectual synthesis patterns in AI research. It traces key papers to their predecessors, identifying 15 distinct thinking patterns that drive breakthroughs. This dataset enables quantitative study of scientific progress and trains next-generation AI res...

🔹 Publication Date: Published on Jan 8

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2601.04577
• PDF: https://arxiv.org/pdf/2601.04577
• Github: https://github.com/AmberLJC/Sci-Reasoning

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

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

#AI #DataScience #MachineLearning #HuggingFace #Research

✨Does Inference Scaling Improve Reasoning Faithfulness? A Multi-Model Analysis of Self-Consistency Tradeoffs

📝 Summary:
Self-consistency improves reasoning accuracy for some models while potentially sacrificing faithfulness, with varying effects across different language models and problem difficulties. AI-generated su...

🔹 Publication Date: Published on Jan 10

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2601.06423
• PDF: https://arxiv.org/pdf/2601.06423

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

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

#AI #DataScience #MachineLearning #HuggingFace #Research

✨Can Textual Reasoning Improve the Performance of MLLMs on Fine-grained Visual Classification?

📝 Summary:
Multi-modal large language models struggle with fine-grained visual classification, and chain-of-thought reasoning harms performance due to increased reasoning length; a new framework called ReFine-RF...

🔹 Publication Date: Published on Jan 11

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2601.06993
• PDF: https://arxiv.org/pdf/2601.06993
• Github: https://github.com/jiezhu23/ReFine-RFT

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

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

#AI #DataScience #MachineLearning #HuggingFace #Research

✨Stochastic CHAOS: Why Deterministic Inference Kills, and Distributional Variability Is the Heartbeat of Artifical Cognition

📝 Summary:
Deterministic inference in LLMs is detrimental, suppressing uncertainty, emergent abilities, and safety awareness by enforcing single-output predictions. This approach misrepresents capabilities and risks. The paper advocates embracing distributional variability as essential for artificial cognit...

🔹 Publication Date: Published on Jan 12

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2601.07239
• PDF: https://arxiv.org/pdf/2601.07239

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

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

#AI #DataScience #MachineLearning #HuggingFace #Research

✨A Rising Tide Lifts All Boats: MTQE Rewards for Idioms Improve General Translation Quality

📝 Summary:
GRPO-style fine-tuning with MTQE models as rewards improves idiom translation by 14 points while enhancing general translation and cross-lingual capabilities. AI-generated summary Non-compositional ex...

🔹 Publication Date: Published on Jan 9

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2601.06307
• PDF: https://arxiv.org/pdf/2601.06307

🔹 Models citing this paper:
• https://huggingface.co/ishikaa/Chinese_llama8b-da
• https://huggingface.co/ishikaa/Chinese_llama8b-qe-cons
• https://huggingface.co/ishikaa/Chinese_llama8b-qe-pos

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

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

#AI #DataScience #MachineLearning #HuggingFace #Research

✨SPINAL -- Scaling-law and Preference Integration in Neural Alignment Layers

📝 Summary:
SPINAL diagnoses how DPO alignment reshapes representations layer by layer, revealing geometric localization of preference gradients in final decoder blocks and enabling practical auditing of alignmen...

🔹 Publication Date: Published on Jan 8

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2601.06238
• PDF: https://arxiv.org/pdf/2601.06238

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

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

#AI #DataScience #MachineLearning #HuggingFace #Research

✨Artificial Entanglement in the Fine-Tuning of Large Language Models

📝 Summary:
Using Artificial Entanglement, this paper finds that LLM fine-tuning like LoRA creates distinct internal parameter entanglement. Yet, external attention outputs are robust and similar to full fine-tuning. This no hair property explains LoRAs effectiveness.

🔹 Publication Date: Published on Jan 11

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2601.06788
• PDF: https://arxiv.org/pdf/2601.06788

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

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

#AI #DataScience #MachineLearning #HuggingFace #Research

✨How Do Large Language Models Learn Concepts During Continual Pre-Training?

📝 Summary:
Large language models develop concept circuits during continual pretraining that exhibit learning and forgetting patterns, with semantically similar concepts showing stronger interference and varying ...

🔹 Publication Date: Published on Jan 7

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2601.03570
• PDF: https://arxiv.org/pdf/2601.03570

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

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

#AI #DataScience #MachineLearning #HuggingFace #Research

✨On the Non-decoupling of Supervised Fine-tuning and Reinforcement Learning in Post-training

📝 Summary:
Supervised fine-tuning SFT and reinforcement learning RL in large language model post-training cannot be decoupled. Separating them causes performance degradation because RL increases SFT loss, and SFT lowers RL reward.

🔹 Publication Date: Published on Jan 12

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2601.07389
• PDF: https://arxiv.org/pdf/2601.07389

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

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

#AI #DataScience #MachineLearning #HuggingFace #Research

✨Ministral 3

📝 Summary:
Ministral 3 is a series of parameter-efficient dense language models available in three sizes 3B, 8B, 14B with three variants each. Designed for compute-constrained applications, they are trained via Cascade Distillation and include image understanding capabilities.

🔹 Publication Date: Published on Jan 13

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2601.08584
• PDF: https://arxiv.org/pdf/2601.08584

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

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

#AI #DataScience #MachineLearning #HuggingFace #Research

✨End-to-End Video Character Replacement without Structural Guidance

📝 Summary:
MoCha enables controllable video character replacement using a single frame mask through condition-aware RoPE and a comprehensive data construction pipeline with specialized datasets. AI-generated sum...

🔹 Publication Date: Published on Jan 13

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2601.08587
• PDF: https://arxiv.org/pdf/2601.08587

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

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

#AI #DataScience #MachineLearning #HuggingFace #Research

✨JudgeRLVR: Judge First, Generate Second for Efficient Reasoning

📝 Summary:
Reinforcement learning with verifiable rewards is enhanced through a judge-then-generate paradigm that improves both efficiency and accuracy in mathematical problem-solving. AI-generated summary Reinf...

🔹 Publication Date: Published on Jan 13

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2601.08468
• PDF: https://arxiv.org/pdf/2601.08468

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

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

#AI #DataScience #MachineLearning #HuggingFace #Research

✨The Confidence Dichotomy: Analyzing and Mitigating Miscalibration in Tool-Use Agents

📝 Summary:
Tool-integrated language model agents exhibit different calibration behaviors based on tool type, with a reinforcement learning framework improving both task accuracy and reliable uncertainty estimati...

🔹 Publication Date: Published on Jan 12

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2601.07264
• PDF: https://arxiv.org/pdf/2601.07264

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

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

#AI #DataScience #MachineLearning #HuggingFace #Research

✨ArenaRL: Scaling RL for Open-Ended Agents via Tournament-based Relative Ranking

📝 Summary:
Reinforcement learning for large language model agents suffers from discrimination collapse in open-ended tasks due to pointwise scalar scoring, which ArenaRL addresses through relative ranking and pa...

🔹 Publication Date: Published on Jan 10

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2601.06487
• PDF: https://arxiv.org/pdf/2601.06487
• Github: https://github.com/Alibaba-NLP/qqr

✨ Datasets citing this paper:
• https://huggingface.co/datasets/Alibaba-NLP/Open-Travel
• https://huggingface.co/datasets/Alibaba-NLP/Open-DeepResearch

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

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

#AI #DataScience #MachineLearning #HuggingFace #Research

✨Motion Attribution for Video Generation

📝 Summary:
Motive is a gradient-based data attribution framework that identifies influential video clips for motion improvement in text-to-video models through motion-weighted loss masking. AI-generated summary ...

🔹 Publication Date: Published on Jan 13

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2601.08828
• PDF: https://arxiv.org/pdf/2601.08828
• Project Page: https://research.nvidia.com/labs/sil/projects/MOTIVE/

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

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

#AI #DataScience #MachineLearning #HuggingFace #Research

✨SnapGen++: Unleashing Diffusion Transformers for Efficient High-Fidelity Image Generation on Edge Devices

📝 Summary:
An efficient diffusion transformer framework for mobile and edge devices that maintains high-generation quality while reducing computational costs through compact architecture, elastic training, and k...

🔹 Publication Date: Published on Jan 13

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2601.08303
• PDF: https://arxiv.org/pdf/2601.08303

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

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

#AI #DataScience #MachineLearning #HuggingFace #Research

✨Aligning Text, Code, and Vision: A Multi-Objective Reinforcement Learning Framework for Text-to-Visualization

📝 Summary:
A reinforcement learning framework for text-to-visualization generation that improves chart quality and code execution by optimizing multiple objectives using post-execution feedback. AI-generated sum...

🔹 Publication Date: Published on Jan 8

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2601.04582
• PDF: https://arxiv.org/pdf/2601.04582

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

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

#AI #DataScience #MachineLearning #HuggingFace #Research

✨VLingNav: Embodied Navigation with Adaptive Reasoning and Visual-Assisted Linguistic Memory

📝 Summary:
VLingNav enhances embodied navigation through linguistic-driven cognition with adaptive reasoning and visual-assisted memory, achieving state-of-the-art performance and zero-shot transfer to real robo...

🔹 Publication Date: Published on Jan 13

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2601.08665
• PDF: https://arxiv.org/pdf/2601.08665
• Project Page: https://wsakobe.github.io/VLingNav-web/

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

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

#AI #DataScience #MachineLearning #HuggingFace #Research