✨Shape of Thought: When Distribution Matters More than Correctness in Reasoning Tasks
📝 Summary:
We present the surprising finding that a language model's reasoning capabilities can be improved by training on synthetic datasets of chain-of-thought (CoT) traces from more capable models, even when ...
🔹 Publication Date: Published on Dec 24
🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2512.22255
• PDF: https://arxiv.org/pdf/2512.22255
==================================
For more data science resources:
✓ https://t.me/DataScienceT
#AI #DataScience #MachineLearning #HuggingFace #Research
📝 Summary:
We present the surprising finding that a language model's reasoning capabilities can be improved by training on synthetic datasets of chain-of-thought (CoT) traces from more capable models, even when ...
🔹 Publication Date: Published on Dec 24
🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2512.22255
• PDF: https://arxiv.org/pdf/2512.22255
==================================
For more data science resources:
✓ https://t.me/DataScienceT
#AI #DataScience #MachineLearning #HuggingFace #Research
✨Self-Evaluation Unlocks Any-Step Text-to-Image Generation
📝 Summary:
Self-E is a novel self-evaluating text-to-image model trained from scratch that supports any-step generation and combines local learning with self-driven global matching to achieve high quality even a...
🔹 Publication Date: Published on Dec 26
🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2512.22374
• PDF: https://arxiv.org/pdf/2512.22374
• Project Page: https://xinyu-andy.github.io/SelfE-project/
==================================
For more data science resources:
✓ https://t.me/DataScienceT
#AI #DataScience #MachineLearning #HuggingFace #Research
📝 Summary:
Self-E is a novel self-evaluating text-to-image model trained from scratch that supports any-step generation and combines local learning with self-driven global matching to achieve high quality even a...
🔹 Publication Date: Published on Dec 26
🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2512.22374
• PDF: https://arxiv.org/pdf/2512.22374
• Project Page: https://xinyu-andy.github.io/SelfE-project/
==================================
For more data science resources:
✓ https://t.me/DataScienceT
#AI #DataScience #MachineLearning #HuggingFace #Research
✨LeVo: High-Quality Song Generation with Multi-Preference Alignment
📝 Summary:
LeVo enhances lyrics-to-song generation. It uses an LM to parallelly model mixed and dual-track audio tokens for vocal-instrument harmony and sound quality. Direct Preference Optimization improves musicality and instruction following, outperforming existing methods.
🔹 Publication Date: Published on Jun 9
🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2506.07520
• PDF: https://arxiv.org/pdf/2506.07520
• Project Page: https://levo-demo.github.io/
• Github: https://github.com/tencent-ailab/songgeneration
🔹 Models citing this paper:
• https://huggingface.co/tencent/SongGeneration
• https://huggingface.co/waytan22/SongGeneration-v1.5-beta
• https://huggingface.co/chaitnya26/SongGeneration-fork
✨ Spaces citing this paper:
• https://huggingface.co/spaces/tencent/SongGeneration
• https://huggingface.co/spaces/NeoPy/SongGeneration
• https://huggingface.co/spaces/Open-Hat-Lab/Song-Generator
==================================
For more data science resources:
✓ https://t.me/DataScienceT
#AI #DataScience #MachineLearning #HuggingFace #Research
📝 Summary:
LeVo enhances lyrics-to-song generation. It uses an LM to parallelly model mixed and dual-track audio tokens for vocal-instrument harmony and sound quality. Direct Preference Optimization improves musicality and instruction following, outperforming existing methods.
🔹 Publication Date: Published on Jun 9
🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2506.07520
• PDF: https://arxiv.org/pdf/2506.07520
• Project Page: https://levo-demo.github.io/
• Github: https://github.com/tencent-ailab/songgeneration
🔹 Models citing this paper:
• https://huggingface.co/tencent/SongGeneration
• https://huggingface.co/waytan22/SongGeneration-v1.5-beta
• https://huggingface.co/chaitnya26/SongGeneration-fork
✨ Spaces citing this paper:
• https://huggingface.co/spaces/tencent/SongGeneration
• https://huggingface.co/spaces/NeoPy/SongGeneration
• https://huggingface.co/spaces/Open-Hat-Lab/Song-Generator
==================================
For more data science resources:
✓ https://t.me/DataScienceT
#AI #DataScience #MachineLearning #HuggingFace #Research
arXiv.org
LeVo: High-Quality Song Generation with Multi-Preference Alignment
Recent advances in large language models (LLMs) and audio language models have significantly improved music generation, particularly in lyrics-to-song generation. However, existing approaches...
✨DreamOmni3: Scribble-based Editing and Generation
📝 Summary:
DreamOmni3 introduces scribble-based editing and generation for more flexible image creation beyond text prompts. It proposes new tasks, data synthesis, and a joint input scheme using colored scribbles on source images for precise localization and complex edits.
🔹 Publication Date: Published on Dec 27
🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2512.22525
• PDF: https://arxiv.org/pdf/2512.22525
• Github: https://github.com/dvlab-research/DreamOmni3
==================================
For more data science resources:
✓ https://t.me/DataScienceT
#AI #DataScience #MachineLearning #HuggingFace #Research
📝 Summary:
DreamOmni3 introduces scribble-based editing and generation for more flexible image creation beyond text prompts. It proposes new tasks, data synthesis, and a joint input scheme using colored scribbles on source images for precise localization and complex edits.
🔹 Publication Date: Published on Dec 27
🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2512.22525
• PDF: https://arxiv.org/pdf/2512.22525
• Github: https://github.com/dvlab-research/DreamOmni3
==================================
For more data science resources:
✓ https://t.me/DataScienceT
#AI #DataScience #MachineLearning #HuggingFace #Research
✨End-to-End Test-Time Training for Long Context
📝 Summary:
This paper introduces End-to-End Test-Time Training TTT-E2E for long-context language models. It uses a standard Transformer that continually learns from context at test time, compressing information into its weights. TTT-E2E scales well with context length and offers constant inference latency, ...
🔹 Publication Date: Published on Dec 29
🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2512.23675
• PDF: https://arxiv.org/pdf/2512.23675
• Github: https://github.com/test-time-training/e2e
==================================
For more data science resources:
✓ https://t.me/DataScienceT
#AI #DataScience #MachineLearning #HuggingFace #Research
📝 Summary:
This paper introduces End-to-End Test-Time Training TTT-E2E for long-context language models. It uses a standard Transformer that continually learns from context at test time, compressing information into its weights. TTT-E2E scales well with context length and offers constant inference latency, ...
🔹 Publication Date: Published on Dec 29
🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2512.23675
• PDF: https://arxiv.org/pdf/2512.23675
• Github: https://github.com/test-time-training/e2e
==================================
For more data science resources:
✓ https://t.me/DataScienceT
#AI #DataScience #MachineLearning #HuggingFace #Research
✨GraphLocator: Graph-guided Causal Reasoning for Issue Localization
📝 Summary:
The issue localization task aims to identify the locations in a software repository that requires modification given a natural language issue description. This task is fundamental yet challenging in a...
🔹 Publication Date: Published on Dec 27
🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2512.22469
• PDF: https://arxiv.org/pdf/2512.22469
==================================
For more data science resources:
✓ https://t.me/DataScienceT
#AI #DataScience #MachineLearning #HuggingFace #Research
📝 Summary:
The issue localization task aims to identify the locations in a software repository that requires modification given a natural language issue description. This task is fundamental yet challenging in a...
🔹 Publication Date: Published on Dec 27
🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2512.22469
• PDF: https://arxiv.org/pdf/2512.22469
==================================
For more data science resources:
✓ https://t.me/DataScienceT
#AI #DataScience #MachineLearning #HuggingFace #Research
Media is too big
VIEW IN TELEGRAM
✨GR-Dexter Technical Report
📝 Summary:
GR-Dexter introduces a hardware-model-data framework for bimanual dexterous-hand robot manipulation using VLA models. It combines a new 21-DoF hand, teleoperation for data, and diverse datasets. This framework achieves strong performance and robust generalization in real-world manipulation tasks.
🔹 Publication Date: Published on Dec 30, 2025
🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2512.24210
• PDF: https://arxiv.org/pdf/2512.24210
• Project Page: https://byte-dexter.github.io/gr-dexter/
==================================
For more data science resources:
✓ https://t.me/DataScienceT
#Robotics #DexterousManipulation #VLA #RobotHardware #MachineLearning
📝 Summary:
GR-Dexter introduces a hardware-model-data framework for bimanual dexterous-hand robot manipulation using VLA models. It combines a new 21-DoF hand, teleoperation for data, and diverse datasets. This framework achieves strong performance and robust generalization in real-world manipulation tasks.
🔹 Publication Date: Published on Dec 30, 2025
🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2512.24210
• PDF: https://arxiv.org/pdf/2512.24210
• Project Page: https://byte-dexter.github.io/gr-dexter/
==================================
For more data science resources:
✓ https://t.me/DataScienceT
#Robotics #DexterousManipulation #VLA #RobotHardware #MachineLearning
✨PhyGDPO: Physics-Aware Groupwise Direct Preference Optimization for Physically Consistent Text-to-Video Generation
📝 Summary:
Recent advances in text-to-video (T2V) generation have achieved good visual quality, yet synthesizing videos that faithfully follow physical laws remains an open challenge. Existing methods mainly bas...
🔹 Publication Date: Published on Dec 31, 2025
🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2512.24551
• PDF: https://arxiv.org/pdf/2512.24551
• Project Page: https://caiyuanhao1998.github.io/project/PhyGDPO/
• Github: https://github.com/caiyuanhao1998/Open-PhyGDPO
==================================
For more data science resources:
✓ https://t.me/DataScienceT
#AI #DataScience #MachineLearning #HuggingFace #Research
📝 Summary:
Recent advances in text-to-video (T2V) generation have achieved good visual quality, yet synthesizing videos that faithfully follow physical laws remains an open challenge. Existing methods mainly bas...
🔹 Publication Date: Published on Dec 31, 2025
🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2512.24551
• PDF: https://arxiv.org/pdf/2512.24551
• Project Page: https://caiyuanhao1998.github.io/project/PhyGDPO/
• Github: https://github.com/caiyuanhao1998/Open-PhyGDPO
==================================
For more data science resources:
✓ https://t.me/DataScienceT
#AI #DataScience #MachineLearning #HuggingFace #Research
✨Scaling Open-Ended Reasoning to Predict the Future
📝 Summary:
This work trains language models for open-ended future prediction using a new dataset synthesized from news. Their OpenForecaster 8B model matches larger proprietary models in accuracy, calibration, and consistency. All resources are open-sourced.
🔹 Publication Date: Published on Dec 31, 2025
🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2512.25070
• PDF: https://arxiv.org/pdf/2512.25070
• Project Page: https://www.openforecaster.github.io
• Github: https://github.com/OpenForecaster/scaling-forecasting-training
==================================
For more data science resources:
✓ https://t.me/DataScienceT
#LLMs #FuturePrediction #AI #OpenSourceAI #MachineLearning
📝 Summary:
This work trains language models for open-ended future prediction using a new dataset synthesized from news. Their OpenForecaster 8B model matches larger proprietary models in accuracy, calibration, and consistency. All resources are open-sourced.
🔹 Publication Date: Published on Dec 31, 2025
🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2512.25070
• PDF: https://arxiv.org/pdf/2512.25070
• Project Page: https://www.openforecaster.github.io
• Github: https://github.com/OpenForecaster/scaling-forecasting-training
==================================
For more data science resources:
✓ https://t.me/DataScienceT
#LLMs #FuturePrediction #AI #OpenSourceAI #MachineLearning
✨Fantastic Reasoning Behaviors and Where to Find Them: Unsupervised Discovery of the Reasoning Process
📝 Summary:
This paper introduces RISE, an unsupervised framework using sparse auto-encoders to discover and control LLM reasoning behaviors. It identifies interpretable reasoning vectors like reflection and backtracking, enabling targeted interventions and discovery of novel behaviors without retraining.
🔹 Publication Date: Published on Dec 30, 2025
🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2512.23988
• PDF: https://arxiv.org/pdf/2512.23988
==================================
For more data science resources:
✓ https://t.me/DataScienceT
#LLM #AI #MachineLearning #AIReasoning #Interpretability
📝 Summary:
This paper introduces RISE, an unsupervised framework using sparse auto-encoders to discover and control LLM reasoning behaviors. It identifies interpretable reasoning vectors like reflection and backtracking, enabling targeted interventions and discovery of novel behaviors without retraining.
🔹 Publication Date: Published on Dec 30, 2025
🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2512.23988
• PDF: https://arxiv.org/pdf/2512.23988
==================================
For more data science resources:
✓ https://t.me/DataScienceT
#LLM #AI #MachineLearning #AIReasoning #Interpretability