Data science research papers
Kanalga Telegramβda oβtish
Machine learning and data science research papers Key ML and AI papers with code and GitHub repos. Simple way to follow current research. Join π https://rebrand.ly/bigdatachannels DMCA: @disclosure_bds Contact: @mldatascientist
Ko'proq ko'rsatish
Tarmoq:3 154
Obunachilar
+624 soatlar
+417 kun
+13030 kun
Ma'lumot yuklanmoqda...
O'xshash kanallar
Taglar buluti
Kirish va chiqish esdaliklari
---
---
---
---
---
---
Obunachilarni jalb qilish
Sentabr '26
Sentabr '26
+18
0 kanalda
Avgust '26
+171
0 kanalda
Get PRO
Iyul '26
+149
0 kanalda
Get PRO
Iyun '26
+97
0 kanalda
Get PRO
May '26
+104
1 kanalda
Get PRO
Aprel '26
+114
0 kanalda
Get PRO
Mart '26
+101
0 kanalda
Get PRO
Fevral '26
+82
0 kanalda
Get PRO
Yanvar '26
+118
9 kanalda
Get PRO
Dekabr '25
+115
0 kanalda
Get PRO
Noyabr '25
+112
0 kanalda
Get PRO
Oktabr '25
+43
0 kanalda
Get PRO
Sentabr '25
+7
0 kanalda
Get PRO
Avgust '25
+3
0 kanalda
Get PRO
Iyul '25
+3
0 kanalda
Get PRO
Iyun '25
+2
0 kanalda
Get PRO
May '25
+3
0 kanalda
Get PRO
Aprel '25
+15
0 kanalda
Get PRO
Mart '25
+112
0 kanalda
Get PRO
Fevral '25
+153
0 kanalda
Get PRO
Yanvar '25
+187
0 kanalda
Get PRO
Dekabr '24
+179
0 kanalda
Get PRO
Noyabr '24
+165
0 kanalda
Get PRO
Oktabr '24
+136
0 kanalda
Get PRO
Sentabr '24
+108
0 kanalda
Get PRO
Avgust '24
+114
0 kanalda
Get PRO
Iyul '24
+139
0 kanalda
Get PRO
Iyun '24
+115
0 kanalda
Get PRO
May '24
+132
1 kanalda
Get PRO
Aprel '24
+109
0 kanalda
Get PRO
Mart '24
+146
0 kanalda
Get PRO
Fevral '24
+183
0 kanalda
Get PRO
Yanvar '24
+228
0 kanalda
Get PRO
Dekabr '23
+171
1 kanalda
Get PRO
Noyabr '23
+28
0 kanalda
Get PRO
Oktabr '23
+28
0 kanalda
Get PRO
Sentabr '23
+504
0 kanalda
| Sana | Obunachilarni jalb qilish | Esdaliklar | Kanallar | |
| 03 Sentabr | +3 | |||
| 02 Sentabr | +7 | |||
| 01 Sentabr | +8 |
Kanal postlari
RAGU: A Multi-Step GraphRAG Engine with a Compact Domain-Adapted LLM
π
Publication Date: Jul 13, 2026
πPaper PDF: https://arxiv.org/pdf/2607.11683.pdf
π Code:https://github.com/huggingface
| 2 |  OPSD-V: On-Policy Self-Distillation for Post-Training Few-Step Autoregressive Video Generators
π
Publication Date: Jul 9, 2026
π Paper: https://arxiv.org/pdf/2607.08766
π» Project Page: https://meigen-ai.github.io/OPSD-V/
π Description:
The paper proposes a method called On-Policy Self-Distillation for Post-Training Few-Step Autoregressive Video Generators, or OPSD-V, which aims to improve the quality of videos generated by few-step autoregressive video diffusion models. The problem with existing models is that they can produce long videos with low latency, but the quality of the video degrades over time due to error accumulation and weakened motion dynamics.
#AutoregressiveVideoGeneration #VideoDiffusionModels #PostTrainingOptimization | 126 |
| 3 |  Does VLA Even Know the Basics? Measuring Commonsense and World Knowledge Retention in Vision-Language-Action Models
π
Publication Date: Jun 17, 2026
π Paper: https://arxiv.org/pdf/2606.19297.pdf
π Code: N/A
π Description:
Act2Answer protocol evaluates embodied vision-language-action models by having agents answer questions through physical actions, revealing knowledge retention and generalization patterns across different semantic categories. | 189 |
| 4 |  Scaling the Horizon, Not the Parameters: Reaching Trillion-Parameter Performance with a 35B Agent
π
Publication Date: Jun 29, 2026
π Paper: https://arxiv.org/pdf/2606.30616.pdf
π Code: N/A
π Description:
Agents-A1, a 35B Mixture-of-Experts Agentic Model, achieves trillion-parameter-level performance through long-horizon trajectory scaling and heterogeneous agent ability scaling via a three-stage training approach involving supervised fine-tuning, domain-level teacher models, and multi-teacher distil... | 241 |
| 5 |  PhysisForcing: Physics Reinforced World Simulator for Robotic Manipulation
π
Publication Date: Jun 26, 2026
π Paper: https://arxiv.org/pdf/2606.28128.pdf
π Code: https://github.com/huggingface
π Description:
PhysisForcing enhances embodied video generation by enforcing physical consistency through pixel-level trajectory alignment and semantic-level relational alignment losses in a DiT-based framework. | 271 |
| 6 |  LiveEdit: Towards Real-Time Diffusion-Based Streaming Video Editing
π
Publication Date: Jun 25, 2026
π Paper: https://arxiv.org/pdf/2606.26740.pdf
π Code: N/A
π Description:
A novel streaming video editing framework enables causal, frame-by-frame editing with stable long-horizon preservation and real-time responsiveness through a three-stage distillation pipeline and AR-oriented mask cache. | 285 |
| 7 |  BlockPilot: Instance-Adaptive Policy Learning for Diffusion-based Speculative Decoding
π
Publication Date: Jun 30, 2026
π Paper: https://arxiv.org/pdf/2606.31315.pdf
π Code: N/A
π Description:
Speculative decoding with adaptive block size selection improves inference efficiency by predicting optimal block sizes from prefilling representations, achieving significant speedup with minimal overhead. | 313 |
| 8 |  DOPD: Dual On-policy Distillation
π
Publication Date: Jun 29, 2026
π Paper: https://arxiv.org/pdf/2606.30626.pdf
π Code: https://github.com/huggingface
π Description:
DOPD addresses privilege illusion in on-policy distillation by dynamically routing token-level supervision between teacher and student policies based on advantage gaps and probabilities, improving capability transfer in large and vision-language models. | 332 |
| 9 |  Dockerless: Environment-Free Program Verifier for Coding Agents
π
Publication Date: Jun 26, 2026
π Paper: https://arxiv.org/pdf/2606.28436.pdf
π Code: https://github.com/huggingface
π Description:
A Dockerless environment-free agentic patch verifier improves code patch evaluation accuracy and enables effective post-training without execution-based verification costs. | 366 |
| 10 |  Agentic Abstention: Do Agents Know When to Stop Instead of Act?
π
Publication Date: Jun 27, 2026
π Paper: https://arxiv.org/pdf/2606.28733.pdf
π Code: https://github.com/huggingface
π Description:
Agentic abstention involves determining when an AI agent should cease interaction under uncertainty, requiring sequential decision-making across multiple environments and task types. | 392 |
| 11 |  Orca: The World is in Your Mind
π
Publication Date: Jun 29, 2026
π Paper: https://arxiv.org/pdf/2606.30534.pdf
π Code: https://github.com/huggingface
π Description:
Orca establishes a unified world latent space through next-state-prediction modeling using multimodal data and demonstrates superior performance in downstream tasks compared to specialized baselines. | 391 |
| 12 |  Grouped Query Experts: Mixture-of-Experts on GQA Self-Attention
π
Publication Date: Jun 18, 2026
π Paper: https://arxiv.org/pdf/2606.20945.pdf
π Code: https://github.com/huggingface
π Description:
Grouped Query Experts (GQE) improves Transformer efficiency by selectively activating query heads based on token content while maintaining key-value cache benefits of grouped-query attention. | 400 |
| 13 |  PlanBench-XL: Evaluating Long-Horizon Planning of LLM Tool-Use Agents in Large-Scale Tool Ecosystems
π
Publication Date: Jun 21, 2026
π Paper: https://arxiv.org/pdf/2606.22388.pdf
π Code: https://github.com/huggingface
π Description:
PlanBench-XL evaluates large language model agents' ability to plan and adapt in complex tool-rich environments with limited visibility and dynamic disruptions. | 421 |
| 14 |  Are We Ready For An Agent-Native Memory System?
π
Publication Date: Jun 23, 2026
π Paper: https://arxiv.org/pdf/2606.24775.pdf
π Code: https://github.com/huggingface
π Description:
Large language model agents' memory systems have evolved into complex data management frameworks requiring systematic evaluation across multiple modules and workloads to understand their performance characteristics and trade-offs. | 425 |
| 15 |  EnterpriseClawBench: Benchmarking Agents from Real Workplace Sessions
π
Publication Date: Jun 22, 2026
π Paper: https://arxiv.org/pdf/2606.23654.pdf
π Code: https://github.com/huggingface
π Description:
EnterpriseClawBench presents a benchmark for enterprise agents based on real-world sessions with 852 reproducible tasks, emphasizing comprehensive evaluation metrics beyond single performance scores. | 400 |
| 16 |  OpenRath: Session-Centered Runtime State for Agent Systems
π
Publication Date: Jun 17, 2026
π Paper: https://arxiv.org/pdf/2606.19409.pdf
π Code: https://github.com/huggingface
π Description:
OpenRath introduces a PyTorch-like programming model for multi-agent systems using Session as a central runtime abstraction that enables explicit fork, merge, and replay operations while recording comprehensive execution state. | 414 |
| 17 |  MemSlides: A Hierarchical Memory Driven Agent Framework for Personalized Slide Generation with Multi-turn Local Revision
π
Publication Date: Jun 15, 2026
π Paper: https://arxiv.org/pdf/2606.17162.pdf
π Code: https://github.com/huggingface
π Description:
MemSlides presents a hierarchical memory framework for personalized presentation agents that separates long-term user profiles, working memory for session constraints, and tool memory for reusable execution experiences to enable stable personalization and reliable local edits. | 441 |
| 18 |  Formalizing Latent Thoughts: Four Axioms of Thought Representation in LLMs
π
Publication Date: May 7, 2026
π Paper: https://arxiv.org/pdf/2606.27378.pdf
π Code: N/A
π Description:
An axiomatic evaluation framework reveals systematic failures in latent thought representations of LLMs across multiple reasoning tasks, demonstrating that current representations fail to satisfy fundamental functional axioms consistently across different model architectures. | 396 |
| 19 | EnterpriseClawBench: Benchmarking Agents from Real Workplace Sessions
π
Publication Date: Jun 22, 2026
π Paper: https://arxiv.org/pdf/2606.23654.pdf
π Code: https://github.com/huggingface
π Description:
EnterpriseClawBench presents a benchmark for enterprise agents based on real-world sessions with 852 reproducible tasks, emphasizing comprehensive evaluation metrics beyond single performance scores. | 418 |
| 20 |  Heterogeneous Scientific Foundation Model Collaboration
π
Publication Date: Apr 30, 2026
π Paper: https://arxiv.org/pdf/2604.27351.pdf
π Code: https://github.com/Violet24K/Eywa
π Description:
Eywa is a heterogeneous agentic framework that extends language-centric systems to scientific foundation models by integrating domain-specific models with language-based reasoning interfaces for improved performance across diverse scientific domains. | 444 |
