✨VideoFlexTok: Flexible-Length Coarse-to-Fine Video Tokenization

📝 Summary:
VideoFlexTok enables efficient video representation through variable-length token sequences that capture abstract information first, followed by fine-grained details, allowing for reduced computationa...

🔹 Publication Date: Published on Apr 14

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2604.12887
• PDF: https://arxiv.org/pdf/2604.12887
• Github: https://github.com/apple/ml-videoflextok

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

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

#AI #DataScience #MachineLearning #HuggingFace #Research

✨PersonaVLM: Long-Term Personalized Multimodal LLMs

📝 Summary:
PersonaVLM introduces a framework for long-term personalized multimodal LLMs. It remembers interactions, reasons multi-turn using retrieved memories, and aligns responses with evolving user personality. This novel method significantly outperforms baselines and GPT-4o on a new evaluation benchmark.

🔹 Publication Date: Published on Mar 20

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2604.13074
• PDF: https://arxiv.org/pdf/2604.13074
• Project Page: https://personavlm.github.io/
• Github: https://github.com/MiG-NJU/PersonaVLM

🔹 Models citing this paper:
• https://huggingface.co/ClareNie/PersonaVLM

✨ Datasets citing this paper:
• https://huggingface.co/datasets/ClareNie/Persona-MME
• https://huggingface.co/datasets/ClareNie/PersonaVLM-Dataset

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

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

#LLM #MultimodalAI #PersonalizedAI #AIResearch #MemoryAI

✨VEFX-Bench: A Holistic Benchmark for Generic Video Editing and Visual Effects

📝 Summary:
VEFX-Bench offers a large human-annotated video editing dataset and VEFX-Reward, a specialized model for quality assessment. This benchmark allows standardized comparison, showing current models struggle with instruction following and edit locality.

🔹 Publication Date: Published on Apr 17

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2604.16272
• PDF: https://arxiv.org/pdf/2604.16272
• Project Page: https://xiangbogaobarry.github.io/VEFX-Bench/

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

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

#VideoEditing #VFX #AI #ComputerVision #Benchmarks

✨Qwen3.5-Omni Technical Report

📝 Summary:
Qwen3.5-Omni is a large multimodal model excelling in audio-visual understanding and generation, achieving SOTA results across many benchmarks. It features a Hybrid Attention MoE architecture, introduces ARIA for improved speech synthesis, and exhibits a new Audio-Visual Vibe Coding capability.

🔹 Publication Date: Published on Apr 17

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2604.15804
• PDF: https://arxiv.org/pdf/2604.15804

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

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

#MultimodalAI #AIResearch #DeepLearning #GenerativeAI #SpeechSynthesis

✨ArtifactNet: Detecting AI-Generated Music via Forensic Residual Physics

📝 Summary:
ArtifactNet detects AI-generated music by analyzing codec-specific artifacts in audio signals using a lightweight neural network and codec-aware training. It achieves superior performance and efficiency compared to existing methods, establishing forensic physics as a new detection paradigm.

🔹 Publication Date: Published on Apr 17

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2604.16254
• PDF: https://arxiv.org/pdf/2604.16254
• Project Page: https://demo.intrect.io

🔹 Models citing this paper:
• https://huggingface.co/intrect/artifactnet

✨ Datasets citing this paper:
• https://huggingface.co/datasets/intrect/artifactbench

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

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

#AI #MachineLearning #AIMusic #DigitalForensics #AudioProcessing

✨PRL-Bench: A Comprehensive Benchmark Evaluating LLMs' Capabilities in Frontier Physics Research

📝 Summary:
PRL-Bench is a new benchmark evaluating LLMs' end-to-end capabilities in theoretical and computational physics research. It uses 100 curated papers to assess exploration-oriented, long-horizon workflows. Current LLMs perform poorly, revealing a significant gap in autonomous scientific discovery.

🔹 Publication Date: Published on Apr 16

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2604.15411
• PDF: https://arxiv.org/pdf/2604.15411

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

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

#LLMs #PhysicsResearch #ScientificDiscovery #AI #Benchmarking

✨Cut Your Losses! Learning to Prune Paths Early for Efficient Parallel Reasoning

📝 Summary:
STOP is a systematic, learnable token-level path pruning method for Large Reasoning Models. It improves efficiency and accuracy, outperforming baselines and scaling across compute budgets to reduce futile paths in parallel reasoning.

🔹 Publication Date: Published on Apr 17

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2604.16029
• PDF: https://arxiv.org/pdf/2604.16029
• Project Page: https://bijiaxihh.github.io/STOP/
• Github: https://github.com/bijiaxihh/STOP

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

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

#AI #LLM #MachineLearning #ParallelReasoning #ModelEfficiency

✨GTA-2: Benchmarking General Tool Agents from Atomic Tool-Use to Open-Ended Workflows

📝 Summary:
GTA-2 is a new benchmark for General Tool Agents, covering both atomic and real-world, open-ended workflows. It shows frontier models struggle significantly, especially on workflows. The study emphasizes that execution frameworks are crucial for performance, more so than just model capacity.

🔹 Publication Date: Published on Apr 17

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2604.15715
• PDF: https://arxiv.org/pdf/2604.15715
• Github: https://github.com/open-compass/GTA

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

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

#AIAgents #BenchmarkingAI #LLMs #AIWorkflows #AIResearch

✨Learning Adaptive Reasoning Paths for Efficient Visual Reasoning

📝 Summary:
Existing visual reasoning models often overthink, using redundant steps. AVR is an adaptive framework that dynamically chooses efficient reasoning formats. It reduces token usage by 50-90 percent while maintaining accuracy.

🔹 Publication Date: Published on Apr 16

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2604.14568
• PDF: https://arxiv.org/pdf/2604.14568
• Github: https://github.com/RunRiotComeOn/AVR

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

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

#VisualReasoning #AI #MachineLearning #Efficiency #DeepLearning

✨Repurposing 3D Generative Model for Autoregressive Layout Generation

📝 Summary:
LaviGen is a 3D layout generation framework that repurposes 3D generative models. It uses an adapted 3D diffusion model for autoregressive generation, explicitly modeling geometric relations and physical constraints. This achieves superior, more plausible 3D layouts 65% faster than previous methods.

🔹 Publication Date: Published on Apr 17

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2604.16299
• PDF: https://arxiv.org/pdf/2604.16299
• Project Page: https://fenghora.github.io/LaviGen-Page/
• Github: https://github.com/fenghora/LaviGen

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

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

#3DGeneration #DiffusionModels #GenerativeAI #ComputerGraphics #DeepLearning

✨Web Retrieval-Aware Chunking (W-RAC) for Efficient and Cost-Effective Retrieval-Augmented Generation Systems

📝 Summary:
Web Retrieval-Aware Chunking (W-RAC) introduces a cost-efficient framework for web document processing that reduces LLM token usage and hallucination risks through structured content representation an...

🔹 Publication Date: Published on Jan 8

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2604.04936
• PDF: https://arxiv.org/pdf/2604.04936

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

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

#AI #DataScience #MachineLearning #HuggingFace #Research

✨Maximal Brain Damage Without Data or Optimization: Disrupting Neural Networks via Sign-Bit Flips

📝 Summary:
Deep neural networks exhibit catastrophic vulnerability to minimal parameter bit flips across multiple domains, which can be identified and mitigated through targeted protection strategies. AI-generat...

🔹 Publication Date: Published on Apr 16

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2502.07408
• PDF: https://arxiv.org/pdf/2502.07408
• Project Page: https://mkimhi.github.io/DNL/
• Github: https://github.com/IdoGalil/maximal-brain-damage

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

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

#AI #DataScience #MachineLearning #HuggingFace #Research

✨AccelOpt: A Self-Improving LLM Agentic System for AI Accelerator Kernel Optimization

📝 Summary:
AccelOpt is a self-improving LLM agentic system that autonomously optimizes kernels for AI accelerators using iterative generation and optimization memory, achieving significant throughput improvement...

🔹 Publication Date: Published on Apr 15

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2511.15915
• PDF: https://arxiv.org/pdf/2511.15915
• Project Page: https://ppl.stanford.edu/accelopt.html
• Github: https://github.com/zhang677/AccelOpt

🔹 Models citing this paper:
• https://huggingface.co/Genghan/sft-qwen-7b-instruct_GRPO_nki_pure_0920_cluster3
• https://huggingface.co/Genghan/deepseek-coder-33b-instruct_GRPO_nki_pure_0907_cluster1
• https://huggingface.co/Genghan/sft-deepseek-coder-33b-instruct_GRPO_nki_pure_0921_cluster4

✨ Datasets citing this paper:
• https://huggingface.co/datasets/Genghan/NKIBench

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

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

#AI #DataScience #MachineLearning #HuggingFace #Research

✨NTIRE 2026 Challenge on Video Saliency Prediction: Methods and Results

📝 Summary:
This paper overviews the NTIRE 2026 Challenge on Video Saliency Prediction. Participants developed automatic saliency map prediction for videos using a novel 2,000-video dataset with crowdsourced fixations. Over 20 teams submitted, and all challenge data is now publicly available.

🔹 Publication Date: Published on Apr 16

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2604.14816
• PDF: https://arxiv.org/pdf/2604.14816
• Project Page: https://www.codabench.org/competitions/12842/
• Github: https://github.com/msu-video-group/NTIRE26_Saliency_Prediction

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

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

#VideoSaliency #ComputerVision #NTIRE #MachineLearning #SaliencyPrediction

✨TIPSv2: Advancing Vision-Language Pretraining with Enhanced Patch-Text Alignment

📝 Summary:
Enhanced vision-language models achieve superior dense patch-text alignment through improved pretraining techniques including patch-level distillation, modified masked image objectives, and optimized ...

🔹 Publication Date: Published on Apr 13

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2604.12012
• PDF: https://arxiv.org/pdf/2604.12012
• Project Page: https://gdm-tipsv2.github.io/

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

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

#AI #DataScience #MachineLearning #HuggingFace #Research

✨(1D) Ordered Tokens Enable Efficient Test-Time Search

📝 Summary:
This paper demonstrates that 1D ordered, coarse-to-fine token structures enhance test-time search in autoregressive models. These tokens allow better verifier evaluation of intermediate states, improving scaling and enabling training-free text-to-image generation through pure test-time search. To...

🔹 Publication Date: Published on Apr 16

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2604.15453
• PDF: https://arxiv.org/pdf/2604.15453
• Project Page: https://soto.epfl.ch/
• Github: https://github.com/EPFL-VILAB/search-over-tokens

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

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

#AI #DataScience #MachineLearning #HuggingFace #Research

✨TwinTrack: Post-hoc Multi-Rater Calibration for Medical Image Segmentation

📝 Summary:
TwinTrack framework addresses pancreatic cancer segmentation ambiguity through post-hoc calibration of ensemble probabilities to empirical mean human response, improving calibration metrics on multi-r...

🔹 Publication Date: Published on Apr 17

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2604.15950
• PDF: https://arxiv.org/pdf/2604.15950

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

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

#AI #DataScience #MachineLearning #HuggingFace #Research

✨EdgeDetect: Importance-Aware Gradient Compression with Homomorphic Aggregation for Federated Intrusion Detection

📝 Summary:
EdgeDetect enables efficient and secure federated intrusion detection for 6G-IoT environments through gradient binarization and homomorphic encryption, achieving high accuracy with reduced communicati...

🔹 Publication Date: Published on Apr 16

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2604.14663v1
• PDF: https://arxiv.org/pdf/2604.14663

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

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

#AI #DataScience #MachineLearning #HuggingFace #Research

The first bot in Telegram that offers free
Udemy coupons https://t.iss.one/UdemySybot

✨Elucidating the SNR-t Bias of Diffusion Probabilistic Models

📝 Summary:
Diffusion models suffer from an SNR-timestep bias during inference, impairing generation quality. A differential correction method is proposed that processes frequency components separately. This significantly improves generation quality across various models with minimal computational cost.

🔹 Publication Date: Published on Apr 17

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2604.16044
• PDF: https://arxiv.org/pdf/2604.16044
• Github: https://github.com/AMAP-ML/DCW

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

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

#AI #DataScience #MachineLearning #HuggingFace #Research